Labour Markets In Transition : The Case Of Eastern Europe 9781845446383, 9780861766970

Citation preview

International Journal of Manpower

ISSN 0143-7720 Volume 23 Number 1 2002

Labour markets in transition: the case of Eastern Europe Guest Editor Belton M. Fleisher Paper format International Journal of Manpower includes eight issues in traditional paper format. The contents of this issue are detailed below.

Internet Online Publishing with Archive, Active Reference Linking, Emerald WIRE, Key Readings, Research Register, Institution-wide Licence, E-mail Alerting Service and Usage Statistics. Access via the Emerald Web site: http://www.emeraldinsight.com/ft See overleaf for full details of subscriber entitlements.

Access to International Journal of Manpower online

2

Editorial advisory board ___________________________

3

Abstracts and keywords ___________________________

4

Guest editorial ____________________________________

6

Has the shift toward markets hurt ethnic minorities? Changes in ethnic earnings differentials in Bulgaria’s early transition Lisa Giddings __________________________________________________

9

Ownership and employment in Russian industry: 1992-1995 Susan J. Linz___________________________________________________

32

Regional diversity of employment structure and outflows from unemployment to employment in Poland Aleksandra Rogut and Tomasz Tokarski ____________________________

62

Worker reallocation during Estonia’s transition to market Milan Vodopivec ________________________________________________

77

About the authors _________________________________

98

This issue is part of a comprehensive multiple access information service

CONTENTS

International Journal of Manpower 23,1 2

International Journal of Manpower online An advanced knowledge resource for the entire organization Access via the Emerald Web site – http://www.emeraldinsight.com/ft

Subscribers to this journal benefit from access to a fully searchable knowledge resource stretching far beyond the current volume and issue. International Journal of Manpower online is enhanced with a wealth of features to meet the need for fast, effortless and instant access to the core body of knowledge. Furthermore, this user-friendly electronic library may be networked throughout the subscribing organization to maximize the use and value of the subscription. This is augmented with advanced search facilities and ‘‘choice of access’’ through a variety of alternative journal gateways.

Emerald online access includes: Institution-wide Licence Our liberal licence allows everyone within your institution to access the journals via the Internet, making your subscription more cost-effective. It has been designed to provide a comprehensive, simple system with minimum administration. Access can be granted by IP range or username and password. ATHENS authentication is enabled.

Research Register Research Registers are Internet-based databases where you can identify inside information on research activity worldwide. Whether you are seeking information or you wish to provide details of your own research, you will find this feature invaluable.

Emerald Alert The table of contents e-mail alerting service will e-mail you the contents page of any chosen journal whenever the latest issue becomes available online. For further information please go to http://www.emeraldinsight.com/ alerts

Support Resources A comprehensive range of resources is available online that help users learn how to effectively use online information resources and that helps information professionals market resources to their users. For further information please go to http://www.emeraldinsight.com/ support

Current Issue + Journal Archive Internet access via Emerald Fulltext to information as it becomes available online and to material from past volumes. Users can choose to browse the database for relevant articles or search using the powerful search engine provided.

Active Reference Linking Direct links from the journal article references to abstracts of the most influential articles cited. Where possible, this link is to the full text of the article.

Emerald WIRE A fully searchable subject specific database, brought to you by Emerald Management Reviews, providing article abstracts from the world’s top management, computing and engineering journals. Directly relating to the journal subject area, the database is updated monthly and gives users details of how to obtain the full text of original articles.

Usage Statistics Online Journal Usage Statistics are now available. This feature allows Emerald Administrators to download their usage statistics with regard to their organization’s journal usage. Usage Statistics allow you to review the value of electronic dissemination of your journal subscriptions throughout your organization. They can also help determine the future trends for information within your organization. For further information go to http://www.emeraldinsight.com/stats

Key Readings Abstracts of articles, relating to keywords, are selected to provide readers with current awareness of interesting articles from other publications in the field. The abstracts are available online and are updated regularly upon publication of the latest issue.

Choice of Access Electronic access to this journal is available via a number of channels, enabling end users’ libraries to reach the content through their preferred delivery system. The Emerald Web site – http://www.emeraldinsight.com/ft – is the recommended means of electronic access as it provides fully searchable and high value-added access to the complete content of the journal. Refer to the next section for ‘‘how to access’’ via the Emerald Web site. Subscribers can also access and search the article content of this journal through the following services: Ebsco Online http://www.ebsco.com/online Huber E-Journals http://e-journals.hanshuber.com/english Minerva http://www.minerva.at/ OCLC Firstsearch ‘‘Electronic Collections Online’’ http://www.uk.oclc.org/oclc/menu/eco.htm RoweCom’s ‘‘Information Quest’’ http://www.informationquest.com SilverPlatter http://www.silverplatter.com SwetsBlackwell’s ‘‘SwetsNetNavigator’’ http://www.swetsnetnavigator.nl

How to access this journal through Emerald Organizations must first register for online access (instructions provided at http://www.emeraldinsight. com/register), after which the content is available to everyone within the organization’s domain. To access this journal’s content, simply log on either from the journal homepage or direct through the Emerald Web site. Emerald Customer Support Services For customer service and technical help, contact: E-mail: [email protected] Telephone: (44) 1274 785278 Fax: (44) 1274 785204

EDITORIAL ADVISORY BOARD Professor David J. Bartholomew London School of Economics, UK Professor Derek Bosworth Manchester School of Management, UMIST, UK Professor Martin Carnoy School of Education, Stanford University, USA Professor Peter Dawkins Melbourne Institute for Applied Economic and Social Research, Melbourne University, Australia Professor John Fyfe W.S. Atkins plc, Epsom, UK

Professor Lord Richard Layard Centre for Economic Performance, London School of Economics, UK Professor John Mangan University of Queensland, Brisbane, Australia Professor Stephen L. Mangum Ohio State University, Ohio, USA

Professor Morley Gunderson University of Toronto, Canada Professor Thomas J. Hyclak Lehigh University, Bethlehem, USA

Professor Abraham (Rami) Sagie Bar-Ilan University, Israel

Professor Susan E. Jackson Rutgers University, New Jersey, USA

Professor David Sapsford Management School, Lancaster University, UK Professor P.J. Sloane University of Aberdeen, Aberdeen, Scotland Professor Klaus F. Zimmerman Department of Economics, University of Bonn, Germany

Professor Geraint Johnes Management School, Lancaster University, UK

3

Professor Noah M. Meltz Netanya Academic College, Israel Professor Barrie Pettman International Management Centres, UK, and Founding Editor of International Journal of Manpower

Professor Harish C. Jain McMaster University, Canada

Editorial advisory board

International Journal of Manpower, Vol. 23 No. 1, 2002, p. 3. # MCB UP Limited, 0143-7720

International Journal of Manpower 23,1 4

International Journal of Manpower, Vol. 23 No. 1, 2002, Abstracts and keywords. # MCB UP Limited, 0143-7720

Has the shift toward markets hurt ethnic minorities? Changes in ethnic earnings differentials in Bulgaria’s early transition Lisa Giddings Keywords Bulgaria, Ethnic groups, Wages, Differentiation Relies on cross-sectional survey data from 1986 and 1993 to explain an increase in the ethnic Turk-ethnic Bulgarian earnings differential in Bulgaria in the country’s early transition. Empirical evidence indicates that the ethnic Turks closed both the gap in the number of years of education and experience acquired during this time. Further, the Turks began to enter the growing commerce and transportation industries in the early transition. Shifts in the wage structure, however, favored the ethnic Bulgarians, and these changes outweighed ethnic Turk gains in the measured characteristics. In addition to these shifts, an increase in the overall level of inequality in the labor market punished t h o s e a t t h e l o w - e n d o f th e w a g e distribution, exacerbating the existing ethnic earnings differential. These results imply however, that the ethnic Turks are responding to market signals and if it continues, this trend will diminish the ethnic earnings gap.

Ownership and employment in Russian industry: 1992-1995 Susan J. Linz Keywords Privatization, Employment, Manufacturing industry, Russia, Divestment What impact did privatization have on employment in Russian industry? Utilizes data collected from a panel of 6,205 civilian manufacturing firms in the Central, Volga, North Caucasus, Northern and Western Siberian regions of Russia to explore in more detail the relationship between changes in ownership and employment in Russian industry between 1992 and 1995. In particular, investigates whether change in ownership structure is relatively more important than industry, region, or the

competitive position of the firm in explaining variation in the employment response to changing output conditions during the initial stage of Russia’s transition from plan to market.

Regional diversity of employment structure and outflows from unemployment to employment in Poland Aleksandra Rogut and Tomasz Tokarski Keywords Labour market, Unemployment, Employment, Poland Analyses factors determining the outflows from unemployment to employment across regions in Poland over the years 1992-98, employing the concept of the augmented matching function. Explores also the influence of the economic growth and the employment structure of the regional labour markets in Poland. Concludes that the values of outflows from unemployment to employment are closely and positively related to the number of unemployed and the number of vacancies, as well as to the economic growth rate; and that the employment structure of regional labour markets has a strong impact on outflows from unemployment to employment. The more a regional employment structure resembles the structure in European G7 countries, the higher the outflows from unemployment to employment.

Worker reallocation during Estonia’s transition to market Milan Vodopivec Keywords Market economy, Labour market, Estonia Based on consecutive labor force surveys, this study examines labor market dynamics during the first decade of the Estonian transition to market. The results show that, similar to other transition economies: Estonia’s employment and labor force was reduced; patterns of mobility profoundly

changed – labor market flows intensified and previously nonexistent transitions emerged; and some groups of workers were disproportionally affected, chief among them the less educated and ethnic minorities. But Estonian fundamental free market reforms also produced labor market outcomes that differ significantly from those in other transition economies – above all, the intensity of worker and job flows in Estonia’s transition have surpassed

those in most other transition economies. This was achieved by deliberate policies aimed at stimulating job creation and employment, above all by low employment protection and other policies geared toward increasing employability and strengthening the incentives of workers. Moreover, under the dynamic Estonian labor market adjustment, marginal groups have fared better than those in more protective labor markets of other transition economies.

Abstracts and keywords

5

International Journal of Manpower 23,1 6

International Journal of Manpower, Vol. 23 No. 1, 2002, pp. 6-8. # MCB UP Limited, 0143-7720

Guest editorial About the Guest Editor Belton Fleisher received his PhD degree in economics from Stanford University in 1961. He was on the faculty of the University of Chicago 1961-65 and joined the Ohio State University economics faculty in 1965, where he is currently professor. In 1989 and 1990 he taught economics at the Renmin (People’s) University of China in Beijing. In 1997 he received a Competitive Research Award, for a paper on the Underpricing of Chinese IPOs, from the Sandra Ann Morsilli Pacific-Basin Capital Markets Research Center – a paper presented at the Center’s Ninth Annual International Finance Conference in Shanghai, China. Professor Fleisher has authored and co-authored numerous articles in professional journals and has written seven books, including Labor Economics: Theory and Evidence and Policy. Since 1990, his research has focused on economic growth, financial markets, labour markets, and productivity in the Chinese economy. He currently serves on the editorial board of the Journal of Comparative Economics, and he is a co-editor of China Economic Review.

Labour markets in transition: the case of Eastern Europe In 1989 a political revolution in the Soviet Union and its Eastern European satellites led to the dismantling of Communist controls and transition toward free markets in goods and services and factors of production. Surely the most critical link between the new opportunities for entrepreneurial profits and the welfare of average citizens has been the transformation of labour markets. Workers had to trade secure, if low, wages for greater earnings opportunities accompanied by sharply increased income risk due to involuntary unemployment. The severity and balance of the tradeoff has varied considerably among the affected countries. Perhaps a reason that Russia, by far the largest and most powerful, delayed formal market reform until January 2, 1992 was recognition that it faced the most daunting obstacles to successfully integrating workers and firms under a system of markets. Evidence that such a fear would have been warranted is found in average annual growth rates of per-capita GNP. World Bank data show an annual decline of 6.6 per cent in the Russian Federation between 1989 and 1999, compared, for example, to positive growth of 4.7 per cent in Poland, with most other countries falling in between. Even Poland’s superior economic-growth performance did not prevent high unemployment, however. As productivity growth approached 5.5 per cent and GDP growth slowed to less than 5 per cent after 1998, unemployment in Poland increased, and by the year 2000 had reached 16 per cent of the labour force. Not only has there been substantial variation in economic growth rates, but also there have been many differences in measures taken to protect workers from drowning in the sea of competition. Much research has investigated the degree to which such policies, as well as initial conditions, have affected the success of transition. The four papers in this volume of the International Journal of Manpower all shed light on some aspect of the success achieved in the Russian Federation and in three Eastern European economies in the early stages of transition. In the first paper, Lisa Giddings investigates a special case of a general problem addressed by many labour economists: how do disadvantaged groups fare under labour-market competition? As Giddings points out in her introduction, despite the high frequency of various ethnic minorities in Eastern

Europe and the Russian Federation, very little research has focused on their experience during the transition. The particular subject of her investigation is Bulgaria’s ethnic Turkish minority. Over the period 1986-93, the average Turkish worker’s earnings declined from approximately 85 per cent of an average Bulgarian’s earnings to approximately 80 per cent. The principal reason for this decline appears to be a disproportionately small representation of ethnic Turks in transportation, commerce, and service industries, which experienced relatively high earnings growth. Services, which were suppressed under Communism, experienced relative expansion, and relative wages grew in response to an increase in the relative demand for labour. However, it appears that ethnic Turks were not prevented from finding employment in the growing sector. If past discrimination had suppressed their employment there, they appear to have experienced improved employment opportunities after market reform. An intriguing result is that while the marginal rate of return to schooling increased during the early transition for ethnic Bulgarians, it declined for ethnic Turks. I might speculate that less educated ethnic Turks are more mobile than their better educated counterparts because of smaller specific human capital investments and thus were able to take better advantage of improved earnings opportunities during early transition. Whatever the answer, this intriguing puzzle invites further investigation and I hope that publication of this paper will elicit further research if, indeed, Giddings is not already well into it. Susan Linz’s paper tackles the giant country with the giant problem, the Russian Federation. The major challenge of Linz’s research is to explain why, in the presence of dramatic production declines, did employment in Russian industry fall so little? According to official Goskomstat statistics (www.nap.edu/html/transform/ch9_t2.htm) the employed population fell by only 7 per cent between 1992 and 1995, while GDP was falling 6.2 per cent per year between 1989 and 1999. She has developed a panel of data consisting of 2,033 firms that were in business in both 1992 and 1995, employing over 3,000,000 workers in 1992. Industries represented include food processing, machine building, light industry, wood/forestry/paper, construction materials, and printing. Regions include Central, Volga, North Caucasus, Western Siberia, and Northern. Linz analyzes employment changes as resulting from two sources: the initial deviation from optimal employment and subsequent changes in output. Her empirical results show that firms with high initial labour productivity reduced employment less than those with what was evidently greater hidden unemployment in their workforces. She also finds that newly privatized firms responded much more sensitively to changing output than did firms that remained state-owned. She also finds that firms exposed to more competitive pressures, e.g. due to location and/or exposure to international import/export competition appear to have reduced costs more effectively than their more sheltered counterparts. Noting that Linz was forced to confine her analysis to firms that did not cease to exist or change industry during her sample period, we hope that future research will soon tackle the question of the role of firm dynamics in Russia’s employment puzzle.

Guest editorial

7

International Journal of Manpower 23,1 8

Poland’s economic transition has been, judging from the data on economic growth, far less traumatic than Russia’s. Nevertheless, Poland experienced an initial ‘‘demand shock’’ and rising unemployment, followed, after 1992, by both output growth and declining unemployment. Aleksandra Rogut and Tomasz Tokarski, like Susan Linz, offer an interpretation of employment dynamics, focusing in their paper on determinants of regional diversity in flows from unemployment to employment. They aim to document factors that facilitated relatively rapid adjustment of the labour force to the competitive pressures of the reformed economy and labour market. Their primary tool of analysis is an augmented matching function, which relates the regional reemployment of unemployed workers to a variable that measures how closely the employment structure in the region resembles the employment structure of more advanced (specifically, G7) European countries. Holding this measure of regional labourforce matching constant, they also find that outflows from unemployment to employment are larger when the number of unemployed workers and job vacancies are larger and when economic growth is greater. Rogut and Tokarski draw several policy implications form their research which are relevant not only for Polish policy makers but also for officials in other transitional economies that seek to reduce the negative labour-force impacts of the move from controlled to free labour markets. One of them is the need to establish a housing market. They find that there is a housing barrier in Poland that severely limits the geographical mobility of workers, thus accentuating any mismatching of regional employment structure with the regional structure of labour demand They find that informal job seeking outside the formal labour-office bureaucracy has been an important means of leaving unemployed status. This would seem to imply that facilitation of informal information flows is a good idea. They also advocate government policies that subsidize bringing jobs to people in areas where the employment structure is particularly badly matched with existing job vacancies. Research reported in the fourth paper, by Milan Vodopivec, is in the nature of a ‘‘natural experiment,’’ in that he compares the labour-force reallocation process in Estonia, which adopted relatively liberal labour-market policies in its transition, with that in Slovenia, which attempted much more to protect its workers from the risks of the new market economy. Vodopivec reports that labour-force dynamics (job creation, accessions, quits, job-to-job moves) during transition have been high, approaching and even exceeding those mature market economies; they have been much higher than in Slovenia. He believes that Estonia’s overall favorable labour-market experience during transition and the experience of certain groups of workers, specifically, young workers and workers on fixed-term contracts, support the theoretical model of Blanchard (1998), which attributes labour-market stagnation to regulations that protect workers from exposure to unemployment risk. On the negative side, Vodopivec notes that ethnic minorities appear to have suffered during Estonia’s transition. In the light of Giddings’ paper in this issue, a close comparison of the Bulgarian with the Estonian experience would be an interesting exercise. Belton Fleisher

The research register for this journal is available at http://www.emeraldinsight.com/researchregisters

The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0143-7720.htm

Has the shift toward markets hurt ethnic minorities? Changes in ethnic earnings differentials in Bulgaria’s early transition

Shift toward markets

9

Lisa Giddings University of Wisconsin-La Crosse, College of Business Administration, La Crosse, Wisconsin, USA Keywords Bulgaria, Ethnic groups, Wages, Differentiation Abstract Relies on cross-sectional survey data from 1986 and 1993 to explain an increase in the ethnic Turk-ethnic Bulgarian earnings differential in Bulgaria in the country’s early transition. Empirical evidence indicates that the ethnic Turks closed both the gap in the number of years of education and experience acquired during this time. Further, the Turks began to enter the growing commerce and transportation industries in the early transition. Shifts in the wage structure, however, favored the ethnic Bulgarians, and these changes outweighed ethnic Turk gains in the measured characteristics. In addition to these shifts, an increase in the overall level of inequality in the labor market punished those at the low-end of the wage distribution, exacerbating the existing ethnic earnings differential. These results imply however, that the ethnic Turks are responding to market signals and if it continues, this trend will diminish the ethnic earnings gap.

I. Introduction Has the transition from plan to market in East and Central Europe imposed greater hardship on ethnic minorities? Labor market studies to date have demonstrated (with some degree of variance among countries) that economies in the transition are experiencing wage decompression, increasing returns to schooling and increased inequality as the countries shift toward markets. Very little research has focused on the effects of the transition on ethnic groups in East and Central Europe. Studies on changes in wage differentials and changes in industry and occupational segregation have mainly covered differences between men and women[1]. Given evidence from other contexts suggesting that ethnic and racial minorities bear the brunt of drastic economic changes, this absence in the literature is curious. Nothing about the transformation from plan to market would imply differently. In 1986 the average ethnic Turkish worker in Bulgaria earned approximately 85 percent as much per month as the average ethnic Bulgarian worker. By 1993, Turks earned about 80 percent as much per month. The goal of this essay is to account for the increase in the ethnic earnings gap in Bulgaria’s early transition within the broader context of the wageThe author would like to thank Mieke Meurs, Bob Lerman, Steve Pressman, T.J. Brooks, Belton Fleisher and two anonymous reviewers for their excellent comments and suggestions on earlier drafts of this paper.

International Journal of Manpower, Vol. 23 No. 1, 2002, pp. 9-31. # MCB UP Limited, 0143-7720 DOI 10.1108/01437720210421286

International Journal of Manpower 23,1 10

structure changes documented in transitioning economies. In particular, the analysis distinguishes between the parts of the increased ethnic earnings gap caused by factors that are ‘‘group-specific’’ from the component caused by changes in the structure of wages, where ‘‘group-specific’’ factors represent the degree to which differences in skills between ethnic Bulgarians and ethnic Turks or differences in the industry in which one group dominates over the other contribute to the increase in the earnings gap. ‘‘Wage-structure’’ factors consist of the portion of the gap explained by changes in skill prices (such as returns to education or returns to working in a particular industry), and changes in the overall level of inequality in an economy. The empirical research relies on cross-sectional survey data from 1986 and 1993. This allows for a comparison between pre and early transition in Bulgaria, as the initial measures of the economic reform were implemented in 1991 when prices and wages were liberalized. The 1986 survey is unique in that it is the only existing pre-transition survey from Bulgaria, thereby enabling an analysis of the heretofore unexplored Bulgarian wage structure prior to 1989. The change in the earnings differentials between 1986 and 1993 is then decomposed using the Juhn et al. (1991) technique in order to identify the important factors involved in causing the change. The earnings decomposition indicates that the returns to schooling increased in Bulgaria’s early transition. Despite recent Turkish catch-up in years of schooling, Bulgarians still acquire more education, and the increased returns to schooling favored the Bulgarians over the Turks. With more of an asset that has become more valuable, Bulgarians expanded their earnings gap over the Turks. Furthermore, the decomposition shows increased relative earnings in the commerce and service industries. Similarly, although the Turks began to enter into the service industry, the Bulgarians dominated both of these growing industries and increased their relative earnings as a result. These results imply that the ethnic Turks are responding to market signals by obtaining more education and entering into growing industries. If this trend continues, it is expected that the ethnic earnings gap will diminish as the country continues to move toward a market economy. This essay is organized as follows. Section II provides a brief history of the ethnic Turks in the region, as well as background on the period of the study including an overview of the transition in Bulgaria, economic indicators of the period of the study, and changes in labor market policies. Section III presents descriptive statistics. Section IV provides a description of the methodology employed. Empirical results are presented in Section V and concluding remarks in Section VI. II. Background Located in the Balkan Peninsula, modern Bulgaria is one of the smaller countries of Europe, with a population of approximately 9 million people and an area of 43,000 square miles. A centrally planned economy coupled with a single-party political system dominated by the Communist Party characterized

the Bulgarian post-war socio-economic policy until the ousting of President Todor Zhivkov[2] in November of 1989. The labor market consisted of a welleducated workforce, guaranteed employment, and it prioritized a strategy of rapid industrial growth. An ideology of equality limited wage inequality, and workers enjoyed a strong social safety net that included extensive child-care benefits and paid maternity leave. Cracks emerged in the economy by the mid1980s and official statistics indicated a marked decline in productivity and national output, together with high rates of inflation. This set the stage for the collapse of central planning and the introduction of market mechanisms. During the immediate period following the fall of Zhivkov in November 1989, marking the formal end of central planning, Bulgaria’s budget deficit grew, the money supply remained unconstrained, and average nominal wages grew by 173 percent until the end of 1990 (Enev and Koford, 1997, p. 86). In early 1991 a package of fiscal and monetary policies was introduced in the Polish style ‘‘big bang shock therapy’’ to open and stabilize the economy (World Bank, 1996, p. 23). In February, wages were decentralized and the prices of 90 percent of goods were liberalized[3]. These liberalization and stabilization programs were introduced under extremely unfavorable conditions. The Bulgarian economy had been highly dependent on Council of Mutual Economic Assistance (CMEA) markets and was, consequently, heavily penalized by their collapse[4]. Not only were trade relations with the Commonwealth of Independent States (CIS) halted, the UN embargo against the former Yugoslavia limited Bulgaria’s access to Western European markets. The Gulf crisis worsened the situation as Iraq’s debt was Bulgaria’s biggest outstanding foreign loan (Beleva et al., 1993, p. 31). Like other countries in transition, Bulgaria experienced a huge one-time price jump when prices were initially liberalized and real output fell. Table I shows annual inflation jumping from 22 percent to over 300 percent in 1991, decreasing to 82 percent in 1992 and further to 73 percent 1993. Registered unemployment rates in Bulgaria were among the highest of any of the East and

Nominal GDPa Real GDPa Labor productivityb Average annual inflation ratesc Wages in state sectora (percent change in annual average) Unemploymenta (% of labor force)

1989

1990

1991

1992

1993

1994

39.58

– 8.7

45.39 (15%) 90.90 (– 9%) – 3.2 22.0 31.7

135.71 (199%) 80.30 (– 11%) 1.5 333.5 165.6

200.83 (48%) 74.40 (– 6%) 0.9 82.0 103

298.93 (49%) 72.60 (– 2%) – 0.8 72.8 59

543.47 (82%) 73.60 (1.4%) 3.5 89.0 50.8

–

1.5

11.5

15.6

16.4

12.8

100.00

Notes: aSource: Jones and Miller (1997); bSource: The World Bank, 1996, p. 172; cPercentage increases in the CPI. Source: World Bank, 1996; dRegistered unemployment rate (%). Source: National Employment Service, quoted in Bobeva (1997, p. 29)

Shift toward markets

11

Table I. GDP, inflation, unemployment and labor productivity (percentage increases over previous year)

International Journal of Manpower 23,1

Central European countries in transition (Beleva et al., 1995, p. 219). National average of annual unemployment rose from pre-1989 negligible levels to over 11 percent in 1991, over 15 percent in 1992, and peaked at 17 percent in 1993[5]. The foreign trade account reached a low in 1993 at a deficit of $530 million and then became positive again in 1994 at $152 million (Bobeva, 1997, p. 25).

12

A brief history on the experience of the ethnic Turks in Bulgaria Ethnic Bulgarians and ethnic Turks in Bulgaria have experienced a checkered relationship dating back at least to the Ottoman Empire when the Turks conquered Bulgaria at the end of the fourteenth century (Crampton, 1997, p. 31). After Bulgarian independence in 1878, many Turks remained, comprising 9.7 percent of the population[6]. Since that time ethnic relations have undergone waves of intolerance in which the Turks[7] were denied the political, social, and economic rights accorded to them in the provisions of the Treaty of Berlin of 1878 (Karapat, 1990, p. 7). The communist takeover in 1944 proved to be a mixed blessing for the ethnic Turks in Bulgaria. Their position in society improved considerably; the number of Turkish schools increased, and local Turkish literature emerged along with periodicals and newspapers in every major city (Karapat, 1990, p. 7). Concurrent with the ethnic Turkish cultural rebirth, however, the Communist Party downplayed ethnic differences in an official program of ‘‘ethnic unification’’ in which minorities in Bulgaria were assimilated into mainstream culture (Eminov, 1997, pp. 52, 84). From 1951 to 1952, approximately 155,000 Turks were forced to leave Bulgaria (Karapat, 1990, p. 4). This was the first of at least six forced expulsions of the ethnic Turks from Bulgaria during communism and part of a larger program including a name-changing campaign designed to eradicate cultural differences in Bulgaria. The Communist Party’s Central Committee designed such measures in order to reach priobshtavne or homogeneity in the population, and to establish edinna bulgarska natsiya or ‘‘one Bulgarian nation’’ (Karapat, 1990, p. 4). The last occurred in the final months of President Zhivkov’s regime, when the Party began a ‘‘regeneration process’’ in an attempt to finally resolve the ‘‘Turkish problem’’ (Eminov, 1997, pp. 96-7). In 1989 many ethnic Turks were expelled, including intellectuals and leaders[8]. These actions have periodically put Bulgaria on the United Nation’s Development Programme’s Human Rights Watch List, and the early transition has changed the situation only marginally (UNDP, 1998). A survey conducted by a sociological collective in June of 1992 found that the ethnic Bulgarians sustain a prejudice against the ethnic Turks. The survey showed that over 50 percent of Bulgarians considered the ethnic Turks to be a ‘‘real danger to national security’’, religious fanatics, and occupying too many important political positions. Over 35 percent supported policies encouraging emigration to Turkey (Kunev, 1992, p. 47). In terms of the labor market, an ideology of equality, coupled with a system of centralized wage-setting limited the level of earnings inequality during

communism. In order to limit any negative social consequences of wage and price liberalization, several new institutions developed in the early transition to limit increases in earnings inequality. These included a tax-based incomes policy, a minimum wage, and collective bargaining at the national, regional, and enterprise levels. A tax-based incomes policy was introduced after price liberalization in 1991 on behalf of a tripartite agreement among government, employers, and the largest national trade union federation (Enev and Koford, 1997, p. 82)[9]. Beginning in the second quarter of 1991, direct controls on wages were applied (Tzanov and Vaughan-Whitehead, 1997, p. 100). Based on actual and expected inflation, the controls imposed a ceiling on the nominal wage bill of individual enterprises in order to limit wage growth (Enev and Koford, 1997). The government planned to enforce the policy by levying a highly progressive tax against excessive wage increases[10]. In reality, enterprises were able to avoid paying the tax (Enev and Koford, 1997). In addition to the incomes policy, public employee wages were controlled through the maintenance of a universal minimum wage for all full-time workers. Under trade union pressure, the state increased the minimum wage in 1990 from 180 to 191 Leva per month. In 1992, the minimum wage increased to 850 Leva per month (equivalent to $30). Although it was adjusted irregularly, the real value of the minimum wage did not keep up with inflation (Tzanov and Vaughan-Whitehead, 1997, pp. 100-101). The real value of the minimum wage based on 1989 Leva fell from 180 in 1989 to 71.3 Leva per month in 1993. During the first years of the communist period (from the 1940s to 1951) collective bargaining over wages was forbidden in Bulgaria while non-wage issues were negotiable. During the early transition (between 1989 and 1993) a pluralist system of industrial relations developed within the context of continuous political upheaval[11]. Trade unions and employers began to distance themselves from the State and the Communist Party (Thirkell and Tseneva, 1992) and union activities were no longer subordinated to the economic goals of the state (Iankova, 2000). By 1993 four levels of negotiation were present in Bulgaria (Iankova, 2000). These included national tripartite negotiations between the government, the national trade union organizations, and the national organizations of employers, industry, and sector level negotiations, regional/local negotiations, and bargaining at the enterprise level including representatives of the state, union, and employer organizations. By July of 1991, 37 percent of state enterprises had negotiated collective bargaining agreements, and by November of that year about 74 percent of the wage-earners were covered by collective agreements. III. Data and descriptive information This analysis is based on two cross-sectional household surveys conducted in Bulgaria, one administered prior to the transition and the other after the initial economic reforms were in place. The 1986 Town and Village Survey, conducted by the Institute of Sociology of the Bulgarian Academy of Sciences in Sofia

Shift toward markets

13

International Journal of Manpower 23,1 14

(Institute of Sociology, 1986), was carried out in conjunction with the national census delivered in the winter of 1985 and contains heretofore unexplored information on the structure of Bulgarian wages prior to the transition. The sample is random and representative of the population. After eliminating those not in the labor force, and those who reported zero earnings[12], the sample contains 6,545 useable cases. The second survey, Social Stratification in Eastern Europe after 1989: General Population Survey (SSEE)[13], was conducted in 1993 by a team of investigators from University of California-Los Angeles (UCLA) in conjunction with Bulgarian officials from the Central Statistical Office. The SSEE survey was carried out through a two-stage clustered sampling technique with stratification by district, municipality, and size of voting section. The field research took place between June and July of 1993 during which 4,921 interviews were completed. The sample contains 2,828 usable cases. The empirical analysis is restricted to employed civilians of working age. In pretransition Bulgaria, this consisted of those aged 16 to 60, while no upper age restrictions are placed on the 1993 sample. Those working in agriculture and the self-employed individuals are included in this analysis. Earnings were not adjusted for hours spent in the labor market in either the 1986 or the 1993 analysis. The number of hours worked was only available in the 1993 survey; however, it is a reasonable assumption that all employed individuals during communism worked full-time. Normal working hours per week were generally longer in Central and Eastern than in Western European countries; in particular, the average workweek in Bulgaria during communism was approximately 40.5 for both men and women (Kroupova, 1990, p. 10, quoted in Hubner, Maier and Rudolph, 1991, p. 34). Adjusting for hours in the 1993 analysis made no significant difference in the results[14]. Respondents in the SSEE survey were permitted to record incomes either as annual or as monthly. The variable used in this analysis includes monthly incomes with imputed missing and implausible values (implausible because they are values that could not realistically occur in Bulgaria given the other characteristics of the respondent, and hence almost certainly constitute errors in reporting or recording). The imputation was made by assigning expected values from a regression of each income variable being imputed on a set of characteristics known to determine incomes[15]. Changes in the ethnic earnings differentials during Bulgaria’s transformation from a planned economy to a mixed-market economy are reported in Table II, which provides information on log earnings for ethnic Bulgarians and ethnic Turks. According to the table, the ethnic earnings gap increased over that time period from 0.1618 to 0.1850, or an increase of 0.0232 log points. This evidence indicates that the log ethnic earnings differential increased by over 14 percent over the period. The effects of the transition on ethnic minorities in the labor market are not consistent across transitioning countries. In Estonia, for example, ethnic Russians are performing relatively better in the early years of the transition.

The mean log wage differential between male ethnic Estonians and male ethnic Russians fell from 0.0186 in 1989 to 0.0028 in 1994 (Kroncke and Smith, 1999, p. 189). The non Slovenian males in Slovenia experienced a slight decline in their earnings premium from 2.3 percent in 1987 to 2 percent in 1992 whereas non Estonian males in Estonia earned less than their Estonian counterparts in both 1989 (2.9 percent less) and 1994 (7.9 percent less) (Orazem and Vodopivec, 2000: p. 291). Data from urban China indicates that earnings premiums for a male national minority fell from 1.8 percent in 1988 to –3.9 percent in 1995 (Gustafsson and Li, 2000, pp. 320-2). Such inconsistent results across the transition may reflect differences in skills and access to education between the majority and minority groups in a country.

Shift toward markets

15

Explaining the increase in the ethnic earnings gap ‘‘Group-specific’’ factors. Economists traditionally approach the question of wage differentials between men, women, and among racial/ethnic groups by focusing on the role of differences in qualifications between gender and ethnic groups, and on differences in the treatment of similarly qualified individuals (discrimination). Segregation by occupation, industry, or sector in which members of different gender, racial, or ethnic groups are disproportionately employed in particular occupations (industries or sectors) can also lead to differences in wages. Such differences between groups in human capital or occupational characteristics, can be referred to as group-specific factors – where groups are based on gender, race, or ethnicity – that influence wage differentials (Blau and Kahn, 1996)[16]. The following tables explore changes in such group-specific factors over time in an attempt to explain some of the increase in the ethnic earnings gap. Table III presents changes in human capital characteristics among ethnic Turks and ethnic Bulgarians between 1986 and 1993. The information in this table indicates that the ethnic Bulgarians acquired more schooling than the Turks in both 1986 and 1993 (10.72 years versus 7.51 years and 11.49 years versus 8.59 years respectively). However, the Turks began to close the education gap by 1993. The Turks also closed the gap in labor market experience. This evidence indicates a probable decline in the ethnic earnings

1986 1993 Gap93 – Gap86

Log of Bulgarian earnings

Log of Turk earnings

Ethnic gap

5.3101 (0.347) 7.6395 (0.457)

5.1483 (0.371) 7.4544 (0.417)

0.1618 0.1850 0.0232

Notes: Data is from a sample of the employed. Log wages are reported in nominal terms and reflect the massive realignment of prices in the transition. Standard deviations are in parentheses

Table II. Changes in log ethnic earnings differentials, Bulgaria, 1986-1993

International Journal of Manpower 23,1 16

Table III. Pre and early transition education and experience distribution of ethnic Bulgarians and ethnic Turks in the labor market, Bulgaria, 1986-1993

gap and therefore, does not explain the observed increase in earnings differentials. In addition to education and experience, systematic differences between the Bulgarians and the Turks in the industry in which they are predominately employed may explain some of the increase in the ethnic earnings gap. Table IV presents the distribution of ethnic Bulgarians and ethnic Turks in the labor market as well as the change in employment in each industry. This information indicates that commerce, services, and transportation experienced the most growth in employment in the early transition. In contrast, manufacturing, construction, agriculture, public administration, and communications saw a decline in overall employment over this time. Employment among the ethnic Bulgarians grew in commerce, services, and transportation in the early transition whereas employment among the ethnic Turks grew in construction, services, transportation, and communications. This indicates that with the exception of the commerce industry, the ethnic Turks, like the ethnic Bulgarians, are moving into the growing industries in the transition. Therefore, even though the Turks do not dominate these industries in an absolute sense, they are adopting to changes in industrial demand and benefiting from growth in services and transportation. As was All Change Years of schooling Experience N

0.939 4.653

Ethnic Bulgarians 1986 1993 Change 10.716 11.485 16.792 21.192 5,464 2,392

Ethnic Turks 1993 Change

7.510 14.557 64

8.591 21.939 259

1.081 7.382

Notes: Columns may not sum to 1.0 due to rounding. Data is from a sample of the employed

All % change

Table IV. Pre and early transition industry distribution of ethnic Bulgarians and ethnic Turks in the labor market, Bulgaria, 1986-1993

0.769 4.4

1986

Manufacturing Construction Agriculture Commerce Services Public administration Transportation Communications N

– 0.1578 – 0.1398 – 0.0684 0.8289 0.4074 – 0.4082 0.2919 – 0.0609

Ethnic Bulgarians 1986

1993

0.3561 0.2972 0.0769 0.0598 0.1332 0.1313 0.0758 0.1421 0.0937 0.1313 0.0686 0.0393 0.0641 0.0861 0.1277 0.1129 5,464 2,392

% change – 0.1654 – 0.2223 – 0.0148 0.8760 0.4009 – 0.4274 0.3445 – 0.1164

Ethnic Turks 1986

1993

0.2775 0.2625 0.0682 0.1042 0.4186 0.3822 0.0481 0.0425 0.0605 0.0811 0.0233 0.0154 0.0496 0.0541 0.0512 0.0579 645 259

% change – 0.0540 0.5282 – 0.0869 – 0.1163 0.3410 – 0.3359 0.0895 0.1320

Notes: Columns may not sum to 1.0 due to rounding; Note that the industry groups refer to 1-digit industry codes. Data from a sample of the employed

the case with education and experience, this evidence does not explain the increase in the ethnic earnings gap. Changes in industrial demand as the economy shifts toward markets may favor certain industries. Such changes could benefit one group over another due to industrial segregation in an economy. Table V presents changes in earnings by industry between 1986 and 1993. Earnings grew fastest in the commerce, services, and communications industries. The disproportionate employment of ethnic Bulgarians in these sectors has contributed to their relative earnings gains in the early transition. However, with the exception of commerce, the Turks are entering into these favored industries. Thus, if this trend continues, the earnings disadvantage of ethnic Turks will wither and perhaps disappear. ‘‘Wage-structure’’ factors. In addition to group-specific factors, labor market institutions associated with the overall wage structure in an economy can affect wage differentials play an important role in generating wage differentials (Blau and Kahn, 1992, 1995, 1996, 1997, 1999). What they call ‘‘wage-structure’’ factors refer to how different characteristics (such as schooling, experience, or industry) are rewarded in an economy, as well as changes in the overall level of inequality in an economy. Table VI reports results from ordinary least squares human capital earnings functions (Mincer, 1974) in order to examine changes in the returns to human capital variables and differences in returns between ethnic Bulgarians and ethnic Turks in the early transition in Bulgaria. The natural log of earnings is the dependent variable and the two regressions include independent variables associated with a traditional human model. The ‘‘basic’’ regression includes continuous variables for years of schooling and years of work experience (and its square). The ‘‘extended’’ regression includes variables for education, work, and dummy variables for region and industry. The regression results indicate that as Bulgaria moved toward a market economy, the returns to a year of schooling among ethnic Bulgarians increased from approximately 3 percent in 1986 to over 4 percent by 1993. In contrast, the returns to a year of schooling fell for ethnic Turks from 3.9 percent to less than

Manufacturing Construction Agriculture Commerce Services Public administration Transportation Communications

All

Bulgarians

Turks

0.4302 0.4220 0.4363 0.5039 0.4730 0.3988 0.4303 0.4494

0.4283 0.4233 0.4218 0.5053 0.4753 0.4018 0.4310 0.4483

0.4495 0.4181 0.4543 0.4924 0.4142 0.3631 0.4252 0.4662

Notes: Data from a sample of the employed. Log wages are reported in nominal terms and reflect the massive realignment of prices in the transition

Shift toward markets

17

Table V. Relative earnings growth by industry of employment (percentage change in nominal log earnings in Bulgaria 1986-93)

International Journal of Manpower 23,1 18

Bulgarian 1986 1993 Basic Extended Basic Extended Intercept Years of schooling Experience Experience2 Sex Married Sofia State Private Const Ag Comrce Svcs Pubadmin Trans Communi

Table VI. Baseline log earnings functions and changes in returns to skill in the early transition, Bulgaria, 1986-1993

R2 F-test n

Turk 1986 1993 Basic Extended Basic Extended

4.7245 4.5750 7.0939 6.8505 4.6152 4.6283 7.3530 7.3036 0.0325 0.0326 0.0421 0.0417 0.0392 0.0291 0.0096 0.0099 (21.98) (22.72) (14.96) (14.53) (8.13) (5.97) (0.78) (0.86) 0.0201 0.0259 0.0071 0.0096 0.0258 0.0237 0.0046 0.0051 (12.48) (17.22) (2.88) (3.91) (5.01) (4.70) (0.67) (0.69) –0.0003 –0.0005 –0.0002 –0.0002 –0.0005 – 0.0005 –0.0001 –0.0001 (–6.26) (–11.61) (–3.07) (–3.48) (–3.15) (–3.60) (–1.00) (–0.71) 0.2662 0.2100 0.2424 0.1122 (33.38) (12.23) (8.37) (1.86) 0.0037 0.0062 0.0958 0.0339 (0.35) (0.29) (2.22) (0.40) 0.0403 0.1477 – 0.2045 – 0.0003 (3.48) (6.64) (–1.42) (–0.01) 0.0481 0.0836 – 0.0270 0.0015 (2.99) (2.45) (–0.75) (0.02) – 0.0165 0.2762 – 0.2691 0.1794 (–0.09) (6.34) (–1.96) (1.69) – 0.0121 – 0.0259 0.0079 –0.0293 (–0.85) (–0.62) (0.16) (–0.31) – 0.0620 – 0.0420 – 0.1508 – 0.1471 (–3.78) (–1.33) (–3.97) (–2.01) – 0.1882 – 0.0347 – 0.1755 0.0479 (–12.91) (–1.11) (–3.41) (0.23) – 0.1606 – 0.0535 – 0.0733 – 0.2980 (–12.64) (–1.90) (–1.53) (–2.75) 0.0275 – 0.0445 – 0.0434 – 0.3511 (1.70) (–1.04) (–0.56) (–1.53) – 0.0568 0.0010 – 0.0345 – 0.0233 (–3.40) (0.03) (–0.60) (– .21) – 0.1402 – 0.1151 – 0.1809 – 0.1267 (–11.54) (–4.14) (–3.62) (– .07) 0.1327 0.3503 0.0958 0.2096 0.1301 0.2905 0.0126 0.1094 273.61 197.94 94.19 37.40 27.93 16.47 1.41 3.02 5,464 5,464 2,392 2,392 645 645 259 259

Notes: Data are from a sample of the employed. Log wages are reported in nominal terms and reflect the massive realignment of prices in the transition. T-statistics are in parentheses, Log of monthly earnings is dependent variable

1 percent. Changes in returns to schooling are likely to have a significant impact on the ethnic earnings gap. The magnitude of the effects depend both on differences in educational attainment and quality of education between the ethnic Turks and the ethnic Bulgarians[17]. Not unlike other transitioning economies, the returns to experience in the labor market fell in the early transition for both Bulgarians and Turks. Changes in the returns to human capital favored the ethnic Bulgarians over the ethnic Turks. The extended log earnings function presented in Table VI controls for demographic factors such as marital status and living in the capital city, Sofia, as well as sector of employment (private, state, or cooperative) and one-digit industry codes of employment[18]. There is a debate regarding the appropriateness of

including industry or occupational dummies in earnings regressions. One side argues that earnings differentials may be the result of job discrimination against women/racial minorities/ethnic minorities who want to enter into high-paying industries or occupations[19]. With ethnic Turks being concentrated in low-wage industries or occupations, the inclusion of industry or occupational dummies could lead to an under-estimation of actual wage discrimination against ethnic Turks who have the same schooling and experience as Bulgarians. The possibility of occupational and/or industry segregation should be considered when interpreting the extended earnings regression estimates reported in Table VI. The results presented in Table VI indicate that changes in the pay to workers in particular industries differed by ethnicity. As compared to the benchmark industry, manufacturing, the pay in agriculture, commerce, services, transportation, and communications grew for both the ethnic Bulgarians and the ethnic Turks. The greatest gain among the ethnic Bulgarians was made in transportation. Among the ethnic Turks, the greatest growth occurred in the communications industry where pay grew from 17 percent less than in the benchmark industry to 4.8 percent more. These results indicate that the Turks are benefiting from the changes in industrial demand that favored the commerce industry. A change in the overall level of inequality in an economy affects ethnic earnings differentials. An absolute increase in inequality can penalize workers at the low-end of the distribution even if relative inequality is unchanged. Table VII presents evidence of a widening relative distribution of wages in Bulgaria in the early transition. The ratio of average top-decile to bottom-decile earnings for both 1986 and 1993 are presented as are percentage changes in the ratio by industry. Earnings differences between the very rich and the very poor in Bulgaria grew between 1986 and 1993. The increased inequality grew most among those working in construction, agriculture, and commerce. The widening of the wage distribution indicates that the position of those at the low-end of the earnings distribution in 1986 worsened by 1993. Ratio of top decile to bottom decile 1986 1993 All Manufacturing Construction Agriculture Commerce Services Public administration Transportation Communication

3.2962 3.2464 3.1843 3.4455 2.8711 3.2869 3.2738 3.0778 3.1576

Notes: Data is from a sample of the employed

5.3405 4.9553 5.5345 6.0528 6.0281 5.0059 5.1988 4.8135 5.2140

Shift toward markets

19

Change 2.0443 1.7089 2.3502 2.6073 3.1570 1.7190 1.9250 1.7357 2.0564

Table VII. Changes in the wage structure: ratios of top decile earnings to bottom decile earnings by ethnicity, education, and industry

International Journal of Manpower 23,1 20

In summary, group-specific factors and wage-structure factors both indicate that the ethnic Turks are adapting to the signals of the market economy by obtaining more years of education and labor market experience, and they are entering into the growing industries in the transition. Further, despite evidence that the Turks continue to have fewer years of education and are disproportionately under-represented in the growing industries of commerce, services, and transportation, trends indicate their movement toward these areas. Changes in the returns to education that benefited the Bulgarians to a greater extent, and the increase in overall inequality in the economy that hurt those at the low-end of the wage distribution could explain the small increase in the ethnic earnings gap in Bulgaria’s early transition. IV. Description of the wage decomposition methodology The Juhn et al. (1991) methodology begins by estimating an ethnic Bulgarian wage equation for an ethnic Bulgarian worker i in time t [20]: Yit ¼ Xit
t þ t it þ uit

ð1Þ

where Xit is a vector containing the observable characteristics of an individual ethnic Bulgarian worker and
t gives the coefficients on these characteristics in year t; it is assumed that Eðuit jxit Þ ¼ 0, so that this equation gives mean wages for ethnic Bulgarians with given characteristics. t is the standard deviation of the residual of the ethnic Bulgarian wage function in year t, and it ¼ uit =bt is a ‘‘standardized’’ residual of the ethnic Bulgarian regression (with mean zero and variance one). Changes in t through time reflect changes in within-group inequality. The actual wage differential between ethnic Bulgarians and ethnic Turks is:
Yt ¼ Ybt
Ykt ¼ ðXbt
bt þ bt bt Þ
ðXkt
kt þ kt kt Þ If we assume that ethnic Turks have the same returns to schooling and experience and the same standard deviation of residuals in the wage equation as Bulgarians, but differ in the amount of schooling, experience, and standardized earnings residuals, the earnings differential can then be written as:
Yt ¼ Ybt
Ykt ¼ ðXbt
Xkt Þ
bt þ bt ðbt
kt Þ ¼
Xt
bt þ bt
t ð2Þ where
Xt ¼ ðXbt
Xkt Þ is the difference between ethnic Bulgarians and ethnic Turks in the average of the individual observable characteristics, the term
Xt
t is the predicted gap between ethnic Turks and ethnic Bulgarians in time t.
t ¼ ðbt
kt Þ, is the difference in the average standardized residual for ethnic Bulgarians and ethnic Turks in time t. ft ¼ ðYkt
Xkt
bt Þ=bt where ðYkt
Xkt
bt Þ is the difference between an ethnic Turk’s actual wages and the wages that he would have received had s/he been rewarded for his/her characteristics at the same rate as ethnic Bulgarians.

Using this formulation, wage divergence between ethnic Bulgarians and ethnic Turks between one year, such as year t, and another year, such as year s, can be written as:
Ys

Yt ¼ ð
Xs

Xt Þ
bt þ
Xs ð
bs

bt Þ þ ð
s

t Þbt þ
s ðbs
bt Þ

Shift toward markets

ð3Þ

21 This decomposes the wage convergence into four components of observable and unobservable characteristics. The first term in equation 3 is the portion of the change in the wage gap due to differences in measured characteristics such as years of schooling and experience, or industry, evaluated at fixed ethnic Bulgarian prices, ð
Xs

Xt Þ
bt . The second term measures the amount of the change in the wage differential over time that is attributable to changes in the prices paid to ethnic Bulgarians for those measured characteristics,
Xs ð
bs

bt Þ. The third term denotes changes in ethnic Turk’s relative position in the ethnic Bulgarian residual wage distribution ð
s

t Þbt . The last term reflects differences in residual inequality over time, or the difference in the ‘‘penalty’’ placed on being at a lower position in the wage distribution,
s ðbs
bt Þ[21]. According to Blau and Kahn (1996), the sum of the first and third terms is a reflection of the group-specific factors that contribute to the difference in wages. Group differences in qualifications and group differences in wage rankings at a given level of measured characteristics. The sum of the second and fourth terms is a reflection of the labor market structure. The wagestructure effect measures the impact of cross-time differences in returns to measured and measured characteristics. The sum of the third and fourth terms represents the impact of cross-time differences in the ‘‘unexplained’’ differential in traditional decompositions[22]. V. Results The decomposition of the increase in the ethnic wage differentials between 1986 and 1993 is presented in Table VIII. In order to explain these changes, the differentials are broken down into four separate effects: the observed characteristics effect, the observed price effect, the gap effect, and the unmeasured prices effect. Group-specific factors consist of the ‘‘observed characteristics effect plus’’ the ‘‘gap’’ effect. Wage-structure factors consist of the ‘‘observed price effect’’ plus the ‘‘unmeasured prices effect’’. See Appendix for variable definitions. Group-specific and wage-structure factors. Based on the decomposition of the basic model presented in Table VIII, three of the four components of the Juhn et al. (1991) decomposition are positive, contributing to an increase in the ethnic wage gap[13]. Only the ‘‘observed characteristics effect’’ served to diminish wage differences between ethnic Bulgarians and ethnic Turks. A negative ‘‘observed characteristics effect’’ indicates that changes in the differences between the ethnic Turks and the ethnic Bulgarians in observable

International Journal of Manpower 23,1 22

Table VIII. Analysis of log wages: human capital regression models

Observed characteristics effect (Xs) Years of schooling Experience Experience squared Sex Married Sofia State Private Const Ag Comrce Svcs Pubadmin Trans Communi Observed prices effect Years of schooling Experience Experience squared Sex Married Sofia State Private Const Ag Comrce Svcs Pubadmin Trans Communi Gap effect Unobserved price effect Total change in gap (1993-1986) Sum of group-specific effects Sum of wage-structure effects Sum of unobserved effects

Basic

Full

–0.0393 –0.0101 –0.0600 0.0309

–0.0697 –0.0102 –0.0773 0.0530 –0.0172 –0.0001 0.0029 –0.0038 –0.0007 0.0006 –0.0021 –0.0135 –0.0027 –0.0006 –0.0010 0.0030 0.0909 0.0264 0.0122 –0.0117 0.0054 –0.0002 0.0223 0.0065 0.0121 0.0006 –0.0050 0.0153 0.0053 –0.0017 0.0019 0.0014 –0.0073 0.0092 0.0232 –0.0769 0.1000 0.0020

0.0328 0.0278 0.0097 –0.0048

0.0137 0.0160 0.0232 –0.0256 0.0488 0.0297

Notes: The components of the decomposition are defined as follows: observed characteristics effect = ð
X93

X86 Þ
b86 , observed price effect =
X93 ð
b93

b86 Þ, gap effect ¼ ð
93

86 Þb86 , unobserved price effect ¼
93 ðb93
b86 Þ. Where X is a vector of explanatory variables,
is a vector of estimated coefficients from an ethnic Bulgarian earnings equation,  is a standardized residual,  is the residual standard deviation of ethnic Bulgarian earnings and
denotes the average of the ethnic Bulgarian/ ethnic Turk difference in the variable that follows. The regression includes controls for education, experience and its square, living in the capital city, marital status, sector of the economy, and industry. T-statistics are in parentheses. Data are from a sample of the employed. Log wages are reported in nominal terms and reflect the massive realignment of prices in the transition

characteristics such as the number of years of schooling or the industry in which one group tends to participate diminished earnings differences. In other words, obtaining additional years of schooling, for example, benefited the Turks and improved their relative earnings position. This is not surprising given that the Turks began to close the gap in number of years of schooling between 1986 and 1993 and entirely closed the experience gap. The other three components, the ‘‘observed prices effect’’, the ‘‘gap effect’’ and the ‘‘unobserved prices effect’’ are positive, serving to increase the ethnic earnings gap. These three components indicate that changing returns to education, experience, and industry benefited the ethnic Bulgarians. Further, changes in overall inequality in the economy penalized the ethnic Turks to a greater extent than the ethnic Bulgarians. In all, changes in ‘‘group-specific factors’’ helped the ethnic Turks whereas changes in ‘‘wage-structure factors’’ hurt the Turks. However the wagestructure effect was larger than the group-specific effect such that the gains made by the ethnic Turks in measured characteristics were more than lost by the changes in wage structure and inequality that benefited the Bulgarians. Importantly, little of the differential is explained by what is sometimes referred to as ‘‘discrimination’’ (the gap effect plus the unobserved prices effect). Particularly in the ‘‘extended’’ model which includes industry dummies and other demographic characteristics. Observed characteristics. Looking first at the ‘‘observed characteristics’’ category, Table VIII shows that the impact of measured characteristics is negative for both models, indicating that the ethnic Turks in 1993 have relatively favorable levels of schooling, experience, and industry of employment as compared to 1986. Members of the ethnic Turkish minority experienced a relative gain in observable characteristics contributing to a decline in the ethnic wage differentials over the period by 0.0393 and 0.0697 log points in each model respectively. Disaggregating the ‘‘observed characteristics’’ effect into its component parts in the extended model shows that, as expected, both years of education and experience contributed to a decline in the ethnic earnings differential. Similarly, participation in agriculture, commerce, services, public administration and transportation also diminished the gap. In particular, participation in the growing commerce industry caused the gap to decline by 0.0135 log points (the largest effect among the industries). Observed prices. In contrast to the ‘‘observed characteristics effect’’, the ‘‘observed prices effect’’ increased the ethnic wage differential in each model by 0.0328 and 0.0909 log points respectively. This effect is the largest of the four components of the decomposition in the ‘‘extended model’’. This indicates that rising returns to schooling, sector of the economy, and industry, for example, favored the ethnic Bulgarians over the ethnic Turks in the transition. Focusing on the extended model and breaking down the ‘‘observed prices effect’’, changing returns to schooling exacerbated the ethnic earnings differential in

Shift toward markets

23

International Journal of Manpower 23,1 24

the transition. Changes in the returns to schooling favored the Bulgarians and increased the gap by 29 percent (0.0264/0.0909). Changes in returns to the previously identified ‘‘growth’’ industries (commerce, services, and transportation) also contributed to an increase in the ethnic earnings differential in Bulgaria’s early transition. These three industries alone comprise of 0.0225 log points, translating into over 24 percent of the observed prices effect (0.0225/0.0909). This indicates that rising returns to commerce, services, and transportation favored the ethnic Bulgarians over the ethnic Turks in the transition. One interpretation of the large observed price effect (particularly in the ‘‘extended model’’) on the change in the ethnic wage differential is that this is due to the persistent high level of industry segregation by ethnicity. Blau and Kahn (1997) point out that the observed price effect may reflect discrimination if segregation by industry ‘‘crowds’’ women or members of a particular racial/ ethnic group into certain industries in the economy, thereby decreasing wages (see also Bergmann, 1974). As of 1993, the ethnic Turks continued to be disproportionately represented in agriculture (see Meurs, 1998; Meurs and Giddings, 1999). Thus, although the ethnic Turks are entering the growth industries, they remain disproportionately underrepresented and the gains benefited the Bulgarians to a greater extent. The gap effect. The ‘‘gap’’ effect measures the contributions of each year’s ethnic Turkish placement in the ethnic Bulgarian residual wage distribution to the change in the ethnic wage differential between 1986 and 1993. This effect increased the ethnic pay gap by 0.0137 log points in the ‘‘basic model’’, and decreased it by 0.0073 in the full model. That is, the ethnic Turks moved up relative to the Bulgarian residual wage distribution. Unmeasured prices. The ‘‘unmeasured prices’’ effect served to increase the ethnic wage gap over time in both models by 0.0160 and 0.0092 log points respectively. On net, however, controlling for education, experience, demographic characteristics, sector, and industry, in the full model the ‘‘gap’’ effect and the ‘‘unmeasured prices’’ effect increased the ethnic gap by only 0.0297 log points in the basic model and 0.0020 log points in the extended model. This evidence indicates that what is sometimes considered to be a measure of discrimination can explain little of the increase in the ethnic earnings gap. Group-specific components in the extended model reduced the ethnic earnings gap by 0.0769 log points, while wage-structure factors increased the gap by 0.1000 log points. Assuming that price changes affected the ethnic Bulgarians and the ethnic Turks similarly, rising inequality in the transition reclaimed more than the gains that the ethnic Turks would have made had the price changes not occurred. In other words, although the Turks were attempting to ‘‘swim upstream’’ by closing the education and experience gap and improving their situation in terms of observed characteristics, they were ‘‘swept downstream’’ by the tide of increasing inequality in the economy. Changes in returns to the observed characteristics that favored the Bulgarians

more than counteracted the gains the Turks made over the period. In summary, changing returns to schooling, experience, sector, and industry in the transition favored the ethnic Bulgarians over the ethnic Turks, causing the ethnic wage differential to increase. VI. Conclusions The ethnic wage gap between the Turks and their Bulgarian counterparts increased as the country began its transition from plan to market. Economists find that differences in personal characteristics such as the amount of education, labor market experience, and participation in certain industries (group-specific factors) are usual suspects in causing wage differentials. Additionally, differences in the returns to such characteristics and changes in the overall structure of wages (wage-structure factors) can contribute to earnings differentials. This analysis explored the degree to which groupspecific and wage-structure factors contributed to the increase in the ethnic earnings differential. Descriptive analyses showed the ethnic Turks closed both the gap in the number of years of education obtained as well as the number of years of labor market experience between 1986 and 1993. Furthermore, the ethnic Turks began to enter the growing commerce and transportation industries in the early transition. This evidence indicates that the ethnic Turks, like the ethnic Bulgarians began responding to market signals, obtaining more education and reallocating resources in response to changes in industry demand. These changes do not explain the increase in the ethnic earnings gap, but point instead to an improvement in the Turk’s relative position in the labor market. In addition to these personal characteristics, this analysis explored the effect of changes in the wage structure as related to level of education and industry. In other words, how changes in the remuneration to different levels of education and industry affected the ethnic wage gap. The results indicated that shifts in the wage structure favored the ethnic Bulgarians, and the wage structure changes explain the increase in the earnings gap. The returns to schooling favored the ethnic Bulgarians, increasing during the early transition to a greater degree than for the ethnic Turks. Similarly, the returns to the growing commerce, services, and transportation industries contributed to the increase in the earnings gap, favoring the Bulgarians. In addition to these shifts, an increase in the overall level of inequality in the labor market punished those at the low-end of the wage distribution, the ethnic Turks, exacerbating the existing earnings differential. These results imply that both the Bulgarians and the Turks are responding to market signals in the early transition. Although the gains made by the ethnic Turks in schooling, labor market experience, and shifts toward growing industries were outweighed by changes in the wage structure that favored the Bulgarians. These trends point toward an eventual improvement in the Turks’ relative position in the labor market.

Shift toward markets

25

International Journal of Manpower 23,1 26

Notes 1. See Brainerd (1998), Obgloblin (1999), Orazam and Vodopivec (1995, 2000), Newell and Riley (1996), Hunt (1998), Bird, Schwarze and Wagner (1994), Krueger and Pischke (1995), Glinskaya and Mroz (2000), Liu, Meng and Zhang (2000), Giddings (2000). 2. President Todor Zhivkov was the long-time communist leader of Bulgaria. 3. The government continued to monitor prices of 14 essential commodities, four goods were priced by the state, and a special policy was introduced to control energy prices (Pishev, 1992). At this time approximately 1,050 firms were privatized and about 80 percent of citizens turned their coupons into investment funds (Pohl, Anderson, Claessens and Djankov, 1997, p. 12). The government retained a majority of the shares among larger firms. 4. GDP declined by 16 percent as a result (Enev and Koford, 1997). 5. Poland followed in 1993 at about 16.0 percent and Hungary at about 14.0 percent (Lenkova, 1997, p. 301). 6. This statistic is based on 1992 Census figures. There are other ethnic minorities living in Bulgaria comprising a smaller portion of the population. The next two largest minority groups include ‘‘Pomaks’’ (Muslim Bulgarians) and Roma. Due to data availability, this analysis cannot explore the effects on minority groups other than the ethnic Turks. 7. I will use the terms ‘‘Turk’’ and ‘‘ethnic Turk’’ and ‘‘Bulgarian’’ and ‘‘ethnic Bulgarian’’ interchangeably throughout the paper. 8. Only approximately 100,000 Turks later returned. It is important to note that this exodus may have influenced the relative position of the ethnic Turks in the post-communist Bulgarian economy. If, for example, many Turkish intellectuals, or leaders fled Bulgaria, leaving behind the less-skilled or poorly educated, measured ethnic inequality would increase as a result. 9. For a detailed description of the wage determination and wage policy in Bulgaria during the period 1990-93, see ILO-CEET and EC (1994) and Bankov (1993). 10. The tax on excess wage fund increases between February and June of 1991 was determined by the following scheme: if the increase was < 1 percent above the fund, a 100 percent marginal tax rate was placed on the increment; if between 1 and 2 percent the marginal tax rate was 200 percent, if between 2 and 3 percent the tax was 400 percent and if above 3 percent, an 800 percent marginal tax rate. 11. Between November of 1989 and December of 1991 alone, the country saw four changes in government. 12 There were 22 cases that reported zero earnings in the sample of the employed. 13. The Bulgarian survey is part of a larger multi-country comparative research project including five other Central and East European countries. 14. The results of the 1993 analysis after adjusting for hours as well as a description of how the data were adjusted to account for hours, are available from the author upon request. 15. Despite the fact that using expected values rather than values that reflect the variance around the regression surface overstates the relation between income and its determinants, because so few cases were imputed, little distortion was introduced. Cases that could not be imputed because data were missing for the independent variables in the regression equations were assigned a value of 0 and dropped from the analysis. Results from the analysis using non-imputed values show a slightly larger increase in the ethnic earnings gap and are available from the author upon request. 16. Francine Blau and Lawrence Kahn refer to these factors as ‘‘gender-specific’’. Because this essay focuses on ethnic differences, the term ‘‘group-specific’’ will be used here instead.

17. Note that while the regression can account for differences in the number of years of schooling, it cannot account for differences in the quality of schooling obtained by ethnic Turks and ethnic Bulgarians which may significantly contribute to differences in earnings between the two groups. Differences in quality of schooling fall into the unmeasured category of the regressions and later in the earnings decompositions. 18. Charles Kroncke and Kenneth Smith (1999) included the ability to speak English in a similar regression examining the wage effects of ethnicity in Estonia. They hypothesized that English language ability would command a higher wage premium in the transition due to changes in industrial demand and in new trading partners. While this is an entirely plausible hypothesis for the Bulgarian case, the 1993 Bulgarian SSEE sample would not allow for inclusion of English language ability as an independent variable, as none of the ethnic Turks in the sample spoke English. 19. Andrew Gill (1994) showed that racial differences in access to high-paying occupations played a significant role in racial wage differentials in the USA. 20. Note that this discussion will reference the ethnic wage gap in Bulgaria, but could be easily applied to any gender or racial/ethnic wage differential. Subscripts ‘‘b’’ and ‘‘k’’ refer to the ethnic Bulgarians and the ethnic Turks respectively. 21. It is important to note that although this decomposition breaks down the residual factors into two components, as opposed to the Blinder (1973) or Oaxaca (1973) techniques, the residual components still represent that portion of the difference in wages that cannot be explained by the explanatory variables in the wage equation. In other words, regardless of the fact that the residual is decomposed, it is still a residual, and it is still a measure of our ignorance and dependent on the structure of the regression equation. 22. This method of decomposing wage differentials is subject to the familiar index problem. The year that one chooses as a base may alter the specific values obtained for each of the four components of the decomposition. Regardless of what base year is used, the overall results are robust. Similarly, whether or not one chooses to base the decomposition on an ethnic Turk wage equation versus an ethnic Bulgarian wage equation will also affect the specific results. Following other studies, this formulation bases the decomposition on the ethnic Bulgarian regression because it is expected that differences over time in discrimination against ethnic minorities will affect the ethnic Bulgarian regression coefficients to a lesser extent. 23. Note that a negative number indicates that the factor serves to decrease the ethnic earnings gap, or a relative improvement of the standing of the ethnic Turks. A positive number indicates that the factor serves to increase the ethnic earnings gap, contributing to a relative decline in standing for the ethnic Turks. References Bankov, G. (1993), ‘‘The cabinet new views on the policy concerning the incomes of the population’’, paper presented at the ILO Conference on Restructuring Labor Practices in Bulgarian Industry, 18-20 May, Sofia. Bankov, G. (1994), ‘‘The incomes policy for 1994 and the mechanisms of its implementation’’, Economic Thought, No. 3, pp. 3-25 (in Bulgarian). Beleva, I., Jackman, R. and Nenova-Amar, M. (1995), ‘‘Bulgaria’’, in Commander, S. and Coricelli, F. (Eds), Unemployment, Restructuring, and the Labor Market in Eastern Europe and Russia, Economic Development Institute of the World Bank, Washington, DC. Beleva, I., Bobeva, D., Dilova, S. and Mitchkovski. A. (1993), ‘‘Bulgaria: labour market trends and policies’’, in Fischer, G. and Standing, G. (Eds), Structural Change in Central and Eastern Europe: Labour Market and Social Policy Implications, Center for Co-operation with the Economies in Transition, Organisation for Economic Co-operation and Development.

Shift toward markets

27

International Journal of Manpower 23,1 28

Bergmann, B.R. (1974), ‘‘Occupational segregation, wages and profits when employers discriminate by race or sex’’, Eastern Economic Journal, Vol. 1, pp. 103-10. Bird, E.J., Schwarze, J. and Wagner, G. (1994), ‘‘Wage effects of the move towards free markets in Germany’’, Industrial and Labour Relations Review, Vol. 47 pp. 390-400. Blau, F.D. and Kahn, L.M. (1992), ‘‘The gender earnings gap: learning from international comparisons’’, American Economic Review, Vol. 82 No. 2, pp. 533-8. Blau, F.D. and Kahn, L.M. (1995), ‘‘The gender earnings gap: some international evidence’’, in Freeman, R. and Katz, L. (Eds), Differences and Changes in Wage Structures, University of Chicago Press, Chicago, IL. Blau, F.D. and Kahn, L.M. (1996), ‘‘Wage structure and gender earnings differentials: an international comparison’’, Economica, Vol. 63, pp. S29-S62. Blau, F.D. and Kahn, L.M. (1997), ‘‘Swimming upstream: trends in the gender wage differential in the 1980s’’, Journal of Labor Economics, Vol. 15 No. 1, pp. 1-42. Blau, F.D. and Kahn, L.M. (1999), ‘‘Understanding international differences in the gender pay gap’’, paper presented at the annual meetings of the American Economic Association, New York, NY. Blinder, A. (1973), ‘‘Wage discrimination: reduced form and structural estimates’’, Journal of Human Resources, Vol. 8 No. 4, pp. 436-55. Bobeva, D. (1997), ‘‘Employment policies and programmes in Bulgaria’’, in Godfrey, M. and Richards, P. (Eds), Employment Policies and Programmes in Central and Eastern Europe, International Labour Office, Geneva. Brainerd, E. (1998), ‘‘Winners and losers in Russia’s economic transition’’, American Economic Review, Vol. 88 No. 5, pp. 1094-115. Crampton, R.J. (1997), A Concise History of Bulgaria, Cambridge University Press, London. Creed, G.W. (1998), Domesticating Revolution: From Socialist Reform to Ambivalent Transition in a Bulgarian Village, The Pennsylvania State University Press, University Park, PA. Eminov, A. (1997), Turkish and other Muslim Minorities in Bulgaria, Routledge, New York, NY. Enev, T. and Koford, K. (1997), ‘‘Incomes policies in Bulgaria’’, in Jones, D.C. and Miller, J. (Eds), The Bulgarian Economy: Lessons from Reform During Early Transition, Ashgate, Aldershot, pp. 81-98. Giddings, L.A. (2000), ‘‘Does the shift to markets impose greater hardship on women and minorities? Three essays on gender and ethnicity in Bulgarian labor markets’’, PhD dissertation, American University. Gill, A.M. (1994), ‘‘Incorporating the causes of occupational differences in studies of racial wage differentials’’, The Journal of Human Resources, Vol. 29 No. 1, pp. 20-41. Glinskaya, E. and Mroz, T.A. (2000), ‘‘The gender gap in wages in Russia from 1992 to 1995’’, Journal of Population Economics, Vol. 13, pp. 353-86. Gustafsson, B. and Li, S. (2000), ‘‘Economic transformation and the gender earnings gap in urban China’’, Journal of Population Economics, Vol. 13, pp. 305-29. Hubner, S., Maier, F. and Rudolph, H. (Eds) (1991), Women’s Employment in Central and Eastern Europe: Status and Prospects, ILO/OECD Technical Workshop: Evolving Labour Markets, Social Policy and Industrial Relations in Eastern Europe. Hunt, J. (1998), ‘‘The transition in East Germany: when is a ten point fall in the gender wage gap bad news?’’, National Bureau of Economic Research, Inc., Working Paper Number 6167. Iankova, E.A. (2000), ‘‘Multi-level bargaining during Bulgaria’s return to capitalism’’, Industrial and Labor Relations Review, Vol. 54 No. 1, pp. 115-37. ILO-CEET and European Commission (1994), The Bulgarian Challenge: Reforming Labor Market and Social Policy, Country Objective Review, Budapest.

Institute of Sociology (1986), Town and Village Study Survey Data, Bulgarian Academy of Sciences, Sofia. Jackman, R. and Pages, C. (1993), ‘‘Wage policy and inflation in Eastern Europe’’, paper prepared for a World Bank Conference, 7-8 October. Jones, D.C. (1991), ‘‘The Bulgarian labour market in transition’’, International Labour Review, Vol. 130 No. 2, pp. 231-48. Jones, D.C. (1992), ‘‘The transformation of labor unions in Eastern Europe: the case of Bulgaria’’, Industrial and Labor Relations Review, Vol. 45 No. 3, pp. 452-70. Jones, D.C. and Kato, T. (1997), ‘‘The nature and the determinants of labor market transitions in former communist economies: evidence from Bulgaria’’, Industrial Relations, Vol. 36 No. 2, pp. 229-54. Jones, D.C. and Meurs, M. (1991a), ‘‘Worker participation and workers’ self-management in Bulgaria’’, Comparative Economic Studies, Vol. 33 No. 4, pp. 48-71. Jones, D.C. and Meurs, M. (1991b), ‘‘On entry in socialist economies: evidence from Bulgaria’’, Soviet Studies, Vol. 43 No. 2, pp. 311-28. Jones, D.C. and Miller, J.B. (1997), Early Transition in Bulgaria: Review and Evaluation, Ashgate, Aldershot. Juhn, C., Murphy, K.M. and Pierce, B. (1991), ‘‘Accounting for the slowdown in black-white wage convergence’’, in Kosters, M.H. (Ed.), Workers and Their Wages, American Enterprise Institute Press, Washington, DC. Karapat, K.H. (1990), ‘‘Introduction: Bulgarian way of nation building and the Turkish minority’’, in Kemel H. Karapat (Ed.), The Turks of Bulgaria: The History, Culture and Political Fate of a Minority, The Isis Press, Istanbul. Kroncke, C. and Smith, K. (1999), ‘‘The wage effects of ethnicity in Estonia’’, Economics of Transition, Vol. 7 No. 1. Kroupova, A. (1990), ‘‘Women, employment and earnings in Central and East European countries’’, Tripartite Symposium on Equality of Opportunity and Treatment for Men and Women in Employment in Industrialized Countries, unpublished manuscript, Prague. Krueger, A.B. and Pischke, J.S. (1995), ‘‘A comparative analysis of East and West German labor markets: before and after unification’’, in Freeman, R.B. and Katz, L.F. (Eds), Differences and Changes in Wage Structures, University of Chicago Press, Chicago, IL. Kunev, K. (1992), ‘‘Etniceskite predrasuduci i etnokulturnata situacija’’, in Kunev, K. (Ed.), Etnokulturnata Situacija v Bulgarija, Sofia. Lenkova, C. (1997), ‘‘Bulgarian labor market during early period of transition’’, in Jones, D.C. and Miller, J. (Eds), The Bulgarian Economy: Lessons from Reform During Early Transition, Ashgate, Aldershot, pp. 301-27. Liu, P.W., Meng, X. and Zhang, J. (2000), ‘‘Sectoral gender wage differentials and discrimination in the transitional Chinese economy’’, Journal of Population Economics, Vol. 13, pp. 331-52. Mason, B. (1995), ‘‘Industrial relations in an unstable environment: the case of central and Eastern Europe’’, European Journal of Industrial Relations, Vol. 1 No. 3, pp. 341-67. Meurs, M. (1998), ‘‘Imagined and imagining equality in East Central Europe: gender and ethnic differences in the economic transformation of Bulgaria’’, in Pickles, J. and Smith, A. (Eds), Theorizing Transition: The Political Economy of Post-Communist Transformations, Routledge, London, pp. 330-72. Meurs, M. and Giddings, L. (1999), ‘‘When the margin becomes the core: occupational stratification and the impact of the economic transition in Bulgaria on women and ethnic minorities’’, paper presented at the annual meetings of the International Association for Feminist Economics, Ottawa.

Shift toward markets

29

International Journal of Manpower 23,1 30

Mincer, J. (1974), Schooling, Experience and Earnings, National Bureau for Economic Research, Boston, MA. Newell, A. and Reilly, B. (1996), ‘‘The gender wage gap in Russia: some empirical evidence’’, Labor Economics, Vol. 3. Oaxaca, R. (1973), ‘‘Male-female wage differences in urban labor markets’’, International Economic Review, Vol. 14 No. 3, pp. 693-709. Obgloblin, C.G. (1999), ‘‘The gender earnings differential in the Russian transition economy’’, Industrial and Labor Relations Review, Vol. 52 No. 4, pp. 602-27. Orazem, P.F. and Vodopivec, M. (1995), ‘‘Winners and losers in transition: returns to education, experience and gender in Slovenia’’, World Bank Economic Review, Vol. 9, pp. 201-30. Orazem, P.F. and Vodopivec, M. (2000), ‘‘Male-female differences in labor market outcomes during the early transition to market: the cases of Estonia and Slovenia’’, Journal of Population Economics, Vol. 13, pp. 283-303. Petkov, K. and Thirkell, J. (1991), Labor Relations in Eastern Europe: Organizational Design and Dynamics, Routledge, London. Pishev, O. (1992), ‘‘Bulgaria: political economy’’, paper for the Hoover Institution Conference on Economy, Society and Democracy, May 7-9, Washington, DC. Pohl, G., Anderson, R.E., Claessens, S. and Djankov, S. (1997), ‘‘Privatization and restructuring in Central and Eastern Europe: evidence and policy options’’, World Bank Technical Paper No. 368, The World Bank, Washington, DC. Poulton, H. (1989), Minorities in the Balkans, Minority Rights Group, Report No. 82. Rutkowski, J.J. (1996), ‘‘Labor markets and poverty in Bulgaria: a background paper prepared for the Bulgaria poverty assessment study of the World Bank’’, unpublished manuscript, The World Bank, Washington, DC. Schienstock, G. and Traxler, F. (1994), ‘‘Economic transformation and institutional change: a cross-national study in the conversion of union structures and politics in Eastern Europe’’, International Journal of Comparative Labour Law and Industrial Relations, Winter, pp. 313-38. Slomp, H., van Hoof, J. and Moerel, H. (1996), ‘‘The transformation of industrial relations in some Central and Eastern European countries’’, in Van Ruysseveldt, J. and Visser, J. (Eds) Industrial Relations in Europe: Traditions and Transitions, Sage, London, pp. 337-57. Szelenyi, I. and Treiman, D.J. (1994), ‘‘Social stratification and mobility in Eastern Europe after 1989: general population survey’’ (computer file), Los Angeles, CA. Thirkell, J. and Tseneva, E.A. (1992), ‘‘Bulgarian labour relations in transition: tripartism and collective bargaining’’, International Labour Review, Vol. 131 No. 3, pp. 355-66. Thirkell, J., Scase, R. and Vickerstaff. S. (Eds) (1995), Labour Relations and Political Change in Eastern Europe, UCL Press, London. Tzanov, V. and Vaughan-Whitehead, D. (1997), ‘‘Macroeconomic effects of restrictive wage policy in Bulgaria: empirical evidence for 1991-95’’, in Jones, D.C. and Miller, J. (Eds), The Bulgarian Economy: Lessons from Reform during Early Transition, Ashgate, Aldershot, pp. 99-125. UNDP (1998), National Human Development Report: Bulgaria 1998: The State of Transition and Transition of the State, Sofia, United Nations Development Program. World Bank (1996), From Plan To Market: World Development Report 1996, World Bank, Washington, DC. Appendix.

Shift toward markets Variable

Definition

lny y s priminc primcomp techcomp voccomp gencomp unicomp ex exsq married sofia state private const ag comrce svcs pubadmin trans communi

The natural log of monthly labor earnings in respondent’s primary occupation Average monthly earnings from employment The average number of years of schooling obtained Incomplete primary education Complete primary degree, possibly some secondary, but incomplete Completed technical secondary degree Completed vocational secondary degree Completed general secondary degree Completed university degree Number of years of total labor market experience Years of labor market experience squared Dummy variable for marital status. Married = 1, not married = 0 Dummy variable for living in the capital city, Sofia, 1= yes, 0 = no Dummy variable for employment in the state sector Dummy variable for employment in the private sector Dummy variable for employment in the construction industry Dummy variable for employment in the agriculture industry Dummy variable for employment in the commerce industry Dummy variable for employment in the services industry Dummy variable for employment in the public administration industry Dummy variable for employment in the transportation industry Dummy variable for employment in the communications industry

31

Table AI. Variable definitions

The research register for this journal is available at http://www.emeraldinsight.com/researchregisters

International Journal of Manpower 23,1 32

The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0143-7720.htm

Ownership and employment in Russian industry: 1992-1995 Susan J. Linz Michigan State University and William Davidson Institute, University of Michigan, East Lansing, Michigan, USA Keywords Privatization, Employment, Manufacturing industry, Russia, Divestment Abstract What impact did privatization have on employment in Russian industry? Utilizes data collected from a panel of 6,205 civilian manufacturing firms in the Central, Volga, North Caucasus, Northern and Western Siberian regions of Russia to explore in more detail the relationship between changes in ownership and employment in Russian industry between 1992 and 1995. In particular investigates whether change in ownership structure is relatively more important than industry, region, or the competitive position of the firm in explaining variation in the employment response to changing output conditions during the initial stage of Russia’s transition from plan to market.

Russia began the transition from a planned economy to a market economy in January 1992. To facilitate a speedy transformation to a market economy, mass privatization (narodnaia privatizatsiia) was adopted. Firms employing two hundred or fewer workers could be sold outright; vouchers were distributed free-of-charge to all Russian citizens to be exchanged for shares in larger companies (Bogomolov, 1993). Manufacturing firms designated for privatization[1] were required to register with federal authorities as a joint stock company and submit their plan for the distribution of shares[2]. Mass privatization, ‘‘creating millions of owners rather than owners of millions’’ as President Boris Yeltsin described it (Nelson and Kuzes, 1994), was to be completed by December 1994. Politically, the objective was to give each person a vested interest in the successful emergence of a market-oriented society[3]. Economically, the objective of transferring ownership from state to non-state authorities was not only to improve the operation and performance of individual firms by allowing (or forcing) them to make profits, but also to improve the federal budget situation[4]. Debate continues over the relative success of Russia’s mass privatization program in achieving these objectives (Aslund, 1995a, b; Blasi et al., 1997; Boycko et al., 1995; Bush, 1994; Frydman et al., 1993; Ickes and Gaddy, 1998; Jeffries, 1996; Perevalov et al., 2000; Radygin, 2000).

International Journal of Manpower, Vol. 23 No. 1, 2002, pp. 32-61. # MCB UP Limited, 0143-7720 DOI 10.1108/01437720210421295

Financial support for data collection and entry provided by the William Davidson Institute of the University of Michigan, an All-University Research Initiation Grant from Michigan State University, and a short-term travel grant from the International Research and Exchanges Board, with funds provided by the US Department of State (Title VIII) and the National Endowment for the Humanities. None of these organizations is responsible for the views expressed in this paper. I thank Jeff Biddle, Belton Fleisher, John Giles, Harry Holzer, Robert DeJong, Gary Krueger, Robert Rasche and two anonymous readers for helpful comments as this paper progressed.

This paper addresses the question, what impact did privatization have on employment in Russian industry? Privatized firms, no longer to receive subsidies from central authorities and facing reduced demand during the chaotic environment caused by dismantling the planning bureaucracy, were expected by Russian and Western analysts alike to release a substantial number of workers[5]. Yet, aggregate data provided by the state statistical agency, Goskomstat, indicate very little reduction in employment between 1992 and 1995. Despite new labor codes which no longer required all able-bodied adults to work, and despite the dramatic decline in state ownership – from nearly 90 percent of the industrial sector to less than 20 percent in a time frame not much longer than the gestation period of an elephant – employment remained remarkably stable in all sectors of the Russian economy. The minimal employment response stood in marked contrast to the dramatic decline in industrial output – machine building firms were producing 40 percent of the 1990 level in 1995, light industry output had fallen to 19 percent of the 1990 level. Overall, industrial production in 1995 was reported at 46 percent of the 1990 level (Goskomstat, 1996, p. 249). Similar results – low employment response to large output reductions – are found in other transition economies in the initial stage of their transformation process (Anderson et al., 1997; Basu et al., 1997; Boeri and Keese, 1992; Dobrinksy, 1996; Earle, 1997; Estrin et al., 1995; Jackman, 1994; Jones, 1996; Kajzer, 1995; Lizal and Svejnar, 1997)[6]. Explanations for firms’ failure to adjust employment levels in correspondence to output reductions include both formal and informal institutional and infrastructure changes, or lack thereof (de Melo et al., 1996; de Melo and Gelb, 1996; Heybey and Murrell, 1997; Lavigne, 1995; Linz and Krueger, 1998), as well as deliberate strategies pursued by ‘‘red executives’’ (Commander et al., 1996; Daianu., 1997; Rutkowski, 1996; Linz, 1996; Linz and Krueger, 1996; Standing, 1996; Thornton and Mikheeva, 1996; Thornton, 1997). This paper utilizes data collected from a panel of 6,205 civilian manufacturing firms in the Central, Volga, North Caucasus, Northern and Western Siberian regions of Russia to analyze the impact of privatization on employment. In particular, this paper addresses the question: to what extent was the responsiveness of employment to changes in output between 1992 and 1995 influenced by the change in ownership structure or competitive position of firm at the beginning of Russia’s transition process? Export experience and location in Moscow are used to proxy for the firm’s competitive position[8]. The industry and regional distribution of the 6,205 civilian manufacturing firms in the panel are summarized in Table I. Also reported in Table I are the percentages of firms by industry that had export experience in 1992, as well as the percentage located in Moscow. Table II reports the change in ownership structure of the firms in the panel between 1992 and 1995. These data are utilized to investigate the extent of industry and regional variation in the importance of ownership structure or the competitive position of the firm in

Ownership and employment in Russia 33

International Journal of Manpower 23,1 34

Industry Power Fuel Ferrous/nonferrous metallurgy Machine building Chemicals Wood/forestry/paper Construction materials Light Food Printing Miscellaneousf Total

Exporters Moscow Centrala N. Caucasusb Volgac Northernd W. Siberiae Total % of total % of total 76 64

6 44

33 14

10 0

3 10

133 132

0.0 3.8

12.0 1.5

29 715 99 444 207 650 993 243 45 3,565

1 119 17 40 58 82 281 45 4 697

7 145 32 162 95 116 459 80 3 1,146

3 31 3 141 15 43 145 8 2 401

5 13 7 109 33 34 155 19 3 396

45 1,023 158 896 408 925 2,033 395 57 6,205

19.0 11.2 16.8 4.9 1.0 5.6 0.8 0.5 5.4

11.0 28.0 21.5 3.4 7.6 12.5 4.5 14.2 33.3

Notes: aIncludes Vladimir, Ivanovo, Kostroma, Tver, Yaroslavl’, Bryansk, Kaluga, Orel, Ryanzan, Smolensk, Tula, Moscow (city) and Moscow oblast; bincludes Rostov and Stavropol; cincludes Astrakhan, Samara, Volgograd, Penza and Ulyanovsk; dincludes Arkhangelsk, Vologda and Murmansk; eincludes Novosibirsk and Tomsk; fincludes firms producing an assortment of goods and services: souvenirs and kitchen utensils; leather goods and funeral services, for example

Table I. Sources: Calculations from firm-level data provided in volumes 1-18, BusinessMap 93: Russian Industry Regional distribution of (Moscow: Business Information Agency), and Biznes-Karta 95, select volumes (Moscow: Business firms by industry: 1992 Information Agency)

Industry Power Fuel Ferrous/nonferrous metallurgy Machine building Chemicals Wood/forestry/paper Construction materials Light Food Printing Miscellaneous Total Table II. State ownership by industry: 1992, 1995

1992 % firms state-owned 98.5 99.2 86.7 86.3 87.3 93.9 79.9 80.2 85.0 98.2 75.4 86.9

n 133 132 45 1,015 158 890 408 920 2,028 392 57 6,178

1995 % firms state-owned 9.8 53.8 17.8 30.2 20.2 55.8 21.1 19.2 21.4 87.1 35.1 32.2

n 133 132 45 1,023 158 896 408 925 2,033 395 57 6,205

Sources: Calculations from firm-level data provided in volumes 1-18, BusinessMap 93: Russian Industry (Moscow: Business Information Agency), and Biznes-Karta 95, select volumes (Moscow: Business Information Agency)

explaining the employment response among civilian manufacturing firms during the initial stage of Russia’s transition from plan to market. Documenting output and employment response patterns by Russian firms adds detail not previously available to the transition story. Given rather

remarkable assertions regarding Russia’s success in establishing a market economy (Aslund, 1995a; Djelic and Sachs, 1993, for example), such detail is crucial in mapping out the contours of Russia’s transition experience. In addition, these data provide an opportunity to assess the relative degree of regional self-sufficiency in the Soviet economy (Gregory and Stuart, 1990; Linz, 2000). Were Soviet planners successful in creating an industrial template that was applied to standardize conditions across regions? If so, we would expect to find little regional variation in the initial conditions among firms in this panel. Variation in the results would be expected in regions that more rapidly or pervasively adopted the reform programs[9]. Our main interest, however, is to examine whether firms began to act as profit maximizers in the early stages of the transition process, and, if so, to determine if it was only firms in industries or regions facing significant foreign competition[10]. In this paper, we take workforce size reduction as a crude measure of profit maximizing behavior. The paper is divided into five parts. Part I summarizes changes in ownership, output, and employment in Russian industry using aggregate data published by Goskomstat. Part II specifies the hypotheses to be tested and methodology used to test the hypotheses. Part III describes the panel data used in this analysis. Part IV presents the empirical results. Part V offers concluding remarks. I. Change in ownership, output, and employment in Russian industry The data on change of ownership in Russia’s privatization plan are striking[11]. Goskomstat reports that state ownership (federal and municipal) accounted for only 15.7 percent of Russia’s commercial enterprises in January 1995, compared to over 90 percent before the privatization program began. Joint stock companies accounted for 38.5 percent of the commercial enterprises, partnerships 18.4 percent, private enterprises accounted for 23.4 percent, and cooperatives 2.4 percent. (Goskomstat, 1996, p. 224). The receipts from privatization between 1992 and 1995 totaled 5,479 billion rubles (Goskomstat 1996, p. 235), which was less than 4 percent of the 1992 federal budget[12]. No mention is made of the direct costs incurred to divest the state of property (Barberis et al., 1996; Jeffries, 1996; Lavigne, 1995); nor is much mention made of the magnitude of indirect costs associated with the speed of the divestiture[13]. Change in industrial output It is hard to exaggerate the decline in industrial output associated with Russia’s version of ‘‘shock therapy’’. Goskomstat reports that at the end of 1992, industrial output totaled only 74 percent of the 1990 level (Goskomstat 1996, p. 249). By the end of 1995, industrial output amounted to less than half of the 1990 level. These output declines varied tremendously by industry, however. For example, while in the fuel industry the 1995 level of output was about twothirds of the 1990 level, in the construction materials industry,output in 1995 was only 45 percent of the 1990 level. Within industries, production across

Ownership and employment in Russia 35

International Journal of Manpower 23,1 36

branches of the industry was uneven[14]. Firms in the food processing industry, for example, produced just over half of the 1990 level in 1995; but in the subcaterogy of meat and dairy processing output was less than 40 percent of the 1990 level (Goskomstat, 1996, p. 249). Explanations for the decline in industrial output abound. Ernst et al. (1996), Gaddy (1996), Gregory and Stuart (1998), Jeffries (1996), Lavigne (1995), and many others provide a wealth of information on the explanations underlying the output collapse. On the demand side, state orders fell when the planning bureaucracy was dismantled and defense expenditures diminished. When COMECON, the socialist trading bloc which ultimately became the transition economies of Central and Eastern Europe and the former Soviet Union, demanded payment in hard currency rather than ruble-equivalents, trade among these countries plummeted. On the supply side, the break up of the Soviet Union disrupted the flow of materials – e.g. border controls delayed shipments by Ukrainian firms supplying steel to Russian pipe manufacturers and cotton used in Russian textile firms from Uzbekistan. In addition, rampant inflation and a depreciating currency made business contracts difficult to write and impossible to fulfill[15]. Given all the obstacles to maintaining a functioning economy, it now appears remarkable that Russian managers were able to overcome these obstacles and continue producing, even at the lower post-reform levels. Even more remarkable was their willingness to do so and not receive payment[16]. In contrast to the decline in industrial output, the number of manufacturing firms actually increased in all industries except electric power (Table III), and the percentage increase exceeded 100 percent in all but the fuel and construction materials industries. Among heavy industry firms, the mastodons of the former Soviet economy, this pattern is likely driven in large part by the privatization process. As firms were required to change their ownership structure, they had the option of dividing into multiple units[17]. A puzzle is that, as shown in Table III, industries with the highest percentage increase in the number of firms experienced the highest percentage decrease in profitability[18]. Whether the dis-integration of these large firms resulted in lower profits, or whether profits would have declined even more had the firms not broken into more manageable parts, is a topic for further research. Change in employment In contrast to the collapse of industrial production during Russian privatization, employment remained remarkably stable between 1992 and 1995. Industrial workers accounted for 29.6 percent of the workforce in 1992 and slightly more than 25 percent in 1995 (Goskomstat, 1996, p. 35)[19]. Several explanations may account for this apparent employment stability; including upward bias in employment figures and downward bias in output figures. Regarding an upward bias in employment, firms had an incentive to overreport their workforce size by including ‘‘dead souls’’, individuals not currently working at the company (Standing, 1996), in the paperwork they submitted to

Industry Power Fuel Ferrous/nonferrous metallurgy Machine building Chemicals Wood/forestry/paper Construction materials Light Food Printing Miscellaneous

Mean workforce sizea Capital-labor ratiob Output-labor ratioc Number Number Number of firmsd 1992 1995 of firmse 1992 1995 of firmsf 1992 1995 112 130

870 966 1,542 1,443

79 41

42 898 149 872 393 886 1,955 387 51

870 818 1,436 1,143 1,519 1,371 392 336 409 385 814 692 240 230 127 94 468 339

27 440 83 533 263 519 1,349 185 18

231.6 41.7

14.8 8.3

40 14

19.8 107.1 14.0 166.6 23.0 17.7 10.9 32.2 19.8 22.2 8.3 64.0 14.3 29.4 6.6 9.3 5.0 2.1

20 285 91 180 138 339 745 22 13

Ownership and employment in Russia

0.022 0.048 0.013 0.015 0.013 0.011 0.033 0.008 0.006 0.024 0.030 0.005 0.024

0.011 0.186 0.018 0.026 0.010 0.017 0.028 0.042 0.008

Notes: aCalculated by summing the number of employees in the industry and dividing by the number of firms; bcalculated for each industry by dividing the sum of the reported book value of capital assets by the total number of employees. Capital re-valuations conducted each year by central authorities rather than market forces; ccalculated for each industry by dividing the sum of the total reported value of output by the total number of employees. Output data for 1992 were adjusted to make them comparable to 1995 values; dincludes only those firms reporting workforce size in both 1992 and 1995; eincludes only those firms for which capital-labor ratios could be calculated in both years; fincludes only those firms for which output-labor ratios could be calculated in both years

tax authorities. The excess wage tax in effect between 1992 and 1996 penalized firms for paying wages that were more than six times the federal minimum wage. This practice could be hidden from tax authorities if firms kept employees on the books who were neither working at the firm, nor receiving wages (dead souls). Moreover, firms had an incentive to put workers on unpaid leave rather than release them from the company. Released workers were to be paid an amount equal to their average wage for the three months prior to the date of termination (Clark, 1996; Clarke, 1999; Standing, 1996). Workers on unpaid leave, however, retained their rights to the non-monetary benefits that the company provided: medical clinic, child care facilities, housing, and the like. Thus, even if the likelihood of ‘‘re-employment’’ at their company was near zero, workers on unpaid leave were unlikely to trade off the stream of benefits for a termination payment, especially when they could work elsewhere without reporting their income . . . Downward bias in output figures stems from Russia’s confiscatory tax policy. Prior to 1998, firms were obliged to pay taxes on the value of their output, regardless of whether it ultimately was sold. In a world where tax rates on revenues (not profits) exceeded 35 percent and tax payments were required before sales revenues were forthcoming, there was a strong incentive to under-report output volume and value (World Bank, 1996; EBRD, 1998; Ickes and Gaddy, 1998; Linz, 1997, 1999)[20].

37

Table III. Employment conditions by industry, 1992 and 1995

International Journal of Manpower 23,1 38

Even though measurement bias may lead to an underestimate of reduction in labor – output ratios in Russian firms during the transition, it is likely that employment was maintained at higher levels than pure cost minimization would have implied. One reason for this maintenance of high employment in the presence of ownership change may be the high incidence of employee ‘‘buyouts’’. In more than 84 percent of the cases, privatized firms were taken over by employees (‘‘insiders’’), who were unlikely to adopt practices that would terminate their position in the firm (Earle and Estrin, 1997; Buck et al., 1994, 1996; Frydman et al., 1996; Linz, 1998). Another reason for employment stability is likely to be the Soviet legacy of ‘‘job rights’’ (Granick, 1987). In the Soviet economy, labor codes required virtually all able-bodied adults to work for the state. Labor codes restricted managers’ rights to release workers. Budget allocations covered the wage bill, regardless of a firm’s performance. Moreover, in the Soviet economy, firms were the primary providers of housing, health care, child care, recreation facilities, and numerous goods and services routinely unavailable in normal distribution channels. Managers maintained a larger workforce than would be required to produce the targeted level of output because the cost of employing surplus workers was zero and the benefits were enormous[21]. In addition, Soviet society attached high prestige to managers of large firms (Millar, 1987; Ledeneva, 1998; Standing, 1996). The tradition of the firm’s (manager’s) providing for the well-being of all employees persisted at the beginning of the transition process because: first, social and cultural norms did not change as quickly as the legislation and second, alternative sources for providing wages and benefits were not readily available. Thus managers, steeped in the paternalistic legacy of providing for workers, were reluctant to release workers[22]. II. Hypotheses and methodology The primary objective of this paper is to investigate percentage changes in Russian firms’ employment between 1992, when there was a major change in ownership structure, and 1995. Conceptually, there are two sources of employment change: (1) the initial deviation from optimal (not to say cost-minimizing) employment when the new owners’ objectives are substituted for the state’s objectives; and (2) subsequent changes in output. Regarding the first source, it is hypothesized that firms in general desire fewer workers per unit of output under non-state ownership than they employed at the time of ownership change. The size of this reduction is further hypothesized to depend on firms’ competitive environment after transition. Firms’ ability to adjust will likely vary by industry and region according to differences in institutional and managerial characteristics[23]. Also, firms in regions or industries that are more dependent on exports are expected to more rapidly approach cost-minimizing labor demands, which implies larger initial declines

in employment. Thus I hypothesize that firms with export experience in 1992 and/or firms located in Moscow will have a larger initial employment adjustment than non-exporters located in the provinces. Firms with export experience in 1992 and/or firms located in Moscow were in a stronger competitive position than firms without export experience or firms located in the provinces. To the extent that exporting firms in 1992 had the capital and quality of labor required to compete in global markets, as well as the information and expertise required to complete the transaction, they are also likely to have acted more quickly when the chance came to improve their competitive position at the beginning of the transition process[24]. Firms located in Moscow have consistently had greater access to information and financing (formal and informal) than firms located in the provinces[25]. This paper attempts to assess the relative importance if these influences. The second component of the employment change involves changes in the quantity of labor demanded that stem from the impact of the transition on the demand for the firm’s output. For firms with increases in output, this source of employment increases might, of course, dominate any reduction in labor per unit of output. Firms in the food processing industry, for example, where shortages were particularly severe at the beginning of the transition and expanding production required little if any additional capital investment, are likely candidates for a positive employment response. Whatever the source of output change, I hypothesize that the employment response to a given percentage change in output by state-owned firms will be less than by nonstate-owned firms. The rationale for specifying only two types of ownership structure is twofold. First, despite the fact there were multiple ownership structures in place between 1992 and 1995[26], the allocation of subsidies from central authorities was based on state versus non-state ownership. State-owned firms continued to receive subsidies after 1992; non-state-owned firms were not scheduled to receive subsidies[27]. No longer facing soft budget constraints (Kornai, 1980), it was thought by Russian and Western analysts alike that privatized firms would find it increasingly difficult to sustain employment levels common among state-owned firms in the former Soviet economy[28]. A second rationale for specifying only two types of ownership structures is that, despite the variety of non-state ownership structures permitted in 1992, nearly all non-state-owned manufacturing firms in 1992 and 1995 were employeeowned. Russia’s privatization program in the manufacturing sector simply transferred the firm’s ownership from the state to the employees (see Appendix). As late as the end of 1995, opportunities for outsiders to obtain shares in any given company remained limited: few firms traded shares on the stock market, Russia’s capital market was underdeveloped, laws precluded foreign ownership, and the general lack of information about the financial position of Russian firms precluded outside investors (non-employees) from acquiring shares. Consequently, for the purposes of this analysis, it is

Ownership and employment in Russia 39

International Journal of Manpower 23,1 40

necessary only to consider two ownership structures: state-owned and nonstate-owned. In the regression used to evaluate the relative importance of the several influences on employment change discussed above, the dependent variable is , CHGL, the percentage change in employment between 1992 and 1995. The regressors are as follows: . STATE95, a dummy variable = 1 if a firm was state-owned in 1995,= 0 otherwise. . EXPORT92, a dummy variable = 1 if a firm reported exports in 1992, = 0 otherwise. . MOSCOW, a dummy variable = 1 if firm is located in Moscow, = 0 otherwise. . CHNGQ, the firm’s percentage change in output between 1992 and 1995. . (Q_L92), the firm’s average labor productivity in 1992. . (K_L92) the firm’s capital – labor ratio in 1992. . A dummy variable for each major industry represented in the panel, with machine building the industry dummy omitted from each regresssion estimated[29]. . A dummy variable for each region included in the panel, with the Volga region omitted from each regression estimated[30]. .

Various cross-product terms were included in alternative specifications to capture industry and regional differences in initial responses to privatization and to changing output between 1992 and 1995. The following section describes the panel data used in this analysis. III. Characteristics of firms in the panel Firm-level data published in business directory form by a privately-owned company in Moscow[31] were utilized in this study. The panel was constructed by matching firm-specific registration numbers in 1992 and 1995. The panel was restricted to those civilian manufacturing firms that did not divide into separate or multiple units between 1992 and 1995, as well as to those firms that remained in the same industry in both years[32]. Funding constraints limited this analysis to the Central, North Caucasus, Volga, Northern and Western Siberia regions[33]. The panel includes 2,033 firms in the food processing industry; 1,023 firms in machine building; 925 firms in light industry; and 896 firms in the wood/ forestry/paper industry. Firms in the construction materials industry and printing industry each account for about 6 percent of the total number of firms in the panel. As seen in Table I, more than half of the firms in this panel are located in the Central region, which includes Moscow and its environs, as well as Vladimir, Ivanovo, Kostroma, Tver, Yaroslavl’, Bryansk, Kaluga, Orel, Ryazan, Smolensk, and Tula. Nearly 20 percent of the firms are located in the

Volga region (Volgograd, Astrakhan, Samara, Penza, and Ulyanovsk). Just over 10 percent are located in the North Caucasus region (Rostov and Stavropol). Western Siberia (Novosibirsk and Tomsk) and the Northern region (Arkhangelsk, Vologda, and Murmansk) each account for about 6 percent of the firms. The selection of the 25 locales to include in the panel was based in large part on my previous and ongoing research projects that have involved one or more visits to the majority of these locales. Between 1992 and 1995, as a consequence of the pace of Russia’s mass privatization program, the share of state-owned firms in the panel fell from 87 percent to 32 percent[34]. The decline in state ownership was most pronounced in the power industry: over 98 percent of the firms were state-owned in 1992, less than 10 percent were state-owned in 1995[35]. State ownership continued to account for more than half of the firms in the fuel industry, the wood/forestry/ paper industry, and the printing industry in 1995. What happened to industry employment patterns in the 6,205 civilian manufacturing firms included in this panel? Table IV summarizes by industry in 1992 and 1995, mean workforce size, capital – labor ratios, and output – labor ratios. The employment figure listed in the business directories is the number of employees reported by the firm to local authorities. The firm reports ‘‘fulltime equivalents’’. That is, if a firm employs two part-time workers and each is working one-half the normal work week, that firm reports a single worker. A similar calculation is made when job-share includes more than two workers. In effect, this standardizes for the normal work week, and the increasing use of part-time workers in 1995 as compared to 1992 (Linz, 1997) does not bias the

Industry Power Fuel Ferrous/nonferrous metallurgy Machine building Chemicals Wood/forestry/paper Construction materials Light Food Printing Miscellaneous Total

Ownership and employment in Russia 41

(1) Employmenta (2) Outputb (3) Employmentc (4) Outputd % Number % Number % Number % Number %chgL/ change of firms change of firms change of firms change of firms %chgQ 11.1 – 3.1

112 130

– 14.2 – 28.2 – 11.8 – 18.1 – 13.7 – 18.8 – 2.4 – 9.4 – 38.2 – 12.8

42 898 149 872 393 886 1,955 387 51 5,875

41.5 – 13.5 – – – – – – – – – –

28.4 46.6 37.7 17.2 12.0 84.1 52.8 33.6 80.0 53.2

40 14

7.6 – 2.9

39 14

20 289 92 182 140 343 749 22 14 1,905

– 7.6 – 21.5 – 9.2 – 17.4 – 11.7 – 15.5 1.2 – 19.2 – 26.2 – 8.8

16 264 86 146 95 328 675 21 12 1696

42.6 – 13.5 – – – – – – – – – –

35.5 48.6 38.9 22.4 16.4 89.6 58.5 35.2 93.3 59.4

39 14

0.015 – 0.072

16 264 86 146 95 328 675 21 12 1,696

0.107 0.201 0.154 – 0.073 0.200 0.097 – 0.161 – 0.082 0.365 0.001

Notes: aIncludes all firms reporting employment in both years; bincludes all firms reporting output in both years; cincludes only firms reporting both employment and output in both years; dincludes only firms reporting both employment and output in both years Sources: Calculations from firm-level data provided in volumes 1-18, BusinessMap 93: Russian Industry (Moscow: Business Information Agency), and Biznes-Karta 95, select volumes (Moscow: Business Information Agency)

Table IV. Employment and output changes by industry: 1992-95

International Journal of Manpower 23,1 42

results[36]. Capital stock values were reported by the firm and thus listed in the business directories in current ruble values, but these values between 1992 and 1995 were determined by central authorities rather than by market conditions (Linz, 1999)[37]. Output also was reported in the directories in current ruble values, and has been adjusted for the purposes of this analysis to account for inflation between 1992 and 1995[38]. The capital-labor and output-labor ratios were calculated by dividing the reported capital stock (output) value by the reported employment figure. Percentage changes in employment and output for firms in this panel are reported by industry in Table V. As seen in column 1 of Table V, employment fell in all but the power industry, with firms in the fuel, food processing, and printing industries experiencing a decline of less than 10 percent. Regional variation in employment change was pronounced as seen in Table VI. Output declined most severely in light industry (see column 2), although firms in the machine building industry and food processing industry reported output values in 1995 of only about half of the 1992 value[39]. Only firms in the power industry reported an increase in output between 1992 and 1995 (see column 2, Table V). When the panel is restricted to those firms reporting both employment and output in 1992 and 1995, as seen in columns 3 and 4 of Table V, the same general results emerge: employment increases in the power industry, and declines by the most in machine building, with light industry and wood/forestry/paper not far behind; output expands in the power industry, and declines by the most in light industry, with machine building and food

Industry Power Fuel Ferrous/nonferrous metallurgy Machine building Chemicals Wood/forestry/paper Construction materials Light Food Printing Miscellaneous Total Table V. Winners by industry

(1) Firms reporting employment increases Number % of of firms totala

(3) Firms reporting (2) Firms reporting employment and output increases output increases Number % of Number %chgL/ of firms totalb of firms %chgQ

(4) Firms reporting employment decrease and output increases Number %chgL/ of firms %chgQ

87 41

78 32

27 7

67 50

19 3

0.192 0.067

8 4

14 136 42 195 120 201 795 95 3 1,729

33 15 28 22 30 23 41 24 6 29

5 71 24 54 27 19 108 6 2 350

25 25 26 30 19 6 14 27 14 18

3 25 11 9 13 3 57 1 1 145

0.069 0.769 0.352 0.229 0.378 0.193 0.524 0.202 0.874 0.453

2 46 13 42 14 14 49 5 1 198

– 0.187 – 0.271 – – – – – – – – – –

0.065 0.586 0.205 0.671 0.471 0.494 0.484 0.712 0.281 0.513

Notes: aIncludes only firms reporting employment in both years; bincludes only firms reporting output in both years

processing industry firms reporting values in 1995 of about half of the 1992 values. In the final column of Table V, the responsiveness of employment to changes in output is recorded by industry. Not surprisingly, all industries generate figures significantly less than one. Firms in machine building and construction materials exhibited the most employment response to changing output conditions, with food processing industry firms only slightly lower. The low employment response figures are not dissimilar to figures reported for other transition economies in Central and Eastern Europe (World Bank, 1996). The fact that machine building firms exhibited a relatively high employment response is rather surprising, however, given their legacy. How responsive was employment in the 6,205 civilian manufacturing firms in this panel to output changes between 1992 and 1995? On the surface, these data tell the tale of rather significant employment security in Russia in the first Percentage change

Central region Moscow city Moscow region Vladimir Ivanovo Kostroma Tver Tula Yaroslavl’ Kaluga Orel Ryazan Smolensk Bryansk

Ownership and employment in Russia 43

Number of firms

–23.1 –19.1 –7.6 –15.6 –11.9 –9.2 –13.5 – 12.7 –15.1 –3.3 –9.2 –6.9 –13.2

612 402 176 220 190 374 263 179 183 123 222 243 186

Volga region Volgograd Astrakhan Samara Penza Ulyanovsk

–16.8 –22.5 –6.2 –7.7 0.6

307 122 250 250 187

Northern region Arkhangelsk Vologda Murmansk

–37.1 –18.1 –36.7

138 192 11

North Caucasus Rostov Stavropol

–8.6 7.8

426 264

Western Siberia Novosibirsk Tomsk

–9.6 –17.1

196 159

Table VI. Percentage change in employment, by region, 1992-1995

International Journal of Manpower 23,1 44

stage of the transition process – while output fell on average by more than 50 percent among the firms in this panel (n = 1905), employment reductions averaged just over 12 percent (n = 5875). Restricting the analysis to firms reporting both output and employment figures (n = 1696), these panel data confirm the nearly infinitesimal employment response to output changes (0.001 for all firms in the panel) that is evident in aggregate statistics reported by Goskomstat, as well as findings based on firm-level surveys reported by Blasi et al. (1997), Lehmann et al. (1997), and Linz (1998)[40]. More importantly, however, these data indicate significant industry variation in employment responsiveness. The remainder of this paper summarizes the results of this investigation of the relative importance of ownership change and competitive position of the firm (as measured by export experience and location in Moscow) on the variation in employment response between 1992 and 1995, controlling for the firm’s initial conditions as well as the changing demand (output) conditions. IV. Empirical results Empirical results are reported in Table VII. In the first specification (column 1), a firm’s percent employment change 1992-95 is regressed on its percent employment change: its output-labor ratio in 1992, its capital-labor ratio in 1992, its ownership in 1995, its export experience as of 1992, its location in Moscow, and its major industry designation (with machine building as the omitted industry), and region (with Volga as the omitted region). The estimated constant term supports the assumption that employment in the machine-building industry was the most seriously constrained by prereform planning constraints. Its initial employment decline of 18.3 percent was the largest of all the industries represented in the panel. However, the estimated coefficient of Q:L implies that a 1992 negative deviation of output per worker of 1,000 rubles was associated with a statistically significant initial reduction in employment of 0.17 percent. Similarly, the estimated coefficient of K:L implies that a negative 1,000-ruble deviation of the capital – labor ratio was associated with a statistically significant initial reduction in employment of nearly 0.1[41]. The coefficients on these variables, consistent with an initial adjustment from a condition of ‘‘surplus’’ labor toward cost minimization, bode well for the ultimate success of the transition process. That these results apply to firms in the machine building industry, frequently depicted as a dinosaur with regard to production techniques, as well as to firms in the Volga region, identified as part of the ‘‘Rust Belt’’ of the former Soviet Union, gives some support to Russian and Western analysts and policy makers who claimed success with regard to establishing a market economy by 1995. In what would be considered a socialist stronghold, machine building firms in the Volga region, firms acted in a manner consistent with their capitalist counterparts. The coefficient of CHGQ, percent change in output is highly significant and implies that a 1 percent change in output was associated with a 0.1 percent change in employment over the period 1992-1995. This is a ‘‘small’’ number if

(1) (2) (3) Coefficient t-statistic Coefficient t-statistic Coefficient t-statistic Independent variables Constant –18.297 CHGQa 0.109 Qa_L92 0.166845 K_L92 0.087 State95 6.456 Export92 1.003 Moscow –3.570 Industry variables Power Fuel Ferrous Chemical Wood Constrmt Light Food Print Misc Machine building (omitted dummy)

0.791 4.853 11.514 3.854 12.421 8.398 11.176 17.855 7.574 –2.260

–7.28* 10.84* 5.95* 3.44* 3.19* 0.39 –0.81

–20.302 –7.13* 0.078 3.05* 0.176589 6.19* 0.083 3.22* 6.554 3.23* 0.745 0.30 –2.722 –0.62

0.14 0.63 1.71 1.03 3.58* 2.55* 4.13* 7.47* 0.99 –0.16

3.242 5.421 12.148 6.074 11.536 8.280 10.701 19.891 1.847 –21.266

Interaction variables CHGQ * Industry CHGQPWR CHGQFUEL CHGQFERR CHGQCHEM CHGQWOOD CHGQCONS CHGQLGHT CHGQFOOD CHGQPRNT CHGQMISC CHGNMACHINE Building (omitted) Region variables WSIBERIA NCAUCAS CENTRAL NORTHRN Interaction variables CHGQ*REGION CHGQSIBR CHGQNCAU

–0.000015 –0.000095 –0.000009 0.000042 –0.000017 –0.000014 –0.000007 0.000032 –0.000136 –0.000225

–8.75 7.80 1.93 Dropped

–1.00 3.33* 1.22

–14.737 11.752 3.815 Dropped

0.53 0.69 1.74 1.51 2.56* 2.40* 2.45* 7.26* 0.22 –1.19

–22.922 0.097 –11.720 0.479 6.031 0.325 –5.749

–7.25* 5.54* –0.14 4.43* 3.00* 0.13 –1.29

9.142 22.098 –9.506 9.311 17.376 –7.837 16.587 17.836 –9.731 17.243

1.32 1.78 –0.69 1.68 2.82* –1.06 4.44* 5.47* –0.56 0.33

–17.851 10.141 2.991 Dropped

–1.96** 3.53* 1.45

Ownership and employment in Russia 45

–0.28 –1.24 –0.12 1.07 –0.36 –0.34 –0.22 1.27 –1.63 –1.61

–1.56 3.79* 1.80

–0.000217 –2.44* 0.000055 1.86

–0.218 0.046

–2.56* 1.63 (continued)

Table VII. Employment response regression results, dependent variable percent employment change 1992-95

International Journal of Manpower 23,1 46

(1) (2) (3) Coefficient t-statistic Coefficient t-statistic Coefficient t-statistic CHGQCENT CHGQNRTH (omitted dummy) Interaction variables Q_L92*Industry QLPOWER QLFUEL QLFERR QLCHEM QLWOOD QLCONST QLLIGHT QLFOOD QLPRINT QLMISC Interaction variables K_L92*Industry KLPOWER KLFUEL KLFERR KLCHEM KLWOOD KLCONST KLLIGHT KLFOOD KLPRINT KLMISC

Table VII.

0.000035 Dropped

265.881 552.628 588.587 43.003 120.048 2,539.835 63.397 309.046 2,105.619 1,139.852

Adj R2 = 1463 Observations = 1449

1.70

0.026 Dropped

1.2

0.36 2.17** 1.54 0.41 0.35 3.18* 0.65 3.48* 1.25 1.03

–0.458 –3.97* –0.731 –3.16* 0.629 1.14 –0.261 –1.41 –0.384 –0.76 0.031 0.10 –0.040 –0.17 –0.361 –2.94* 0.556 0.59 –4.762 –0.93 AdjR2 = 1530 Observations = 1449

AdjR2 = 1810 Observations = 1449

Notes: *Significant at 1 per cent; **Significant at 5 per cent

one takes as a benchmark the implications of long-run constant returns to scale. It is important, however, that this estimate be placed in the context of the immense declines in output that occurred during the Russian privatization process. Average decline in output 1992-1995 was 12.6 percent, which is associated with an estimated average decline in employment of about 1.3 percent, in addition to the initial adjustments already described above. How much of the employment response variation is explained by ownership structure and the firm’s competitive position? Holding initial conditions and the change in output constant, a firm’s employment response varied significantly by ownership structure – firms which remained state-owned (STATE95) exhibited a smaller employment reduction than non-state-owned firms (see column 1). That is, among those civilian manufacturing firms that did not subdivide (create multiple firms) during the privatization process, the employment

response to a given change in output, controlling for differences in labor productivity and the initial capital – labor ratio, was greater among privatized and other non-state-owned firms than among state-owned firms. The coefficients on the two variables used to proxy the firm’s competitive position at the beginning of the transition process, EXPORT92 and MOSCOW, were not statistically significant. That the firm’s competitive position, as measured by export experience and location in Moscow, had no significant effect on the firm’s employment response is somewhat surprising if one is expecting, a priori, that these firms would behave like profit-maximizing firms in a developed market economy. The initial expectation of such behavior is shaped by the assumed influence that foreign firms and/or global markets would have had on these Russian firms. It would appear from the results of this analysis, however, that domestic policies and institutions in the first stage of the transition process had a stronger effect on a firm’s employment response than did the firm’s competitive position[42]. That is, there is not enough evidence in these panel data to suggest that firms in Moscow or firms with export experience, despite their greater access to foreign financing, global markets, and hard currency, acted any differently than other firms with regard to their employment response to changing output conditions – the Soviet legacy of job rights (Granick, 1987) sustained employment levels even among firms with (potentially) more experience working in a market environment. How important are industry and region in accounting for the firm’s employment response? The results in column 1 indicate that, during the first stage of Russia’s transition process, there were significant differences by industry and region in the firm’s employment response. That is, holding the change in output and initial conditions constant, firms in the food processing industry (FOOD), light industry (LIGHT), the construction materials industry (CONSTRMT), and the wood/forestry/paper/pulp industry (WOOD) exhibited a significantly smaller employment response than firms in machine building. Similarly, firms in the North Caucasus region (NCAUCAS) exhibited a significantly smaller employment response than firms in the Volga region. Industry variation in the employment response evident in these panel data may be explained by differences in wage payments. That is, despite the fact that real wages fell across all industries between 1992 and 1995, firms in the food processing, light, wood, and construction materials industries had greater access to cash in terms of payment for goods delivered than firms in machine building. Consequently, firms in these industries were more likely to be able to pay workers in a timely manner. At a time when wage arrears were a predominate feature of the economic environment, the ability to pay wages, albeit low in real terms, enabled firms to sustain employment levels despite falling production. Are the industry and regional differences in employment response explained by the differential impact of the transition on output? In Table VII, column 2, regression results are presented for a specification which includes interaction variables that take into account differences in the change in output by industry and region. The three main results obtained in the initial specification remain:

Ownership and employment in Russia 47

International Journal of Manpower 23,1 48

(1) the employment response between 1992 and 1995 among the civilian manufacturing firms in this panel was minuscule, less than 0.1 percent; (2) holding the change in output constant, the employment response was greater among firms with low labor productivity and low capital intensity in 1992; (3) ownership, not competitive position, influenced a firm’s employment response – state-owned firms retained a larger percentage of their workers during the first stage of the transition process than did nonstate-owned firms. The industry interaction variables (CHGQPWR to CHGQMISC) were not significantly different from zero; that is, the differences in employment response by industry were unrelated to industry differences in the percentage change in output between 1992 and 1995. There were, however, significant differences associated with regional differences in output reductions. Holding industry constant, the impact of the transition on output among firms in the North Caucasus region (CHGQNCAU) had a smaller influence on their employment response than among firms in the Volga region – in comparison to firms in the Volga region, firms in Rostov and Stavropol were less likely to release workers and/or workers were less likely to leave firms when output declined. The final specification (column 3) allows for industry differences in the firm’s initial conditions: average labor productivity and the capital-labor ratio in 1992. Including these interaction variables (QLPOWER to QLMISC, and KLPOWER to KLMISC) makes it possible to assess the extent to which a firm’s employment response was influenced by industry differences in the firm’s initial conditions[43]. In this specification, as seen in the results reported in column 3, ownership remains a significant determinant of the firm’s employment response – in percentage terms, state-owned firms released fewer workers than their non-state-owned counterparts; and, as before, neither measure of the firm’s competitive position is significant in explaining the firm’s employment response[44]. Industry differences in initial conditions explain a significant part of the variation in employment response across firms[45]. The results from these panel data indicate that, holding the change in output constant, there is a significant industry difference among firms with high labor productivity and low labor productivity in terms of their employment response – firms in the fuel industry, the construction materials industry, and the food processing industry released fewer workers between 1992 and 1995 than machine building firms with the same labor productivity characteristics. The result that capital-intensive firms in machine building held on to a greater fraction of their workforce than did machine building firms with a low capital-output ratio (K_L92) remains robust. This same result holds for firms in the power industry, the fuel industry, and the food processing industry. However, firms in the fuel industry behaved significantly different

than their machine building counterparts – that is, capital intensive fuel industry firms held on to fewer workers between 1992 and 1995 (the employment response to a given change in output was greater among capital-intensive fuel industry firms than among capital-intensive machine building firms).

Ownership and employment in Russia

V. Conclusions Using a panel of 6,205 civilian manufacturing firms, this paper examines the impact of transition in general and privatization in particular on employment in Russia between 1992 and 1995. The results indicate that the change in ownership was significant in explaining the employment behavior of Russian firms – the employment response by privatized firms to a given change in output was larger than that of firms which remained state-owned over this period. In some respects, this result is surprising. While subsidies from central authorities were not to be paid to privatized firms, many privatized firms, by establishing banks which obtained interest-free loans from the Central Bank, continued to receive indirect subsidies at least until the end of 1995. Consequently, despite their non-state ownership structure, many privatized firms did not face hard budget constraints and thus did not have to act as profit maximizers in order to continue operations (Gregory and Stuart, 1998). While the persistence of ‘‘subsidies’’ could easily have diminished the employment response by firms in industries (consumer goods) or locations (Moscow) subject to foreign competition, this outcome is not strongly evident among the firms in this panel. Among the civilian manufacturing firms that did not divide into multiple units between 1992 and 1995, location in Moscow did not have a significant effect on a firm’s employment behavior. Firms in the food processing industry, and to a somewhat lesser extent, firms in light industry, both of which were immediately affected by foreign competition, exhibited a significantly lower employment response than firms in other industries, especially in comparison to machine building. What are we to conclude from these results? First, this study illuminates the impact that privatization had upon the employment response among Russian manufacturing firms. Despite the fact that privatized firms received vouchers rather than cash during the privatization process which limited their ability to renovate their capital stock and thus to improve the quality of existing products or change the assortment of production, they were significantly more responsive to changing economic conditions than state-owned firms with regard to employment practices. Indeed, despite the fact that ‘‘insiders’’ took over ownership, and that, at least on paper, employment stability in industry remained the norm in the face of dramatic output reductions, the privatized firms in this panel, more so than their state-owned counterparts, adjusted employment in a manner consistent with profit-maximizing behavior. In the process of investigating the importance of ownership structure in explaining employment response among civilian manufacturing firms in Russia, we learned that the firm’s competitive position, as measured by export

49

International Journal of Manpower 23,1 50

experience and location in Moscow, had no significant effect on employment response. One explanation for this result involves the political position of these firms: their ability to extract subsidies from central authorities was no doubt greater than that of firms without hard currency earnings or firms located in the provinces. A second explanation involves the nature of the panel itself. That is, the panel is comprised of firms that did not subdivide into multiple entities between 1992 and 1995. Anecdotal evidence suggests that many firms did subdivide during the privatization process, and that one of the separate new entities created would likely be that division of the original firm that was involved in producing for global markets. Moreover, because Moscow was and continues to be the center for all things related to the successful operation of a business in Russia, it is likely that, all other things equal, firms in Moscow would subdivide in order to capture a particular market niche. The fact that Moscow led the entire country between 1987 and 1992 in the establishment of non-state-owned firms (Linz, 2000) suggests that Muscovites are quick to take advantage of financial opportunities. Consequently, significant employment adjustment could have occurred among firms with export experience in 1992, as well as among firms in Moscow, but given the nature of the panel, would not be picked up by these data. Second, we learned that there was significant variation across industries and regions in the employment response to changing output conditions, but that the employment response differences were not explained by the differential impact of the transition on output across industry. Regional differences in the impact of the transition on output were significant, however. It would be interesting to discover if the employment response varied inversely with the relative magnitude of subsidies granted to the region. Third, these data clearly show that machine building firms, in comparison to firms in other industries, were not automatic losers in Russia’s transition process, despite carrying the burden of a largely obsolete capital stock. Even among firms that did not subdivide into multiple entities, machine building firms exhibited behavior consistent with a cost-minimization strategy: machine building firms with low labor productivity in 1992 released a greater percentage of their workforce than firms with a high labor productivity at the beginning of the transition process; labor intensive machine building firms released a greater percentage of their workforce in response to a given reduction in output than did capital-intensive firms. Interestingly enough, in industries which faced stiffer foreign competition than firms in machine building, food processing, light, and construction materials, for example, the employment response to a given reduction in output was significantly less than for machine building firms. That is, even before the transition began in 1992, traders were bringing to Russia consumer goods and construction materials; goods for which there was an enormous excess demand created by Soviet production and pricing policies. The fact that machine building firms were not laggards with regard to employment response is a

positive sign that the transition from a planned economy to a market economy had gained a strong foothold by 1995. Finally, these firm-level data underscore the importance of understanding institutions in interpreting results emerging from the transition process. The main result that the employment response in the first stage of Russia’s transition process was very small attests to the strength of the Soviet institution of job rights; the commitment by managers and workers alike to maintaining the enterprise despite the elimination of the Soviet labor codes requiring them to do so. As such, these data support the recent revelation by Western policymakers that profit maximizing behavior is not an outcome automatically emerging with privatization. Without doubt, the civilian manufacturing firms in this panel, like all firms in Russia, took advantage of financial opportunities that arose with price liberalization and trade decentralization. They did not, however, automatically abandon the institution of job rights when the labor codes changed. While this point is now taken as rather passe´, during the actual initial stage of Russia’s transition process when privatization was in full swing, the point was either ignored or not understood by policymakers. Only now has hindsight enabled us to approximate 20/20 vision. At the same time, however, these data clearly reveal the Russian managers, even the ‘‘red executives’’, adopted employment policies consistent with a costminimization strategy. That is, firms with low labor productivity in 1992 released a greater fraction of their workforce than firms with high labor productivity; capital-intensive firms retained a greater percentage of their workforce. In comparison to the machine building industry, industries with a narrower production assortment, fuel and food processing, for example, were exceptions. While the overall employment response was small, the fact that it was significantly greater among privatized firms and that it was consistent with the pattern predicted for a profit-maximizing firm, even at a time when budget constraints were not necessarily hard, suggests that the cause of Russia’s ongoing adverse economic conditions lies far outside the realm of employment strategies adopted by manufacturing firms. Notes 1. Goskomstat reports over 61,000 manufacturing firms in Russia in 1992 (Goskomstat, 1996; p. 248). Of these, more than two-thirds were targeted for privatization. Firms excluded from the privatization program were in strategic areas of the economy, such as firms in defense-related production (aviation), pharmaceuticals, and printing, for example (Bush, 1994; Linz, 1994). Expressly targeted for privatization were large and medium-scale ‘‘enterprises whose activity is not efficient’’ (Ekonomika i zhizn’, 1992, p. 5). According to Vasil’ev (1992), vice-chair of Russia’s State Property Management Committee, the 4,500 enterprises slated for ‘‘obligatory privatization’’ had a combined book value equal to the total value of all vouchers distributed in Russia. 2. Within one year, of the 14,500 large-scale firms (employing more than 1,000 workers) targeted for privatization, 11,000 had been converted to joint stock companies. The firm’s privatization plan had to specify one of three share distribution methods, all of which involved trading shares for vouchers. In more than 85 percent of the cases, firms specified

Ownership and employment in Russia 51

International Journal of Manpower 23,1 3.

52 4.

5.

6.

7.

8.

9.

10.

11.

the variant where employees obtained 51 percent of the shares prior to any open auction, with the option of buying additional shares (Ekonomika i zhizn’, 1994, p. 8). In effect, Russia’s privatization program, as applied to the manufacturing sector, simply turned state-owned firms into employee-owned firms without the added benefit of generating additional capital for the firm. Anatoli Chubais, head of Russia’s State Property Management Committee and the overall privatization program, reported in 1993 that the privatization program’s design was ‘‘5 percent economics and 95 percent politics’’ (Blasi et al., 1997; Nelson and Kuzes, 1994). Privatized firms were no longer to receive subsidies from the state, thus state expenditures would be reduced. Moreover, privatized firms making a profit would pay taxes, thus improving the revenue side of the budget. Revenues from the sale of property would also be net additions to the budget, as would the sale of the ‘‘golden shares’’ held by the state (30 percent) in the firms privatized in voucher auctions. Undermining these anticipated gains, head of the State Property Management Committee, Anatoli Chubais reported in 1993 that more than half of all federal property was on the brink of bankruptcy (Kotel’nikova, 1993). Anticipated unemployment estimates ranged from 10 percent to 25 percent of the industrial workforce, based in part on estimates of the magnitude of surplus labor in former state-owned firms, and in part on Poland’s experience with ‘‘shock therapy’’, where unemployment quickly reached double digits. Basu et al. (1997) report, for example, the Czech and Slovak firms registered labor demand – sales elasticities of 0.0 to 0.1 in the initial stage of their transition; Polish and Hungarian firms exhibited labor demand – sales elasticities of 0.3 and 0.6 during the similar stage. These elasticities rose marginally within two years: 0.33 in Slovakia, 0.6 in the Czech Republic, 0.4 in Poland, and 0.65 in Hungary, but remained low in comparison to the expected firm response in a developed market economy. ‘‘Red executives’’ are managers trained in the planned economy, where production not profit (quantity not quality) dominated the reward structure. While the majority of red executives were still in place in 1995 (Linz, 1996), it is not clear that all acted as pilferers rather than paladins (Linz and Krueger, 1996). Firms with export experience in 1992, albeit relatively few in number because of restrictive regulations during the Soviet regime, were positioned at the beginning of the transition period to continue participating in global markets and thus maintain access to hard currency earnings. Firms located in Moscow, the financial, economic, and political hub of Russia, had first access to information and financial resources, and also were first to face competition from foreign firms and products. Resistence in Russia to the reform programs was widely reported in the popular press at the time, with resistence particularly strong among directors of large industrial enterprises (Cline, 1993; Jeffries, 1996; Lavigne, 1995; Rose, 1993), who ultimately formed an organization (the Russian Union of Industrialists and Entrepreneurs) to strengthen their protests. However, certain regions were targeted as reform zones – Moscow and Nizhny Novogorod, for example. Among civilian manufacturing firms foreign competition was most significant between 1992 and 1995 in consumer goods: firms in the food processing industry and light industry, for example. Firms producing electronic devices (TVs, VCRs, computers) were also at risk to foreign competition, but their production assortment would typically place them in the strategic or defense sector of the economy and thus beyond the scope of the data available for this project. In terms of location, firms in Moscow, the financial, communication, transportation, and political hub of the RSFSR, faced foreign competition sooner than firms located in the provinces. See Appendix for a detailed description of Russia’s privatization program.

12. Unpaid taxes by privatized firms annually totaled more than 25 percent of the 1992 federal budget. See, for example, Aukutsionek (1997, 1998), Litisian (1997), Markarov and Kleiner (1996, 2000), Yakovlev (2000), and Vavilov and Kovalishin (2000). 13. Chubais (1994) expressed concerns about future costs incurred as privatized firms go bankrupt. At the time, estimates put over 80 percent of the civilian manufacturing firms in the category of ‘‘potential candidates for bankruptcy’’; more than half were identified as contributing negative value-added. No one mentioned the wide array of potential costs associated with Russia’s privatization program that was essentially designed to result in insiders taking over. 14. In terms of capacity utilization, the differences across industries and within industries are even more dramatic. Among metallurgical firms, for example, average annual capacity utilization was reported as not exceeding 65 percent in 1995; firms in machine building averaged less than 45 percent. In light industry, average annual capacity utilization in 1995 fell below 30 percent. Within the food industry, capacity utilization was under 20 percent for fruit and vegetable processing, but over 86 percent in sugar processing. See Goskomstat (1996, pp. 258-261). 15. That contract law was not on the books at the beginning of the transition, and no appropriate legal institution to enforce the law in place at the time, also contributed to the obstacles privatized firms faced when trying to conduct business in Russia. 16. For detailed discussion, see Ickes and Ryterman (1993), Linz (1997), Linz and Krueger (1996). 17. For firms in heavy industry, where the emphasis in the Soviet economy had been on selfsupply, the production assortment was very diverse. Moreover, heavy industry firms frequently included consumer goods in their production assortment. It was not unusual for several distinctly different companies to be created during the privatization process from a single state-owned firm. 18. Goskomstat (1996, pp. 262-271) reports the profitability of industrial firms in 1992 and 1995 without actually describing the calculations used to generate the results. My own interviews between 1992 and 1994 with dozens of Russian managers suggest that profits, profitability, profits tax and other related phrases did not correspond directly to similar phrases in conventional (Western) economic or accounting usage. 19. A number of anomalies appear in the official statistics, however, which suggest that either employment stability was not the norm or that little effort was made within Goskomstat to utilize consistent definitions. For example, the size of the industrial workforce declined by 19 percent at a time when, overall, the size of the workforce fell by less than 7 percent (p. 35); yet the share of industrial workers in the overall workforce fell by only 4 percent. While the economy-wide reduction between 1992 and 1995 is consistent with the decline in the number of women in the workforce (p. 32), confusing the analysis is a growth in the industrial sector of the number of part-time workers and workers on administrative leave, reaching 10.9 percent and 12.6 percent, respectively, of the average number of payroll workers in 1995 (p. 45). In Goskomstat documentation, and in conversations by the author with Russian enterprise directors, part-time refers to job-sharing which counts as a single position/person, regardless of the number of people involved. Part-time also refers to reduced hours for a person who previously worked full-time; that is, the position went from full to part-time status in the company. Workers on administrative leave are counted as part of the overall workforce (Standing 1996), but it is not clear how they are counted by firms tallying their workforce size or monthly wage bill. Further muddying the waters are Goskomstat (1996) figures on ‘‘new entrants’’ and ‘‘leavers’’ in industry (p. 43). According to the figures provided, the net outflow (job leavers minus new entrants) of industrial workers in 1994 alone accounted for 12 percent of the industrial workforce in that year, followed by a net outflow in 1995 equal to 6.3 percent of the industrial workforce in that year.

Ownership and employment in Russia 53

International Journal of Manpower 23,1 54

20. Falsifying production volume documents submitted to authorities was not a new experience for Russian managers. The practice of misrepresenting output figures is standard lore for the former Soviet economy (Gregory and Stuart, 1990) – enterprise directors routinely over-reported their production figures (Freris 1984; Linz and Martin, 1982; Linz, 1988). While the direction is no longer the same, the regular misrepresentation of output figures appears to be standard operating procedure for ‘‘red executives’’ in Russia’s transition economy, as well as the for the ‘‘new Russians’’, owners of newlycreated firms (Blasi et al., 1997, Ernst et al., 1996, Lavigne, 1995). 21. ‘‘Surplus’’ workers were mobilized to fulfill plan targets, as well as to engage in both formal and informal (legal and illegal) production activities not specified in the plan. ‘‘Surplus’’ workers were required to provide the social benefits which the firm provided. Plan fulfillment and above-plan production generated bonus payments and other financial rewards. For discussion of overfull employment, see Berliner (1957), Granick (1987), Linz and Martin (1982), for example. 22. The author’s interviews with dozens of Russian managers each year between 1992 and 1995 underscore this reluctance to release workers, the exception being female workers, who were viewed as less productive because domestic responsibilities reduced their actual work time. 23. Discussions with Russian analysts and other specialists on business conditions in Russia indicate that, in their view, 75-100 percent of a firm’s adjustment or restructuring practices are defined by managerial characteristics. My own results, based on survey data collected from 120 firms in Taganrog, Russia, suggest that managerial characteristics account for at least half of the success/failure in adopting effective restructuring strategies (Linz, 2001). 24. The vast majority of firms with export experience in 1992 were state-owned enterprises designated as part of the traditional Soviet planned economy to generate export revenues which would be used by central authorities to finance imports. In this environment, where the state had a monopoly over foreign trade activities, neither production cost nor comparative advantage or efficiency were key considerations in the designation of the firm to export. Moreover, in producing to fulfill quantity targets rather than consumer demand targets, Soviet state-owned firms tended to employ technology well below state-of-the-art. For further discussion, see Gregory and Stuart (1990). 25. For discussion of the capital city effect, see Linz (2000) and Linz and Krueger (1996, 1998). 26. The ownership types included: joint stock company, leased firm, cooperative, worker collective, private/partnership, and joint venture. The exceptions to the employee ownership rule are joint ventures and private/partnership ownership structures. Joint ventures and private/partnerships accounted for less than 10 percent of the civilian manufacturing firms in Russia in 1995, however. 27. Implicit in this decision rule by central authorities was the assumption that non-stateowned firms would be more likely than state-owned firms to act as profit-maximizers. That is, both Russian and Western analysts thought that under these conditions non-stateowned firms would be obliged to sell their products at a price higher than cost, and thus feel pressure to find ways to undertake more efficient production; in short, act as profit maximizers. Moreover, since there was no effective legal or tax differentiation among the different types of non-state-owned firms with regard to profit maximization, they are treated as a single group for the purposes of this analysis. 28. That Soviet firms employed ‘‘surplus’’ labor is well-documented in the literature – most recently by Clarke (1999) – although the actual magnitude is subject to debate. 29. Machine building has traditionally been regarded as the foundation and bellwether of the former Soviet economy, accounting for a significant portion of all industrial employees as well as a significant share of all investment allocations. Because machine building firms

30.

31.

32.

33.

34.

35. 36.

are located in virtually all regions of the Russian economy, and reflect both the best and worst of the former Soviet economy, they are used here as the comparison group. The Volga region comprises the second largest group of firms in the panel. It includes firms in all industries, but does not include firms in Moscow (located in Central region). The Volga region is to Russia what the Midwest is to the United States. The firm, Business Information Agency, obtained the data contained in their 18-volume directories from Goskomstat, the Statistical Agency of the Russian Federation. According to the company, the business directories were sold in Dom Knigi, a large bookstore in central Moscow, and in the Business Center of the Slavyanskaya (Radisson) Hotel, as well as at the company’s headquarters. Despite the limited distribution, the objective of the business directories was to facilitate the flow of information to foreign and Russian businesses about the location and basic characteristics of civilian manufacturing firms in Russia. A total of 21,582 firms were included in the 1992 data set, and 12,521 firms in the 1995 data set. Of these, in less than 1 percent of the cases did multiple firms in 1992 report the same registration number; in about 15 percent of the cases, the registration number for a single firm in 1992 was matched with multiple firms in 1995. In about 10 percent of the cases, the industry code changed between 1992 and 1995. For some firms, a logical explanation may underlie the industry change; firms switching from tractor components to wheelchair production (that is, the main product assortment changes in such a way as to require industry reclassification). In other instances, it may simply be a change or inconsistency in the coding: in Rostov alone, 18 firms coded in the power industry in 1992 were listed in the fuel industry in 1995. In Volgograd, three firms in the power industry in 1992 were listed in the machine building industry in 1995. In some instances, the change in industry affiliation appear to be a mistake: in Moscow, one firm was listed in the power industry in 1992 and the food processing industry in 1995. Thus, to simplify the analysis and clarify the results, only those firms appearing in both listings in the same industry are included in this study. The data from the directories was coded and entered into a database by several research assistants, with funding provided by an All-University Research Initiation grant from Michigan State University, and a research grant from the William Davidson Institute of the University of Michigan. I thank Kathleen Beegle, Janet Blake, Stephen Glenister, Elizabeth Harkness, Sarah Linz, and Natalia Smirnova for their yeoman service. Goskomstat figures include both manufacturing and retail organizations, thus a stateownership percentage of 16 percent for all commercial enterprises (Goskomstat, 1996, p. 244) in 1995 and a state-ownership percentage of 32 percent for civilian manufacturing firms in 1995 is not necessarily inconsistent. Where joint stock companies accounted for only 2 percent of the panel in 1992, they accounted for more than 50 percent in 1995. The share of leased and cooperatively-owned firms fell substantially: from 5 percent to less than 1 percent, and from 4 percent to less than 0.5 percent, respectively, as a consequence of the new ownership laws. The privatization legislation, for example, stated that leased companies could be redeemed by their management and workers for a predetermined price, without auction or competition. If the option of buying the leased company was not exercised, leased companies could be transformed into joint stock companies, where employees could buy as many shares as they could afford, selling the remaining shares. That Russia’s voucher privatization program promoted employee buyout is welldocumented in the literature, as is the fact that firms received no cash for shares (Blasi et al., 1997, Frydman et al., 1994; Nelson and Kuzes, 1994). Management turnover among privatized firms between 1992 and 1995 was rare (Linz, 1996) Non-state ownership in 1992 might also be considered a proxy for a firms’s competitive position at the beginning of the transition process. The reported figures give no clear indication of the number of employees actually working, and/or being paid, however. Standing (1996) estimates at least one-in-five employees in

Ownership and employment in Russia 55

International Journal of Manpower 23,1

37.

56 38.

39.

40.

41

42.

43.

industry are on unpaid leave; Lehmann et al. (1997) put this figure at 10 percent. Both agree that more than half of the employees in industry routinely experienced wage delays. The 1995 figures are likely to be biased upward as a consequence of employment practices that leave workers ‘‘on the books’’ even if they no longer show up for work or receive wage payments. Gregory (1998) notes that, given the high percentage of obsolete capital in each firm, if capital stock values were determined by market forces, the sum total of the fixed capital in Russia in 1992 or 1995 would be very small, indeed. A price index of 171.66 for ‘‘all industry’’, calculated from data provided in Russia in Figures (Goskomstat, 1996), was used to adjust 1992 output figures to comparable 1995 prices. This index does not include the percentage increase in prices in 1992 in comparison to 1991, because it is believed that the 1992 output volume figures reported by the firms in this panel are already adjusted for the 1992 price level. For comparative purposes, a second source of price changes was examined. Figures published in Tsenyi v Rossii (Goskomstat, 1996, pp. 152-152), allowed for calculation of the percentage change in price between 1992 and 1995 by industry: power, 163 percent; fuel, 138 percent; ferrous metallurgy (black), 159 percent; chemicals 151 percent; machine building, 152 percent; forestry/wood products, 153 percent; construction materials, 155 percent; light industry, 141 percent; food processing industry, 148 percent; and ‘‘all industry’’, 151 percent. No explanation is offered in either publication (e.g. coverage, inclusive dates) that would justify the difference in the ‘‘all industry’’ price index. The business directories do not provide explicit information about the time frame covered by the output figure, or any other figure (employment, depreciation, value of capital stock, for example). The 1992 directory appears to refer to year-end results. The 1995 directory refers to mid-year data, and is taken to mean the results associated with mid-1994 to mid1995. Thus, in both cases, it is assumed the information refers to annual figures. Because these employment data, whether reported by Goskomstat or by independent researchers, were collected when the excess wage tax was in effect, providing managers with an incentive to ‘‘over-report’’ the size of their workforce, they must be interpreted with caution. This result may be linked to the nature of the panel, which includes firms that did not subdivide or create multiple units. Machine building firms in the Soviet economy placed a high priority on the ability to self-supply necessary components in the production process. Moreover, Soviet planning practices place consumer goods in the production assortment of machine building firms. Thus it is easy to imagine a capital-intensive machine building firm retaining a relatively large fraction of its workforce, but reassigning production priority to lines that in Soviet times received little priority (consumer goods, components). My own interviews with managers of machine building firms indicate that they shifted production priorities within the first year – as much as doubling the production of consumer goods and the production of spare parts. While it is clear that exporting firms were likely to undergo privatization and thus in that regard forego their claim on subsidies, little is known about the extent to which exporting firms as a group were more likely to receive subsidies than firms that did not have export experience in 1992. That is, among the firms in this panel, 236 of the 260 firms with export experience in 1992 were state-owned in 1992; by 1995, only 59 of these firm remained stateowned. Their access to hard currency earnings may have positioned them to acquire special assistance from central or local authorities (officially or unofficially), or, alternatively, may have required them to retain on the books an even greater number of ‘‘dead souls’’ (Standing 1996) in order to avoid the confiscation of the hard currency earnings. Since results from the second regression indicate that the differences in employment response by industry were unrelated to industry differences in the percentage change in

44.

45. 46. 47. 48. 49.

50.

output between 1992 and 1995, the industry-output interaction variables were not included in the third specification. When a third proxy for the competitive position of the firm, non-state ownership in 1992 (nonst92 = 1 if the firm was not state-owned in 1992) was included, the coefficient on the variable was not significant; nor did any of the other coefficients change significantly. An F-test conducted on each set of the coefficients requires the we reject the hypothesis that the coefficients are equal to zero. Zakony RSFSR. ‘‘O privatizatsii gosudarstvennykh i muinitsipal’nykh predpriiatii v RSFSR,’’ (Moscow: Sovietskaia Rossiia), pp. 3-36. A price ceiling was imposed whereby the value of these shares could not exceed six times the minimum wage paid at the company. In this instance, the nominal price was not allowed to exceed 20 times the minimum wage. Shares may be purchased using vouchers or cash. See ‘‘Basic provisions of the program of privatization of state and municipal enterprises in the Russian Federation for 1992’’, Ekonomicheskaya gazeta, No. 2 (January 1992), ‘‘The State Program of Privatization’’, Izvestiya (27 June 1992), and ‘‘State Program for the Privatization of State and Municipal Enterprises,’’ Rossiiskaya gazeta (9 July) 1992. Vouchers, or privatization coupons, which had a nominal value of 10,000 rubles were made available to all Russian citizens in August 1992, with the instruction that all vouchers must be ‘‘spent’’ by December 1994. For detailed descriptions of the voucher scheme, see Bush (1994), Frydman et al. (1993), Djelic (1992), Djelic and Tsukanova (1993), Linz (1994), and Nelson and Kuzes (1994).

References Anderson, J. et al. (1997), ‘‘Which enterprises (believe they) have soft budgets after mass privatization? Evidence from Mongolia’’, Davidson Institute Working Paper No. 83, University of Michigan, East Lansing, MI, October. Aslund, A. (1995a), How Russia Became a Market Economy, Brookings Institution, Washington DC. Aslund, A. (1995b), Russian Economic Reform at Risk, St Martin’s Press, New York, NY. Aukutsionek, S. (1998), ‘‘Barter v Rossiiskoi promyshlennosti’’, Voprosy ekonomiki, No. 2. Aukutsionek, S. (1998), ‘‘Industrial barter in Russia’’, Communist Economies and Economic Transformation, Vol. 10 No. 2, pp. 179-88. Barberis, N. et al. (1996)., ‘‘How does privatization work? Evidence from Russian shops’’, Journal of Political Economy, Vol. 104 No. 4, Winter, pp. 764-90. Basu, S. et al. (1997), ‘‘Employment and wage behavior of enterprises in transitional economies’’, paper presented at William Davidson Institute conference on Labor Markets in Transition, University of Michigan, East Lansing, MI, October. Berliner, J.S. (1957), Factory and Manager in the USSR, Harvard University Press, Cambridge, MA. Blasi, J. et al. (1997), Kremlin Capitalism, Cornell University Press, Ithaca, NY. Boeri, T. and Keese, M. (1992), ‘‘Labour markets and the transition in Central and Eastern Europe’’, OECD Economic Studies, Vol. 10 No. 18, Spring, pp. 133-63. Bogomolov, O. (1993), ‘‘Razdaetsia nicheninoe bogatstvo’’, Nezavisimaia gazeta, No. 13, 23 January. Boycko, M. et al. (1995), Privatizing Russia, Brookings Institution, Washington, DC. Buck, T. et al. (1994), ‘‘Employee buyouts and the transformation of Russian industry’’, Comparative Economic Studies, Vol. 36 No. 2, Summer, pp. 1-15.

Ownership and employment in Russia 57

International Journal of Manpower 23,1 58

Buck, T. et al. (1996), ‘‘The process and impact of privatization in Russia and Ukraine’’, Comparative Economic Studies, Vol. 38 No. 2/3, Summer-Fall, pp. 45-69. Bush, K. (1994) ‘‘Report on privatization’’, RFE/RL Research Report, No. 67, April. Chubais, A. (1994), ‘‘Privatizing state property’’, Moscow News, No. 3, March, p. 10. Clark, C. (1996), ‘‘The transition of labor relations in Russian industry: the influence of regional factors in the iron and steel industry’’, Post-Soviet Geography and Economics, Vol. 37 No. 2, February, pp. 88-112. Clarke, S. (1996), The Russian Enterprise in Transition, Edward Elgar, Brookfield, MA. Clarke, S. (1999), The Formation of a Labour Market in Russia, Edward Elgar, Northampton, MA. Cline, M. (1993), ‘‘Attitudes toward economic reform in Russia’’, RFE/RL Research Report, Vol. 2, May, pp. 43-9. Commander, S. et al. (eds.) (1996), Enterprise Restructuring and Economic Policy in Russia, World Bank, Washington, DC. Daianu, D. (1997), ‘‘Structure and strain in explaining inter-enterprise arrears’’, William Davidson Institute, Working Paper No. 97, University of Michigan, East Lansing, MI, November. de Melo, M. et al. (1996), From Plan to Market: Patterns of Transition, Policy Research Working Paper # 1564, World Bank, Washington, DC. de Melo, M. and Gelb, A. (1996), ‘‘A comparative analysis of twenty-eight transition economies in Europe and Asia’’, Post-Soviet Geography and Economics, Vol. 37 No. 5, May, pp. 265-85. Djelic, B. (1992), ‘‘Mass privatization in Russia: the role of vouchers’’, RFE/RL Research Report, Vol. 1, October, pp. 40-4. Djelic, B. and Sachs, J. (1993), ‘‘Russia: breakthrough year in 1993’’, in Lord, R. (Ed.), Privatization Yearbook 1993, Privatization International, London, pp, 84-5. Djelic, B. and Tsukanova, N. (1993), ‘‘Voucher auctions: a crucial step toward privatization’’, RFE/RL Research Report, Vol. 2, July, pp. 10-18. Dobrinksy, R. (1996), ‘‘Enterprise restructuring and adjustment in the transition to market economy: lessons from the experience of Central and Eastern Europe’’, Economics of Transition, Vol. 4 No. 2, October, pp. 389-410. Earle, J.S. (1997), ‘‘Industrial decline and labor reallocation in Romania’’, paper presented at conference on Labor Markets in Transition Economies, William Davidson Institute, University of Michigan, East Lansing, MI, October. Earle, J.S. and Estrin, S. (1997), ‘‘Privatization versus competition: changing enterprise behavior in Russia’’, World Bank, Washington, DC. Ekonomika i zhizn’ (1992), ‘‘Gosudarstvennaia programma privatizatsii gosudarstvennykh I munitsipal’nykh predpriiatii v Rossiiskoi Federatsii na 1992 godu’’, No. 25, July. Ekonomika i zhizn’ (1994), ‘‘O sotsial’no-ekonomicheskom polozhenii Rossii v 1993 godu’’, No. 6, 8 February Ernst, M. et al. (1996), Transforming the Core, Indiana University Press, Bloomington, IN. Estrin, S. et al. (1995) ‘‘Shocks and adjustment in firms in transition: a comparative study’’, Journal of Comparative Economics, Vol. 21 No. 2, October, pp. 131-53. Freris, A. (1984), The Soviet Industrial Enterprise, St Martin’s Press, New York, NY. Frydman, R. et al. (Eds) (1993), The Privatization Process in Russia, Ukraine and the Baltic States, Central European University Press, London. Frydman, R. et al. (Eds) (1996), Corporate Governance in Central Europe and Russia: Insiders and the State, Vol. 2, Oxford University Press, New York, NY.

Gaddy, C. (1996), The Price of the Past: Russia’s Struggle with the Legacy of a Militarized Economy, Brookings, Washington, DC. Goskomstat (1996), Russia in Figures, Moscow. Granick, D. (1987), Job Rights in the USSR: Their Consequences, New York University Press, New York, NY. Gregory, P. and Stuart, R. (1990), Soviet Economic Structure and Performance, HarperCollins, New York, NY. Gregory, P.R. and Stuart, R.C. (1998), Russian and Soviet Performance and Structure, 6th ed., Addison-Wesley, New York, NY. Gregory, P. and Stuart, R. (1998), Russian and Soviet Economic Performance, Addison Wesley, New York, NY. Heybey, B. and Murrell, P. (1997), ‘‘The relationship between economic growth and the speed of liberalization during transition’’, Department of Economics and IRIS Center Working Paper, University of Maryland-College Park, MD, November. Ickes, B. and Gaddy, C. (1998), ‘‘Russia’s virtual economy’’, Foreign Affairs, Vol. 77 No. 5. Ickes, B. and Ryterman, R. (1993), ‘‘Roadblock to economic reform: inter-enterprise debt and the transition to markets’’, Post-Soviet Affairs, Vol. 9, July-September, pp. 231-52. Izvestiya (1994), ‘‘Privatization and public opinion’’, 2 July. Jackman, R. (1994), ‘‘Economic policy and employment in the transition economies of Central and Eastern Europe: what have we learned?’’, International Labour Review, Vol. 133 No. 3, pp. 327-45. Jeffries, I. (1996), A Guide to Economies in Transition, Routledge, New York, NY. Jones, D.C. (1996), ‘‘The nature and effects of employee ownership and control: evidence from the Baltics, Russia and Bulgaria’’, paper presented at ASSA meetings, San Francisco, CA, January. Kajzer, A. (1995), ‘‘The real-wage – employment relationship and unemployment in transition economies: the case of Slovenia and Hungary’’, Eastern European Economics, Vol. 33 No. 4, July-August, pp. 55-78. Kornai, J. (1980), The Economics of Shortage, North Holland, New York, NY. Kotel’nikova, E. (1993), ‘‘Viktor Chernomyrdin sdelal stavku na Goskomimushchestvo’’, Kommersant Daily, No. 237, 9 December, p. 3. Kucherenko, V. (1992), ‘‘Massovaia privatizatsiia’’, Megapolis-Express, No. 34, 26 August. Lavigne, M. (1995), The Economics of Transition, St Martin’s Press, New York, NY. Ledeneva, A.V. (1998), Russia’s Economy of Favours: Blat, Networking and Informal Exchange, Cambridge University Press, New York, NY. Lehmann, H. et al. (1997), ‘‘Grime and punishment: employment, wages and wage arrears in the Russian Federation’’, paper presented at the William Davidson Institute Conference on Labor Markets in Transition, University of Michigan, East Lansing, MI, October. Linz, S.J. (1988), ‘‘Management’s response to tautness in Soviet planning: evidence from the Soviet interview project’’, Comparative Economic Studies, Vol. 30 No. 1, Spring, pp. 65-102. Linz, S.J. (1994), ‘‘The privatization of Russian industry’’, RFE/RL Research Report, Vol. 3 No. 19 March, pp. 27-35. Linz, S.J. (1996), ‘‘Red executives in Russia’s transition economy’’, Post-Soviet Geography and Economics, Vol. 37 No. 10, November, pp. 633-51. Linz, S.J. (1997), ‘‘Russian firms in transition: champions, challengers and chaff’’, Comparative Economic Studies, Vol. 39 No. 2, Summer, pp. 1-36.

Ownership and employment in Russia 59

International Journal of Manpower 23,1 60

Linz, S.J. (1998), ‘‘Job rights in Russian firms: endangered or extinct institution?’’, Comparative Economic Studies, Vol. 40 No. 4, Winter, pp. 1-32. Linz, S.J. (1999), ‘‘Who is shouldering the burden of transition? An analysis of depreciation rates in Russian industry’’, Comparative Economic Studies, Vol. 41 No. 2-3, Summer-Fall, pp. 1-47. Linz, S.J. (2000) ‘‘Labor productivity in transition: a regional analysis of Russian industry’’, Economic Development and Cultural Change, Vol. 48 No. 4, July, pp. 685-718. Linz, S.J. (2001), ‘‘Restructuring with what success? A case study of Russian firms’’, Comparative Economic Studies, Vol. 43 No. 1, Spring, pp. 75-99. Linz, S.J. and Krueger, G. (1996), ‘‘Russia’s managers in transition: pilferers or paladins?’’, PostSoviet Geography and Economics, Vol. 37 No. 7, September, pp. 397-426. Linz, S.J. and Krueger, G. (1998), ‘‘Enterprise restructuring in Russia’s transition economy: formal and informal mechanisms’’, Comparative Economic Studies, Vol. 40 No. 2, Summer, pp. 5-52. Linz, S.J. and Martin, R.E. (1982), ‘‘Soviet enterprise behavior under uncertainty’’, Journal of Comparative Economics, Vol. 24 No. 1, March, pp. 24-36. Litisian, N. (1997), ‘‘Oborotnye sredstva, protsess obrashcheniia stoimosti kapitala neplatezhi’’, Voprosy ekonomiki, No. 9. Lizal, L. and Svejnar, J. (1997), ‘‘Enterprise investment during the transition: evidence from Czech panel data’’, Davidson Institute Working Paper # 60a, University of Michigan, East Lansing, MI, December. Markarov, V. and Kleiner, G. (1996), Barter v ekonomkie Rossii: osobennosti I tendentsii perekhodnogo perioda, TsEMI RAN, Moscow. Markarov, V. and Kleiner, G. (2000), ‘‘Barter in Russia’’, Problems of Economic Transition, Vol. 42 No. 11, March, pp. 51-79. Millar, J.R. (ed.) (1987), Politics, Work, and Daily Life in the USSR, Cambridge University Press, New York, NY. Nelson, L. and Kuzes, I. (1994), ‘‘An assessment of the Russian voucher privatization program’’, Comparative Economic Studies, Vol. 36 No. 1, Spring, pp. 24-36. Perevalov, Y. et al. (2000), ‘‘Is privatizaiton affecting the activity of enterprises?’’, Problems of Economic Transition, Vol. 42 No. 11, March, pp. 35-50. Radygin, A. (2000), ‘‘The redistribution of property rights in post-privatization Russia’’, Problems of Economic Transition, Vol. 42 No. 11, March, pp. 6-34. Rose, R. (1993), ‘‘The Russian response to privatization’’, RFE/RL Research Report, Vol. 2, November, p. 55. Rutkowski, M. (1996), ‘‘Labor market policies in transition economies’’, MOCT-MOST: Economic Policy in Transitional Economies, Vol. 6 No. 1, pp. 19-38. Standing, G. (1996), Enterprise Restructuring and Russian Unemployment: Reviving Dead Souls, Macmillan, Basingstoke. Thornton, J. (1997), ‘‘Restructuring production without market infrastructure’’, in Nelson, J. et al. (Eds), Transforming Post-Communist Political Economies, National Academy Press, Washington, DC, pp. 133-55. Thornton, J. and Mikheeva, N. (1996), ‘‘The strategies of foreign and foreign-assisted firms in the Russian far east: alternatives to missing infrastructure’’, Comparative Economic Studies, Vol. 38 No. 4, Winter, pp. 85-120. Transition Report (1998), Financial Sector in Transition, European Bank for Reconstruction and Development (EBRD), London. Vasil’ev, D. (1992), ‘‘Privatizatsiia: voprosy i otvety’’, Izvestiya, No. 224, 9 October.

Vavilov, A. and Kovalishin, E. (2000), ‘‘Problems of restructuring Russia’s debt’’, Problems of Economic Transition, Vol. 43 No. 1, May, pp. 6-25. World Bank (1996), From Plan to Market, Oxford University Press. Yakovlev, A. (2000), ‘‘The causes of barter, nonpayments and tax evasion in the Russian economy’’, Problems of Economic Transition, Vol. 42 No. 11, March, pp. 80-96.

Ownership and employment in Russia

Appendix. Russia’s privatization program Russia’s privatization program, written in 1991 and adopted in 1992, was designed to facilitate a widespread and timely transfer of ownership from the state. Even though the second aim listed in the privatization legislation was ‘‘to increase the productivity of enterprises’’,[46] speed and equity, rather than efficiency, dominated discussions and policies related to Russia’s privatization program. Indeed, policymakers argued at the time that speed and equity dominated efficiency considerations in the privatization program, given the political pressures to dismantle the command economy and the central planning bureaucracy before opposition forces could organize sufficiently to halt the process (Kucherenko, 1992; Chubais, 1994; Aslund, 1995a). The emphasis on speed shows up in the simple rules adopted for privatization. According to the Russian privatization law, firms with fewer than 200 employees and/or assets valued at less than 1 million rubles are considered to be ‘‘small’’ and thus subject to auction, tender, private placement, or liquidation (sales of assets). Firms employing 1,000 or more workers or having assets valued at 50 million rubles are eligible for the large-scale privatization options. When registering as a joint stock company, the first step of the privatization process, each firm had to indicate which variant was selected to distribute the shares. In the first variant, workers received free-of-charge 25 percent of the shares and had the right to purchase an additional 10 percent of the shares at a 30 percent discount off the nominal price[47]. Under this variant, management had the option of purchasing 5 percent of the shares, at a price not greater than 2,000 times the minimum wage paid by the firm. Thus, under variant #1, insiders could acquire up to 40 percent of the firm’s shares. The balance was to go to the State Property Committee (or its local equivalent) to be auctioned for vouchers or cash. Under the second variant of share distribution, prior to any auction, employees had the right to purchase up to 51 percent of the shares at the nominal price, paying for additional shares using vouchers or cash. Firms with a workforce between 200 and 1000 employees (or with assets valued between 1 million and 50 million rubles) could select either variant #1 or #2, or a third variant which involved the firm negotiating with the relevant authorities for a one-year agreement that covered production and employment commitments and liabilities of the firm. This variant required the firm to deposit an amount equal to 200 times the minimum wage, for which employees received 20 percent of the voting shares. At the end of the designated year, if the terms of the agreement had been fulfilled, the workers had the right to buy the shares they had been holding at the nominal price[48]. They also had the option of buying an additional 20 percent of the shares at a 30 percent discount. To select the share distribution variant required a two-thirds majority vote of the current workforce; without this majority, the first variant was the default distribution option[49]. The streamlined process for divesting the state of property included no real mechanism for getting accurate information to potential investors, nor any procedure for facilitating the survival of privatized firms in the chaotic transition environment. The emphasis on equity is evident in the vouchers offered free-of-charge to each Russian citizen[50]. Vouchers were touted as a means to substitute for savings lost as a consequence of the January 1992 price liberalization, thus ensuring that all Russians could participate in the privatization process. Hence the claim that vouchers would facilitate an equitable distribution of property to the general population.

61

The research register for this journal is available at http://www.emeraldinsight.com/researchregisters

International Journal of Manpower 23,1 62

The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0143-7720.htm

Regional diversity of employment structure and outflows from unemployment to employment in Poland Aleksandra Rogut and Tomasz Tokarski University of lLo´dz´, lLo´dz´, Poland Keywords Labour market, Unemployment, Employment, Poland Abstract Analyses factors determining the outflows from unemployment to employment across regions in Poland over the years 1992-98, employing the concept of the augmented matching function. Explores also the influence of the economic growth and the employment structure of the regional labour markets in Poland. Concludes that the values of outflows from unemployment to employment are closely and positively related to the number of unemployed and the number of vacancies, as well as to the economic growth rate; and that the employment structure of regional labour markets has a strong impact on outflows from unemployment to employment. The more a regional employment structure resembles the structure in European G7 countries, the higher the outflows from unemployment to employment.

1. Introduction This paper presents a cross section-time series analysis of the relation between employment structure in voivodship[1] (regional) labour markets and the outflows from unemployment to employment over the years 1992-1998, using the concept of an augmented matching function. The augmented matching function relates outflows from unemployment to the employment structure of regional labour markets, the number of unemployed, the number of vacancies, and the economic growth. This paper is a continuation of the research work undertaken by Kwiatkowski and Tokarski, 1997, 1998, and Kaczorowski and Tokarski, 1997, 1998. The matching function separates two causes of outflows from unemployment to employment: first, changes in the sectoral structure of employment and second, changes in the volume of unemployed and number of vacancies. Outflows from unemployment to employment are assumed to depend positively on unemployment, vacancies, economic growth rate, and how closely regional employment structure in Poland matches that in European G7 countries. The paper has the following structure. Section 2, briefly characterises the Polish labour market in the years 1990-98. The regional diversity of employment structure in respective voivodships is presented in Section 3. International Journal of Manpower, Vol. 23 No. 1, 2002, pp. 62-76. # MCB UP Limited, 0143-7720 DOI 10.1108/01437720210421303

A preliminary version of this paper has been presented at the First EALE/SOLE World Conference, Milan, June 2000. The authors are grateful to Prof. Belton M. Fleisher and to an anonymous referee for their comments and suggestions.

Section 4 presents the augmented matching function, which is used in the statistical analyses in Section 5. Section 6 concludes. 2. The Polish labour market in the transition period The Polish labour market in the transition period (1990-98), can be described as follows. (see also e.g. Kwiatkowski and Tokarski, 1997, 1998, 2000 or Kaczorowski and Tokarski, 1997, pp. 31-32): Although open unemployment was practically non-existent in Poland in the centrally planned economy, there was substantial hidden unemployment. The estimates of Rutkowski, 1990, suggest that the hidden unemployment rate in Polish industry in the late 1980s amounted to 25 per cent of total employment. The program of stabilisation and liberalisation of the Polish economy, introduced at the beginning of the year 1990, was connected with both a negative demand shock in the product market (the so-called transition recession), and a change in the market demand structure, described more fully in the next section. After January 1st, 1990, open unemployment has emerged , and by the end of 1991 amounted to more than 11 per cent. It is important to note that the proportional decline in employment in the transition recession (1990-1991) was less than the decline in GDP. Therefore, hidden unemployment must have been increasing as well. Economic growth recovery began in 1992, and was initially accompanied by a further reduction in employment and an increase in unemployment. We conclude that hidden unemployment declined. Since 1995 GDP growth has been accompanied by falling unemployment and increasing employment, but between 1995 and 1998, proportional employment growth has been smaller than GDP growth. This could mean that the hidden unemployment has fallen, or it could represent a mismatch of labour supply and demand. Mismatches might have resulted from substantial changes in the structure of product markets over the years 1990-98, which led to changes in the structure of labour demand, not matched by labour supply adjustments. Such rigidity may have been due to low geographic mobility and low skills mobility related to housing problems and to workers’ unwillingness or inability to retrain. We believe that our analysis of employment structure and its influence on the outflows from unemployment to employment contributes to an understanding of factors responsible for mismatching of the supply and the demand sides of labour markets in Poland. Analysing the volume and rates of quarterly outflows from unemployment to employment the authors conclude as follows. . First, both the quarterly outflows and the rates of these outflows from unemployment to employment over the years 1992-93 were relatively low (see Figures 1 and 2). The low outflows resulted from a reduction in hidden unemployment in the public sector. This reduction translates into GDP growth, low growth rates of labour demand, and low outflows from unemployment to employment.

Employment structure in Poland 63

International Journal of Manpower 23,1 64

Figure 1. Outflows from unemployment to employment in Poland over the years 1992-98 (thousands of people, quarterly)

Figure 2. Rates of outflows from unemployment to employment (percentage, quarterly)

.

.

Second, in 1994-97 both the outflows and their rates tended to increase enabled by rapid GDP growth. Third, in 1998, slower GDP growth was accompanied by a decline in both absolute outflows to employment and, to a lesser extent, in the rates of the outflows.

3. Regional diversity of the employment structure in Poland At the beginning of Poland’s transition to a market economy, both labour and product markets were highly segmented. Some of the important problems were as follows (see also Jajuga et al., 1994; Felbur, 1996; Buga and Kuszewski, 1997 or Lipowski, 1998)[2]: . The Polish economy inherited a structure of demand and supply on the product market that was seriously distorted by an absence of feedback between consumer preferences, the scale of aggregate demand, and market supply conditions, leading to a deformed price structure. Moreover, planners determined product demand, not according to domestic consumers and investors’ needs, but to the demand generated by co-operation within the framework of the Council for Mutual Economic Assistance and the Warsaw Pact. This situation made for a

.

relatively high market share of manufacturing and a rather low market share for services. The program of stabilisation and liberalisation of the economy launched at the beginning of 1990, involved reorienting some market prices. This reorientation brought about a change in the demand structure of the product market. Consequently, this altered both the structure of production and to a minor extent, the employment structure in the economy in general.

Employment structure in Poland 65

When comparing the structure of the gross value added in Poland to the advanced economies, which we characterize by the European G7 countries, namely France, Germany, Great Britain, and Italy, we observe the following: . First, the market share of industry and agriculture in Poland has been relatively high. This high market share stems from a level of development of the Polish economy relative to the G7. . Second, the market share of construction in gross value added in Poland is similar to the share in most well developed countries. . Third, the service sector share is relatively small. The low service sector share is connected to the heritage of the centrally planned economy, the low level of socio-economic advancement, and the relatively low prices of market services in Poland. The structure of the Polish economy measured by average value added and by average employment 1990-96 is shown in Table I. It is clear that employment shares and value-added shares are not in balance, as the structure of employment substantially differed from the structure of value added. As shown in Table II, the employment structure in Poland also differed substantially from that in advanced economies. The divergence of relative production and relative employment in Poland is probably attributable to a number of factors, including: . divergences in the productivity of industry and construction compared to that of services and agriculture; . hidden unemployemt in public industry, construction, and a majority of private agriculture; Sectors Agriculture Industry Construction Services

Gross value-added structure %

Employment structure %

7 38 8 47

25 26 6 43

Source: The authors’ estimates, on the basis of e.g. Nowy szacunek PKB za lata 1985-1995

Table I. Structure of gross value added and employment in Poland in the years 1990-96

International Journal of Manpower 23,1 66

.

the service sector share in added gross value is probaby underestimated, in view of the low service prices relative to product prices.

Our formal measure of the regional diversity of the employment structure in Poland with respect to the G7 countries is based on the following indicator[3]: qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Þ2 þ ðIit

I Þ2 þ ðCit
C
Þ2 þ ðSit
S
Þ2
it ðAit
A ð1Þ where i = 1, 2, . . ., 49 for the respective voivodships, t = 1992, 1993, . . ., 1998, and:
Þ – a share of employment in agriculture in a given voivodship i on Ait ðA December 31st of year t-1 (for European G7 countries, over the years 1991-94);
Iit ðI Þ – the rate of employment in industry in a given voivodship i on December 31st of year t-1 (for European G7 countries, over the years 1991-94);
Þ – the share of employment in construction in a given voivodship i on Cit ðC December 31st of year t-1 (for European G7 countries, over the years 1991-94);
Sit ðS Þ – the share of employment in services in a given voivodship i on December 31st of year t-1 (for European G7 countries, over the years 1991-94)[4]. The lower the value of the distance
it, the more a voivodship employment structure in a given year is similar to the employment structure of European G7 countries and
it equals 0 if regional employment structure is identical to that of G7 countries over the years 1991-94. Regional diversity as measured by
it is presented in Figure 3. A comparison of the employment ‘‘distance’’ measure with regional indices of socio-economic development including unemployment, GDP per capita, urbanisation, and rates of outflows from unemployment to employment (see Figures 4-6 and also Kwiatkowski and Tokarski, 2000; Rogut and Tokarski, 2000 or Radziwill, 1999) reveals the following:

Sectors

Table II. Employment structure in European G7 countries

Agriculture Industry Construction Services

France %

Germany %

Years 1991-94 Great Britain %

Italy %

EuroG7 %

5 21 7 67

4 30 8 58

2 23 7 68

8 23 9 60

5 25 8 63

Source: The authors’ estimates, on the basis of e.g. Nowy szacunek PKB za lata 1985-1995 and Yearbook of Labour Statistics

Employment structure in Poland 67

Figure 3. Divergences between the employment structure in regional labour markets in Poland and European G7 countries (the average in the years 1992-98)

Figure 4. Regional diversity of unemployment rates in Poland (the average in the years 1992-98, percentage)

.

The Warsaw voivodship had the most modern structure over period of time covered in this study. The other ‘‘modern’’ best-performing voivodships were: Szczecin, Gdansk, Wroclaw, L l o´dz´, Poznan, and Koszalin, most of which contain cities with over 500,000 citizens. The

International Journal of Manpower 23,1 68

Figure 5. Regional diversity of GDP per capita in Poland in the years 1995-96 (in PLN, fixed prices from 1995)

Figure 6. Regional diversity of the rates of outflows from unemployment to employment in Poland (the average in the years 1992-98, percentage)

second most ‘‘modern’’ group of voivodships were found in western regions, southwestern regions; Elblag and Olsztyn voivodships were also relatively ‘‘close’’ to the G7 countries in terms of their employment structure. The regions departing furthest from the structure of European G7 countries were mainly eastern voivodships.

.

.

.

. .

We note that the voivodships diverging most from the employment structure in advanced economies were characterised by relatively low unemployment rates (about 9.9-14.9 per cent), which could be explained by high hidden unemployment rate in these essentially agricultural regions. Moreover, the regions with the most ‘‘modern’’ employment structure, had the lowest unemployment rates (e.g. Warsaw and Poznan voivodships). There are regions that exhibited both a relatively modern employment structure (between 0.204 and 0.37), and high unemployment. Examples are Slupsk, Elblag, Olsztyn and Walbrzych voivodships. These voivodships’ very high unemployment rates (over 19 per cent) might have been due to the fact that at the beginning of the transition, state farms were abolished (Slupsk, Elblag, and Olsztyn voivodships), and redundancies were created in mining (Walbrzych voivodship). Another likely reason for the coexistence of an ‘‘advanced’’ employment structure and high unemployment in these regioins is that they are attractive tourist regions (including the coast, the Mazurian Lake District, and the mountainous Kotlina Klodzka), which explains the high share of service employment and a low
in these regions. The voivodships with the most modern labour market demand structure were also the wealthiest, with GDP per capita exceeding 6,900 PLN in 1995 prices (1USD&2.6PLN in 1995), and the voivodships with the most archaic labour market structure were poor or very poor, with GDP per capita generally under 5,500 PLN. The most modern voivodships were also the most urbanised. Regions with the most modern employment structures experienced relatively high rates of outflows from unemployment to employment (understood as the total of the outflows in a year relative to the average unemployment rate in that period). A more formal analysis follows.

4. Augmented matching function and flows out of unemployment The augmented matching function hypothesis relates the outflows from unemployment, the number of unemployed, and the number of vacancies as conditioned by labour-market effectiveness, which is a function of regional market structure
and other variables as specified below (Burda, 1993; p. 5, see also e.g. Stasiak and Tokarski, 1998, p. 79, or Kaczorowski and Tokarski, 1998; pp. 129-30)[5]: OðtÞ ¼ AðtÞF

U ðtÞ; V ðtÞ

where: OðtÞ = the volume of outflows from unemployment in the time t 2 0ðA_ < 0Þ, then in subsequent periods of time the same unemployment volume and number of vacancies will generate higher and higher (lower and lower) outflows from unemployment. This means that if A_ > 0ðA_ < 0Þ, then obtaining the same outflows from unemployment will be possible with a decreasing (increasing) number of the unemployed and of the job vacancies. If the labour market effectiveness coefficient A(t) is determined by some other macroeconomic variables, then the matching function is called an augmented matching function. We assume that: Fð0; V Þ ¼ FðU; 0Þ ¼ 0: @F @F > > 0: @U @V @2F @2F < 0 and < 0: 2 @U @U 2 which means that the matching function F is concave with respect to U and V. The following section presents the estimates of parameters of the exponential matching function: OðtÞ ¼ AðtÞ½U ðtÞu ½V ðtÞv

ð3Þ

where: OðtÞ; AðtÞ; UðtÞ, and V ðtÞ are as defined above, and the labour market effectiveness component is given by function A=A(
, g,t), where g is the @A @A economic growth rate and t is time, with @A @
< 0; @g > 0 and @t > 0. Hence, the closer an employment structure in a labour market approximates the employment structure in European G7 countries, (the lower is
), and the higher is economic growth, the greater is the labour market effectiveness coefficient. The negative relationship between A and
is explained by the fact that the lower
value is the closer is Beveridge curve to the zero coordinates. The hypothesised positive relationship between A and g is based on the assumption that the labour demand dynamic translates only partially into the new job vacancies. This means that under high economic growth, the unregistered vacancies and outflows from unemployment are increasing. The Beveridge curve shifts towards the zero coordinate. The increase in coefficient A in time could be interpreted in economic terms as increased effectiveness of labour market functioning due to factors unspecified in the model. This might be explained by the fact that at the beginning of the transition, neither labour offices, nor the unemployed were accustomed to active job seeking in the market. With development of the labour market and its institutions, the labour offices and the unemployed gained more experience. The labour offices became more efficient in finding existing

vacancies in the market, which (ceteris paribus) enhanced outflows from unemployment. The coefficient A in the matching function (3) is specified as follows: 1nA ¼ 0
1
þ 2 g þ 3 t

ð4aÞ

3 t

ð4bÞ

or: 1nA ¼ 0
1
þ 2 g


71

with: 0 2 0. Differentiating equations (4a) and (4b) with respect to time one can obtain: A_ ð5aÞ ¼
1
_ þ 2 g_ þ a3 A or: A_ 3 ¼
1
_ þ 2 g_ þ 2 A t

ð5bÞ

which shows that: .

.

.

1 is the growth rate of the labour market effectiveness coefficient A with respect to a change in the parameter
, where
_ = 0.01. 2 is the growth rate of the labour-market effectiveness coefficient with respect to a 1 percentage point increase in the economic growth rate. 3 is the growth rate of the labour market effectiveness coefficient with respect to other factors (e.g. improvement in the efficiency of job seeking by the unemployed, or the improved functioning of labour offices). € < 0, so Equation (4b) also implies that, with
_ ¼ g ¼ 0; A_ > 0 and A that under these conditions, labour-market effectiveness increases at a decreasing rate. (See also Kaczorowski, Tokarski, 1998, pp. 129-130).

From equations (3), (4a), and (4b) one can obtain: ln½O ¼ 0
1
þ 2 g þ 3 t þ U ln½U þ V ln½V 

ð6aÞ

3 þ U ln½U þ V ln½V  t

ð6bÞ;

ln½O ¼ 0
1
þ 2 g


Employment structure in Poland

which constitute a basis for the statistical analyses. Estimation results In order to estimate the impact of respective factors on the outflows of the unemployment to employment, we estimate the augmented matching functions derived from equations (6a) and (6b):

International Journal of Manpower 23,1 72

ln½OJit  ¼ 0
1
it þ 2 gt þ 3 t þ U ln½Uit  þ V ln½Vit  þ "it 3 ln½OJit  ¼ 0
1
it þ 2 gt
þ U ln½Uit  þ V ln½Vit  þ "it t

ð7aÞ ð7bÞ

where: OJit = total outflows from unemployment to employment in voivodship iði ¼ 1; 2; . . . ; 49Þ in year t ¼ 1992; 1993; . . . ; 1998 in thousands of people (source: statistical data obtained from the Central Bureau for Statistics, Labour Department); t ¼ the time variable assuming the value 92, 93, . . ., 98 in respective years;
it = the divergence of the employment structure in the regional labour markets and the employment structure in European G7 countries, which is calculated according to equation (1); gt = the GDP growth rate in Poland in year t (source: Nowy szacunek PKB za lata 1985-1995 and Biuletyny statystyczne, 1997-1998 editions); Uit = the average unemployment volume in voivodship i in the year t (in thousands of people), which is defined as the arithmetical means of the unemployment, estimated at the end of every quarter (source: Bezrobocie rejestrowane w Polsce, 1992-1998 editions); Vit ¼ the number of vacancies (in thousands) registered at labour offices in voivodship i in year t (source: statistical data obtained from Central Bureau for Statistics, Labour Department). "it – random terms. The parameters 0 ; 1 ; 2 ; 3 ; U and V in the equations above are interpreted in an analogous way to equations (6a) and (6b). Equations (7a) and (7b) were estimated by ordinary least squares. The estimated parameters of the augmented matching functions (7a) and (7b) are presented in Table III. The estimates of augmented matching functions presented in columns 9 and 10 include dummy variables for Sll upsk, Koszalin, Pill a, and Suwall ki voivodships, which had the highest residuals in the estimates reported in columns 1 through 8. The following conclusions can be drawn from the estimates of the matching functions: . Both specifications (7a) and (7b) account for a high proportion of the variance of outflows from unemployment to employment, as measured by the adjusted R2 . . The time variable is statistically significant in all of the estimations, both in estimations of outflows function (7a) and (7b). This indicates improvement in the balance of supply and demand sides of the labour market due to other factors than those accounted for in the model, as discussed above. Based on the estimates of specification (7a), we project that the annual growth rate of the outflows resulting from improved

Estimated parameters (t-values) Independent variables

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

– 8.757 – 9.233 – 6.881 – 7.310 8.218 9.022 5.968 6.799 – 7.164 6.548 (– 13.91) (– 14.32) (– 10.41) (– 11.31) (15.02) (14.95) (9.41) (10.67) (– 11.98) (11.09) t 0.089 0.096 0.067 0.074 – – – – 0.072 – (14.47) (14.71) (10.02) (11.13) (11.68) – – – – – 1/t – – – – 808.642 870.690 612.501 672.527 – 653.589 (– 14.65) (– 14.92) (– 10.00) (– 11.10) (– 11.65) – – 0.223 – – 0.360 – – 0.232 – – 0.359 – 0.266 – 0.266 (– 2.88) (– 4.87) (– 3.00) (– 4.86) (– 3.78) (– 3.77) g – – 4.666 5.511 – – 4.503 5.331 5.333 5.158 (6.48) (7.68) (6.18) (7.34) (8.02) (7.68) ln U 0.636 0.655 0.637 0.668 0.637 0.657 0.635 0.666 0.655 0.654 (20.58) (20.93) (21.81) (23.05) (20.74) (21.13) (21.80) (23.03) (24.38) (24.35) ln V 0.328 0.276 0.306 0.218 0.326 0.272 0.306 0.219 0.240 0.241 (16.33) (10.28) (15.88) (8.42) (16.27) (10.14) (15.90) (8.44) (9.91) (9.94) Sll upsk – – – – – – – – 0.232 0.232 (4.34) (4.34) Koszalin – – – – – – – – 0.142 0.142 (2.63) (2.63) Pill a – – – – – – – – 0.256 0.256 (4.78) (4.78) Suwall ki – – – – – – – – 0.223 0.223 (4.19) (4.19) 2 0.902 0.904 0.912 0.918 0.902 0.905 0.912 0.918 0.931 0.931 R 0.901 0.903 0.911 0.917 0.902 0.904 0.911 0.917 0.929 0.929 Adj. R2 DW statistic 1.968 1.910 2.177 2.157 1.984 1.925 2.175 2.154 2.042 2.039 No. of obs. 343 343 343 343 343 343 343 343 343 343

Employment structure in Poland

Cons.

.

.

.

efficiency of employment offices and of job-seeking behavior will probably continue to increase (ceteris paribus) in the coming years. The divergence of the structure of employment in the regional labour markets and the employment structure in labour markets of European G7 countries is a key determinant for outflows from unemployment in all estimates concerning this variable. Each approximation of the employment structure on regional labour markets to the employment structure of the European G7 markets by 0.01 raises the growth rate of outflows from unemployment to employment by around 2.2-3.6 per cent. The GDP growth rate is also an important factor. Each percentage point added to the GDP growth rate is translated into the growth in the outflows of around 4.5-5.5 per cent. Both the volume of unemployment and the number of vacancies are statistically important variables determining outflows from unemployment to employment. The elasticities of the outflows from unemployment to employment with respect to the number of unemployed and the number of vacancies averaged around 0.64-0.67 and 0.22-0.33, respectively.

73

Table III. Estimated parameters of augmented matching functions (7a) and (7b)

International Journal of Manpower 23,1 74

.

Due to regionally idiosyncratic factors, outflows from unemployment to employment in Sll upsk, Koszall in, Pila, and Suwall ki voivodships are on average higher than in the other voivodships (respectively by 23.2 per cent, 14.2 per cent, 25.6 per cent, and 22.3 per cent).

Conclusion Both the product market structure and the employment structure in Poland fundamentally differ from the product markets and the labour market structures in advanced economies. These results from the fact that gross value added and the employment structure in Poland are characterised by a high share of agriculture and industry, and a low share of services. The divergence of the structures are affected by the lower GDP per capita in Poland and possibly by the archaic structure of the economy inherited from years of a centrally planned economy. The analysis of the differences in the employment structures of respective voivodship shows that only the Warsaw voivodship structure is similar to the employment structure in advanced economies. Also, labour structures in western and north-western voivodships of Poland are relatively modern. These voivodships are closely related to European G7 countries in terms of the employment structure as opposed to that of the eastern agricultural voivodships. It should be stressed that the agricultural voivodships, characterised by an archaic employment structure, feature a relatively low unemployment rate, a low rate of outflows from unemployment to employment, and a low GDP per capita. Likewise, the high hidden unemployment and a relatively small number of new vacancies (low dynamics of the labour market demand side) further illustrates rural areas of Poland. The voivodships that feature a relatively modern employment structure show a very high rate of outflows from unemployment to employment (with the exceptions of mining Walbrzych and Katowice voivodships). Furthermore, the analysis of the impact of employment structure on outflows from unemployment to employment, based on the augmented matching function, suggests that the more similar the employment structure of the regional labour market is to the structure of European G7 countries, then the higher the outflows from employment to employment. In analysing the changes in the sectoral structure of employment in Poland (see for example Kwiatkowski et al., 2000) one can notice that in the above mentioned period the changes were relatively small. Moreover, in the eastern, agricultural areas of Poland these changes were significantly slower than in the western and the north-western areas of Poland. This means that the mechanisms of market transition in Poland do not lead to convergence in regional development of labour markets. Consequently, housing barriers in Poland generate a very low geographical mobility of the labour force. The inflexibility of labour markets in eastern Poland will worsen without an active regional macroeconomic policy.

An impact of the GDP growth rate on the volume of the analysed outflows is also very important. This suggests that high rates of economic growth lead to an increase in labour demand, which (ceteris paribus) translates into a rise in outflows from unemployment to employment. Another important factor is the positive influence of the time variable on the value of the outflows. The increase is due to both a more intensive job search by the unemployed and to a better functioning of the labour offices. The number of the unemployed and of the vacancies registered at labour offices is an important factor influencing the volume of the outflows concerned. The analyses of elasticities of the outflows from unemployment to employment relative to the number of unemployed and to the number of vacancies illustrate the variation of 0.64-0.67 and 0.22-0.33. This suggests that job seeking other than at labour offices (i.e. directly at firms) has affected the volume of outflows from unemployment to employment in Poland. This was more important than waiting for the jobs offered at labour offices (which, in general, register less financially attractive vacancies). Because sectoral changes in employment structure in rural areas are slower than in more modern voivodships of Poland, so the main directions in active regional policy should be the following. First, the active regional policy should be connected with investments in socio-economic infrastructure, which can create new vacancies in the service sector. Second, the government should use financial support from the EU to activate the development of small and medium size towns. This can be the reason for the changes of employment structure in rural areas, that will generate new jobs mainly in the service sector and make a labour market more flexible. Notes 1. The voivodship is an administrative unit in Poland. 2. Unfortunately there is no possibility to disaggregate further the sectoral structure of employment in Poland (for instance disaggregate industry employment into the employment in high technological industry and other industry) because of the lack of regional statistical data for Poland. 3. One can suppose that more important will be the comparison of the Polish labour market with countries like Portugal, Greece, Turkey, or Spain because of smaller gap in GDP per capita between Poland and these countries (in contrast to the Poland – European G7 countries gap). But the indicator
is a measure of how archaic the sectoral structure is in the regional labour market in Poland compared to that in the most modern European economies. Therefore, European G7 economies are a benchmark. 4. The statistical data concerning the employment structure in Poland come from Roczniki statystyczne wojewo´dztw, Central Bureau for Statistics, Warsaw, 1992-1998 editions (the data for employment from 1997 show the state on 31st September, while the data concerning the employment structure in European G7 countries come from Yearbook of Statistics, ILO, Geneva, 1998. The choice of the period covered by the data on structure of employment in European G7 countries (the years 1991-94) was dictated by accessibility of the statistical data. 5. All the variables presented in this section are continuous and differentiable functions of _ dx time. The notation x_ ðtÞ  dx dt or x  dt will stand for the derivative of x with respect to time

Employment structure in Poland 75

International Journal of Manpower 23,1 76

t, or in economic terms, increase in the value of this variable in time. The notation 2 €xðtÞ  €x  ddt2x will be termed as the second derivative of x with respect to time t. References Bezrobocie rejestrowane w Polsce, Central Bureau for Statistics, Warsaw, 1991-1998 editions. Biuletyn Statystyczny, Central Bureau for Statistics, Warsaw, 1992-1998 editions. Buga, J. and Kuszewski, T. (1997), Zmiany strukturalne w gospodarce Polski i wybranych krajo´w _wiata. Analiza poro´wnawcza, Ekonomista, No. 3, pp. 323-44. Burda, M.C. (1993), Modelling Labor Market Dynamics in Eastern Germany: a Matching Function Approach, INSEAD, Wissenschaftszentrum Berlin and CEPR. Felbur, S. (1996), Struktura gospodarki Polski i jej dostosowania do integracji z Unia˛ Europejska˛, Ekonomista, No. 4, pp. 425-51. Jajuga, K., Panasiewicz, Z. and Strahl, D. (1994), Wzorce zmian strukturalnych w Polsce w latach dziewie˛c´dziesia˛tych, IRiSS, Working Paper, No. 24, Warsaw. Kaczorowski, P. and Tokarski, T. (1997), Restrukturyzacja a odplywy z bezrobocia. Analiza oparta na rozszerzonej funkcji dopasowan´, Wiadomos´ci Statystyczne, No. 11, pp. 22-38. Kaczorowski, P. and Tokarski, T. (1998), ‘‘Niekto´re determinanty odplywo´w z bezrobocia w Polsce robocza’’, in Kwiatkowski, E. (Ed.), Pzepll ywy sill y a efelty altwnej pohtyki parı´stwa na rynku pracy w Polsce, Wydawrictwo Uniwersytetu Lodzkengo, Lodz, pp. 115-38. Kwiatkowski, E. (Ed.) (1998), Przepll ywy sill y roboczej a efekty aktywnej polityki pan´stwa na rynku pracy w Polsce, Wydawnictwo Uniwersytetu L l o´dzkiego, L l o´dz´. ´stwa Kwiatkowski, E. and Tokarski, T. (1997), Efekty polityki pan wobec rynku pracy w Polsce. Analiza na podstawie funkcji dostosowan´, Ekonomista, No. 3, pp. 345-72. Kwiatkowski, E. and Tokarski, T. (1998), ‘‘Macoreconomic effects of active labour market policies in Poland’’, Comparative Economic Research (Central and Eastern Europe), Vol. 1 No. 1 pp. 1-23. Kwiatkowski, E. and Tokarski, T. (2000), ‘‘Employment structure and employment flexibility in Poland in transition’’, International Review of Economics and Business, June. Kwiatkowski, E., Rogut, A. and Tokarski, T. (2000), Regionalne struktury pracuja˛cych a odpll ywy z bezrobocia do zatrudnienia in Wzrost gospodarczy, restrukturyzacja i bezrobocie w Polsce. Uje˛cie teoretyczne i praktyczne, Katedra Ekonomii UL l ,L l doz´. Lipowski, A. (1998), Analiza zmian strukturalnych w trzech podstawowych sektorach gospodarki: rolnictwie, przemys´le i uslugach, Working Paper of INE-PAN, November. Nowy szacunek PKB za lata 1985-1995, Studia i Prace Zakladu Badan´ Spoll eczno-Ekonomicznych, Central Bureau for Statistics and Polish Academy of Sciences, Vol. 263, Warsaw, 1999. Produkt Krajowy Brutto wedll ug wojewo´dztw za 1996 rok, Central Bureau for Statistics in Katowice, Warsaw, Katowice, June, 1998. Radziwill l , A. (1999), Zro´z_ nicowanie regionalne bezrobocia w Polsce. Perspektywy zro´wnowa_zonego rozwoju, Studies & Analyses, No. 197, CASE Foundation, Warsaw. Rocznik statystyczny wojewo´dztw, Central Bureau for Statistics, Warsaw, 1991-98 editions. Rogut, A. and Tokarski, T. (2000), Regionalne zro´z_ nicowanie sytuacji na rynku pracy w Polsce w latach dziewie˛c´dziesia˛tych, Studia Prawno-Ekonomiczne, Volume LXI. Rutkowski, M. (1990), Labour Hoarding and Future Unemployment in Eastern Europe: The Case of Polish Industry, Discussion Paper, LSE, London. Solow, R.M. (1998), ‘‘What is labour market flexibility? What is it good for?, Proceedings of the British Academy, Vol. 97. Stasiak, J. and Tokarski, T. (1998), ‘‘Analiza odpll ywo´w z bezrobocia. Funkcja dopasowan´’’, in Kwiatkowski, E. (Ed.), Przepll ywy sill y roboczej a efekty aktywnej polityki pan´stwa na rynku pracy w Polsce, Wydawnictwo Uniwersytetu L l o´dzkiego, L l o´dz´, pp. 74-92. Yearbook of Labour Statistics (1998), ILO, Geneva.

The research register for this journal is available at http://www.emeraldinsight.com/researchregisters

The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0143-7720.htm

Worker reallocation during Estonia’s transition to market

Worker reallocation in Estonia

Milan Vodopivec The World Bank (HDNSP), Washington, DC, USA

77

Keywords Market economy, Labour market, Estonia Abstract Based on consecutive labor force surveys, this study examines labor market dynamics during the first decade of the Estonian transition to market. The results show that, similar to other transition economies: Estonia’s employment and labor force was reduced; patterns of mobility profoundly changed – labor market flows intensified and previously nonexistent transitions emerged; and some groups of workers were disproportionally affected, chief among them the less educated and ethnic minorities. But Estonian fundamental free market reforms also produced labor market outcomes that differ significantly from those in other transition economies – above all, the intensity of worker and job flows in Estonia’s transition have surpassed those in most other transition economies. This was achieved by deliberate policies aimed at stimulating job creation and employment, above all by low employment protection and other policies geared toward increasing employability and strengthening the incentives of workers. Moreover, under the dynamic Estonian labor market adjustment, marginal groups have fared better than those in more protective labor markets of other transition economies.

1. Introduction Since the collapse of the socialist economic system, there has been an intense research effort geared toward monitoring the progress of transition economies. For various reasons, most studies have focused on Central European economies. The process of transition began somewhat earlier in these countries, making Central European countries useful test cases for other transition economies. Another advantage of these countries was availability of data on individual workers which allowed detailed analysis of their labor markets. Knowledge of labor markets in Bulgaria, the Czech Republic, East Germany, Hungary, Poland, and Slovenia is therefore fairly advanced[1]. Among the conclusions that can be drawn from the Central European experience is that transition to a market economy requires large redeployment of workers from noncompetitive sectors, that these redeployments can have substantial costs in terms of unemployment and lower production, and that real wages can drop substantially in the early years. At the same time, efforts to avoid the costs of transition by encouraging early retirement, maintaining high unemployment benefits or pensions, subsidizing unprofitable firms, or restricting bankruptcies can make the problems worse by taxing the emergence of new firms and employment opportunities. Transition may also have The author is grateful to the Statistical Office of Estonia for providing the data used in the study, to John Haltiwanger and Peter Orazem for numerous insights and comments, and to Debabrata Das and Ylle Petai for excellent research assistance. The paper is part of a project ‘‘Labor Market Adjustment in Estonia’’ financed by the World Bank’s Research Support Budget (RPO 679-71).

International Journal of Manpower, Vol. 23 No. 1, 2002, pp. 77-97. # MCB UP Limited, 0143-7720 DOI 10.1108/01437720210426885

International Journal of Manpower 23,1 78

increased relative productivity of the better educated workers, both because of the decline of low-skill intensive sectors such as manufacturing and agriculture, and because of the increased need for decision-making skills. Much less is known about the progress of transition in countries which were states of the former Soviet Union. There is little data from the period before transition since these countries were previously tied into the Soviet statistical ministry which held all records. Data during transition have been fragmented and incomplete, since labor force surveys have been slow to replace the previous statistical system based on a census of workers. These countries were also less integrated into the international community of scholars, and therefore their available expertise was much more limited than was the case for Central Europe. Finally, transition began later in these countries, and so the attention of researchers concentrated on countries where the transition was at a more advanced stage. Nevertheless, research on former Soviet economies can shed additional light on many important dilemmas faced by transition economies. For example, Central European economies have afforded a much more generous social safety net and more protective labor market policies than former Soviet states, including government sponsored early retirement programs, generous pensions, longer eligibility for and higher replacement rates of unemployment insurance, and longer advance notification of layoffs. These policy differences across countries will allow researchers to compare the relative success of interventionist versus free market labor policies. Moreover, the trade within the former Soviet states was much more integrated than the trade within other economies of the socialist block, and the collapse of trade may have generated an even greater shock to these economies. Therefore, the scale of labor reallocation and displacement caused by transition to market may be larger in the former Soviet states. This paper reviews progress of labor market transition in Estonia, the country that makes a particularly interesting case in that it opted for much more fundamental and intense free market reforms than other transition economies. The country placed few roadblocks in the way of international trade, layoffs, bankruptcies, or foreign ownership of firms. Estonia also set low levels of unemployment benefits, pensions, and minimum wages. Based on consecutive labor force surveys, this study examines patterns of labor reallocation during the Estonian transition to market. First, it examines how employment and unemployment adjusted to changes in labor market demand and supply, and identifies worker and job flows which have been responsible for changes in employment and unemployment. Second, it identifies groups of workers which have been particularly hard-hit by transition. Based on the fact that Estonia introduced more fundamental and intense reforms than other transition economies, the following two working hypotheses about a labor market adjustment in Estonia can be formulated. First, the intensity of worker and job flows in Estonia’s transition has surpassed those in other transition economies, thereby contributing to the fast reallocation of labor

to more productive uses. Second, Estonian dynamic labor market adjustment contributed to more favorable labor market outcomes for marginal groups of workers (young workers and workers with fixed-term appointments) than the adjustment of more protective labor markets of other transition economies. The organization of the paper is as follows. Section 2 describes the data sources. Section 3 focuses on aggregate labor market transitions, and Section 4 examines determinants of labor market transitions, identifying groups of the population which benefited from and those which were adversely affected by transition. Section 5 concludes with reviewing the main findings and policy issues raised by them. 2. Data sources The study’s main data source is the Estonian 1995 Labor Force Survey (retrospectively covering the period of 1989-95); the subsequent 1997 and 1998 surveys are also used. The universe for the sampling was the 1989 census of the Estonian population, for the 1995 survey, and the population database of the Statistical Office, for the later two surveys. The sample size for the 1995 survey was just below one percent of the adult population (12,246 individuals), with 9,608 (77 percent) interviewed. The sample size for the later two surveys was about half of the 1995 survey, with somewhat higher response rate. Most of the nonresponse was attributable to failure to locate an address for the individual, and, in the 1995 survey, to emigration of Non-Estonians following Estonian secession from the former Soviet Union. For each consecutive survey, a completely new sample was drawn. In the 1995 survey, respondents were asked about their labor market status as of January 1989 and all subsequent changes of the status. For each spell of employment, they also reported industry of employment, type of employment, and a number of employer attributes. The survey also elicited information on human capital attributes including education, work experience, and job tenure, and demographic information on age, ethnicity, and gender. Beside standard questions about the current labor market activity, the 1995 survey also asked retrospective questions on wages and employment from the period before transition up to 1995, and the subsequent surveys covered the gaps between the two consecutive surveys. This required recollection of labor activities up to six years before the time of interview, which makes the collected data suspect to recall bias. To minimize this bias, enumerators were carefully trained to cross check answers for employment and unemployment spells to insure consistency. Moreover, research indicates that individuals recall traumatic events more readily, and changes in labor market status are likely to have been particularly memorable in an economy transiting from a system with many years of constant steady employment. Indeed, data validation checks show that the recall bias has been very limited. For example, the data on economic activity from the 1989 census corresponded quite well with the survey responses from the 1995 survey, and the majority of the discrepancies are attributable to changes in labor force definitions[2]. Similarly, the estimates

Worker reallocation in Estonia 79

International Journal of Manpower 23,1 80

of the number of registered unemployed obtained from the surveys quite closely match the data from the registers of Employment Offices (for example, for the second quarter of 1997 the survey estimate is 36,400 and the Employment Office number 35,700 – see Statistical Office of Estonia, 1998). 3. Aggregate worker flows Former socialist economies face a huge task of reallocating labor according to market forces[3]. Below we explore how has this task was implemented in Estonia. We describe trends in aggregate employment, unemployment, and inactivity, and analyze workers flows and transition probabilities. This discussion also allows us to test the hypothesis that Estonia’s fundamental labor market and other reforms produced a more fluid labor market than the one found in other transition economies[4]. Changes in employment, unemployment, and inactivity. Estonia’s transition produced dramatic changes: employment was strongly reduced, and the number of unemployed and inactive individuals increased. Many of the employers reduced their workforces or closed down, reducing employment by 23 percent during 1989-98 (Table I). Worsening of labor market conditions produced a surge of unemployment, from virtually non-existent in 1989 to over 70 thousand – 10 percent unemployment rate – in 1998. The number of inactive individuals also increased from 255 in 1989 to over 330 thousand in the latter half of the 1990s. Because of a strong reduction of labor force participation, the labor force shrunk from 841 thousand in 1989 to 711 thousand ten years later, or by 15 percent. Particularly large adjustments occurred in the early 1990s. Worker flows and transition probabilities. During the transition, flows across labor market states were strongly affected. Unemployment has become an important destination of those exiting employment, and flows to inactivity also strongly increased (see Tables II and III). As a sign of the strong job creation capacity of the Estonian economy, direct job-to-job transitions strongly increased, more than doubling their pre-transition rate during 1992-95, the period of most intensive restructuring. As another significant development, the outflow from unemployment has become an important source of employment, contributing nearly 25 percent of accessions since 1995. It is notable that in spite of obvious tightening of the conditions of the labor market, accessibility of employment increased, and the chances to find a job if unemployed did not deteriorate very much. How dynamic was the Estonian adjustment? Comparison of worker flows in Estonia and Slovenia. Because communism generated large production imbalances, worker – and job! – reallocation is undoubtedly a sine qua non for the overall success of the transition reforms. What can we say about the scale of reallocation that took place in the Estonian economy? Was it large or small? To address this question, we compare worker flows during Estonian transition with those in OECD and other transition countries, particularly with Slovenia[5].

1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 A. Employment Total Men Women B. Unemployment Total Men Women

836 424 412

828 425 404

815 423 392

791 417 374

716 378 338

684 362 321

652 339 313

644 334 309

632 328 304

640 332 308

81 6 3 3

4 2 2

7 4 3

17 8 9

44 25 19

57 28 29

71 41 30

71 41 30

79 42 37

71 40 31

C. Inactivity Total Men Women

255 93 162

269 96 173

282 97 185

294 98 195

320 110 210

329 118 212

339 125 214

339 125 214

336 128 209

333 125 208

D. Labor force Total Men Women

842 427 415

832 427 406

822 427 395

808 425 383

760 403 357

741 390 350

723 380 343

715 375 339

711 370 341

711 372 339

E. Memorandum items Unemployment rate (%) Men Women Labor force participation rate (%) Men Women

Worker reallocation in Estonia

0.7 0.7 0.6

0.5 0.5 0.6

0.9 0.9 0.8

2.1 1.9 2.3

5.8 6.1 5.3

7.6 7.2 8.2

9.8 10.8 8.8

9.9 11.0 8.8

11.1 11.4 10.9

10.0 10.7 9.2

76.7 82.1 71.9

75.6 81.6 70.1

74.5 81.5 68.1

73.3 81.2 66.3

70.3 78.5 63.0

69.2 76.8 62.3

68.1 75.2 61.7

67.8 75.0 61.3

67.9 74.4 62.0

68.1 74.8 62.0

Sources: For 1989-94, ELFS95; for 1995-96, ELFS97; and for 1997-98, ELFS98

Slovenia is chosen because its labor market policies contrast sharply with the Estonian ones. Both countries faced similar initial conditions of transition: they both needed to correct the accumulated production imbalances, and to re-orient their trade to Western partners after their countries disintegrated in the early 1990s. But they differ significantly in the boldness of transition reforms. Estonia followed a liberal approach, imposing few barriers to worker dislocations, meager support of the unemployed, and no effective wage floor. In contrast, Slovenia adopted a much more protective approach, with significant barriers to job dislocation, and rigid wage setting (see Appendix for a more detailed comparison of transition reforms in Estonia and Slovenia). As shown in Table IV, accession and separation rates in Estonia were only slightly below long-term rates in mature market economies, and Estonia’s rates even exceed those of Italy, France, Sweden, and Japan. In contrast, Slovenia’s separation and accession rates were lower then those of any of the comparison OECD countries and much lower than in Estonia (note that the Czech Republic is the only comparison country with lower rates). It has to be remembered that the

Table I. Labor market stocks, 1989-98 (in thousands, as of the beginning of the year)

International Journal of Manpower 23,1 82

Table II. Separation probabilities – employment, unemployment, and inactivity, 1989-97 (percentages)a

1989 1990 1991 1992 1993 1994 1995 1996 1997 A. Percent of individuals employed as of January of the year, who by next January: Stayed in the same employment 87.6 84.1 Changed a job 7.5 9.6 Become unemployed 0.3 0.7 Became inactive 4.6 5.6

79.8 12.1 1.7 6.4

69.6 15.9 4.8 9.7

70.4 16.9 5.3 7.4

72.0 16.3 5.2 6.6

83.7 9.0 4.2 3.0

77.9 81.4 12.2 9.9 4.9 4.7 5.0 4.0

B. Percent of individuals unemployed as of January of the year, who by next January: Stayed unemployed 35.8 45.9 47.1 44.1 Found a job 62.1 43.0 43.2 46.5 Became inactive 2.0 11.1 9.8 9.3

43.4 44.7 12.0

39.4 49.9 10.7

59.8 33.6 6.6

53.5 56.4 37.4 37.2 9.2 6.4

C. Percent of individuals inactive as of January of the year, who by next January: Stayed inactive 84.7 83.1 Found a job 14.5 15.4 Became unemployed 0.8 1.5

83.8 12.8 3.4

79.5 15.6 4.9

90.8 6.2 3.0

86.3 88.7 8.5 7.4 5.1 3.8

82.5 15.5 1.9

82.1 14.3 3.6

Notes: a Separations are defined on the basis of the first transition undertaken after January in the same calendar year. For example, the transition of a person who is employed in January, becomes unemployed in June, and re-employs in November and remains employed in December would be counted as ‘‘became unemployed’’, even though the person was employed both at the beginning and at the end of the year. Based on population aged 15-69 Sources: For 1989-1994 – ELFS95; for 1995-1996 – ELFS97; and for 1997 – ELFS98

reported rates of transition countries relate to the period when high the intensity of labor turnover was highly desirable so as to correct past imbalances, and that the reported rates of the OECD countries cover complete business cycles. Although they were similar in the pre-transition period, Estonian employment accession and separation rates well exceeded Slovenian ones during the period when reforms were fully in place (Figures 1 and 2). Moreover, there is a striking difference in the evolution of job-to-job transitions between the two countries: in Slovenia, the intensity of job-to-job transitions decreased; in contrast, in Estonia the intensity strongly increased (Figure 3). Moreover, the transition probabilities from unemployment were also more favorable in Estonia: the probability of exit to employment was higher and to inactivity much lower than in Slovenia (Figures 4 and 5)[6]. Differences between Slovenia and Estonia extend also to the area of job creation and job destruction. In Estonia both job destruction and, with a lag, job creation rates increased tremendously (see Haltiwanger and Vodopivec, 2002). In contrast, the Slovenian economy witnessed a much lower intensity of both job destruction and job creation (see Figure 6). The most stunning difference is on the job creation side, where the rates of Slovenia never even approach the order of magnitude of the Estonian rates. Except for 1994, job creation rates were at or below 1 percent per year, much lower than in Estonia. The only similarity

1989 1990 1991 1992 1993 1994 1995 1996 1997 A. Percent of individuals employed as of January of the next year, who Stayed in the same employment Changed a job Exited from unemployment Exited from inactivity

during the year: 86.8 83.1 78.6 7.7 9.7 12.1 0.7 0.8 1.2 4.0 4.7 5.3

67.2 15.1 2.8 5.3

68.8 16.8 4.6 5.4

70.6 16.6 5.3 6.4

82.2 9.1 4.5 3.0

76.6 80.6 11.9 10.1 5.3 5.2 4.3 3.9

the year: 28.3 20.6 40.0 56.0 31.8 23.5

16.5 62.4 21.1

33.0 51.5 15.4

33.9 46.0 20.1

59.2 31.9 8.9

47.3 54.6 38.0 31.5 14.7 14.0

during the year: 83.8 82.4 81.4 16.1 17.3 18.2 0.0 0.3 0.4

76.4 22.9 0.7

82.4 15.9 1.7

82.9 14.5 2.6

91.8 6.5 1.7

88.3 90.9 9.8 7.4 2.0 1.7

B. Percent of individuals unemployed as of January of the next year, who during Stayed unemployed 44.3 Exited from employment 33.1 Exited from inactivity 22.6 C. Percent of individuals inactive as of January of the next year, who Stayed inactive Exited from employment Exited from unemployment

Notes: aAccessions are defined on the basis of the last transition undertaken before the January of the next year. For example, the transition of a person who is employed in January, becomes unemployed in June, and re-employs by November and remains employed in December, would be counted as ‘‘Exited from unemployment’’, even though the person was employed both at the beginning and at the end of the year Sources: For 1989-1994 – ELFS95; for 1995-1996 – ELFS97; and for 1997 – ELFS98

between the countries was the declining job destruction rate as the transition progressed. But even at the height of restructuring, the Slovenian job destruction rate remained below 60 percent of the maximum Estonian job destruction rate[7]. How can the above differences in worker and job flows between Estonia and Slovenia be explained? One explanation is offered by the theoretical model of Blanchard (1998), which points to employment protection policies as a source of ‘‘sclerosis’’ of the labor market, that is, of the low intensity of the labor market flows. In particular, his model shows that higher employment protection leads to: reduction of labor turnover as well as lower job destruction and job creation intensity of the economy; longer unemployment duration or, equivalently, lower probabilities of exit from unemployment; and higher unemployment of marginal groups of workers, because of their impaired access to jobs. As described above, the two countries differed sharply in employment protection polices, so the above model can be fruitfully applied to them. Indeed, differences in both worker and job flows between Estonia and Slovenia are fully consistent with the model predictions: as describe above, Estonia has witnessed a much more intense labor turnover, larger job destruction, and job creation rates, and higher probabilities of exit from unemployment. (Evidence that a less ‘‘sclerotic’’ labor market disfavors marginal groups is presented below.)

Worker reallocation in Estonia 83

Table III. Accession probabilities – employment, unemployment, and inactivity, 1989-97 (percentages)a

International Journal of Manpower 23,1 84

Accession rate

Separation rate

Transition countries (average for selected years) Bulgaria (1991, state firms) Czech Republic (1994-98) Estonia (1989-91) Estonia (1992-94) Estonia (1995-97) Hungary (1991, state firms) Poland (1991) Slovenia (1990-96)b Slovenia (1989-95)c

12.9 n.a. 15.5 27.3 19.3 20.6 13.0 13.2 n.a.

31.5 9.0 16.2 29.3 19.0 30.5 28.0 18.2 13.0

OECD countries (average for selected years between 1971 and 1984) Unweighted average of the presented countries Finland France Germany Italy Japan Sweden United Kingdom United States

25.5 37.0 18.6 28 16.7 15.7 18.4 22.1 47.8

26.1 35.5 18.3 28.9 17.5 15.6 18.3 25.6 49.4

Notes: aAccession rate is defined as the annual number of accessions per 100 employees in a calendar year; separation rate is defined as the annual number of separations per 100 employees in a calendar year; b Obtained from group data reported by enterprises; c Obtained from data on individual labor turnover Table IV. International comparison of accession and separation ratesa

Sources: For Bulgaria, Hungary, and Poland: Boeri (1998); for Czech Republic, Sorm and Terrell (1999); for Slovenia: Abraham and Vodopivec (1993), internal material of Statistical Office of Slovenia based on Labor Force Survey and monthly employers’ reports on accessions and separations; for Estonia, own computations based on 1995 Estonian Labor Force Survey; for OECD countries, OECD (1994)

Fluidity of the Estonian labor market and its large job creation capacity is underscored by the countercyclical behavior of both quits and job-to-job transitions. There was a surge of both quits and job-to-job transitions during the period of most intense output reduction in 1992-93. Even during years when output was falling, the percentage of quitters who exited directly to another job remained quite stable. This deviates sharply from the evidence from the West as well as other transition economies, where quits are procyclical (for evidence on Western economies, see OECD, 1994, p. 64; for evidence on Poland, see Blanchard, 1997). As for job-to-job flows, evidence from Slovenia and the Czech Republic (Sorm and Terrell, 1999) shows that these flows were reduced when output declined. The performance of the Estonian labor market is even more remarkable when one considers that throughout the transition, about half of the

Worker reallocation in Estonia 85

Figure 1. Accession rates to employment, Estonia and Slovenia

Figure 2. Separation rates from employment, Estonia and Slovenia

Figure 3. Job-to-job probabilities, Estonia and Slovenia

International Journal of Manpower 23,1 86

Figure 4. Unemployment-toemployment probabilities, Estonia and Slovenia

Figure 5. Unemployment-toinactivity probabilities, Estonia and Slovenia

Figure 6. Job creation and destruction, Estonia and Slovenia

workers who lost jobs through layoffs or bankruptcies transferred to a new job without an intervening period of unemployment. The above discussion highlighted large differences in labor market adjustment between Estonia and Slovenia. Can we say which adjustment path is more desirable, that is, more efficient and fair? I believe the answer is yes. The

above results suggest that the Estonian adjustment has been more efficient (I return to the fairness implications in the next section). Indeed, productive destruction requires that newest technologies and improved management are created and outdated units closed down; because of the vast accumulated initial imbalances, in socialist economies this task was of much larger proportions. High job creation and destruction of the Estonian economy allowed the country to successfully address ‘‘technological sclerosis’’ that permitted many lowproductivity units to survive in other transition economies. That the resulting restructuring was indeed efficient can be attested by the dynamics of Estonian job flows: a surge of job destruction was accompanied by strong job creation, and eventually the rate of job creation even surpassed the rate of job destruction (see the discussion on the cumulative effect on restructuring of a recessionary shock in Caballero and Hammour, 2000). Indeed, many laid off workers found new jobs without experiencing unemployment, the result underscored by large job-to-job transition probabilities. The fact that the job intensity and worker flows in Estonia in the mid-1990s resemble the intensity in mature market economies adds credibility to this argument. Second, the recovery of economic growth in Estonia has been more vigorous than the one in Slovenia. During a four-year period of resumed growth, for example, Estonian output recovered by 26 percent, in contrast to 16 percent recovery of Slovenia. 4. Determinants of transitions: winners and losers in the Estonian labor market The costs of transition to a market economy are unlikely to be borne equally by all members of the society. Some groups of workers (for example, older and less educated workers) may face lower chances of (re)employment; other groups (for example, women and minorities) may be particularly vulnerable to job loss. This section examines these questions by identifying the determinants of labor market transitions, and examining how their influence has changed during the transition. It also compares labor market outcomes of marginal groups with those in other transition economies and thus provides a test of the hypothesis that the Estonian dynamic labor market adjustment contributed to more favorable labor market outcomes for marginal groups of workers. 4.1 Transition probabilities for different demographic and skill groups To determine how individual’s demographic and skill characteristics influence his/her probability of exit from different labor market states, we estimated multinomial logit models for exits from employment, unemployment, and inactivity. To learn how the effects associated with individual’s characteristics have changed during the transition, we separately estimated multinomial logit models for the pre-transition and transition period, and compared the results. Pre-transition results refer to 1989, arguably the last year in which typical conditions of communism prevailed, and transition results to 1994, the year when transition was already in a mature phase. The exception is the analysis of unemployment, for which only the results for the transition period are

Worker reallocation in Estonia 87

International Journal of Manpower 23,1 88

provided, because the number of pre-transition unemployed included in the survey was too low. Below we present evidence on the effects on women, ethnic minorities, and groups that differ by age, education, and type of appointment[8]. Gender. For women, the transition brought mixed and overall worrisome results. On the one hand, during the transition they were less likely than men – other things equal – to lose a job and become unemployed. But this result seems to be driven by the fact that employed women have been more likely to leave workforce (from both employment and unemployment) than men (Tables V and VI). Moreover, in contrast to pre-transition, women have also been less likely than men to enter workforce (Table VII). Ethnic minorities. There are signs that the position of ethnic minorities has worsened during the transition: they are more likely than Estonians to become unemployed, and their access to jobs has decreased. In contrast to pretransition, the ability of Non-Estonians to switch directly to another job during the transition has been significantly below the ability of Estonians, and their likelihood of losing a job and becoming unemployed is much greater (see Table V). Moreover, in comparison to pre-transition, the chances of NonEstonians of exit from inactivity to jobs have been reduced. Note that the command over the Estonian language significantly improves the chances of exit from inactivity to work, but does not increase the chances switching directly to another job or lessen the chances of losing a job. These results, combined with a strong reduction of a relative wages of Non-Estonians (see Noorkoiv et al., 1998), suggest a strong reduction of demand for Non-Estonian workers. It is conceivable that pressures to reduce employment have been more intense in the predominantly Russian populated North-Eastern region of Estonia, but the above results do not exclude the possibility of discrimination of ethnic minorities. Age groups. Transition brought some disadvantages, but also some advantages for young workers. The results show that young workers are more susceptible to losing a job and becoming unemployed than older workers (Table VI). On the other hand, young workers have an advantage over the old ones in accessing jobs if unemployed. Interestingly, in comparison to pre-transition, workers younger than 20 years faced the increased probability of leaving employment and becoming inactive, perhaps returning to school, as a well as lower probability of leaving inactivity for work. Older workers also share a mixed fortune: in comparison to pre-transition, their ability to switch jobs has been curtailed, but they have been less likely to exit from employment to unemployment – and also to inactivity, except at the end of their work career (see below on the effects of work experience on transitions). It is also notable that the share of younger workers than 20 in the workforce even increased in the early years of transition and later decreased. The share of workers older than 60 fell from 8.3 percent in 1989 to 5.7 in 1997, but the share of 50 to 60 year olds stayed remarkably constant.

Probability of staying in the same job P T Baseline probabilityb

74.3

51.3

Difference in probability associated with: Gender Female 3.7 4.6

Probability of exit Probability of exit Probability of exit to another job to unemployment to inactivity P T P T P T 20.0

25.2

2.0

15.3

3.7

8.1

89 – 8.0

– 6.4

0.0

– 4.0

4.2

5.8

0.0

– 8.7

0.6

13.3

– 0.9

1.2

0.9

3.7

0.7

– 3.8

1.2

– 2.6

– 7.2 5.9 7.2 9.7 13.0 15.1 18.2 22.9 35.6 16.1

1.9 – 1.5 0.0 – 2.0 – 7.4 – 7.8 – 10.5 – 10.0 – 9.9 – 20.0

2.4 1.8 0.0 1.2 0.1 0.9 – 4.1 – 11.4 – 21.8 – 6.3

0.3 1.2 1.8 1.7 1.7 1.6 1.4 1.6 1.5 2.0

– 0.9 – 3.8 – 2.9 – 4.9 – 5.8 – 8.6 – 7.8 – 8.2 – 12.8 – 9.7

Education (primary education excluded) Elementary – 7.2 5.4 Secondary – 3.1 1.4 Special secondary – 1.1 0.3 University – 7.2 0.3

6.3 5.3 3.9 9.1

2.2 8.6 9.8 15.3

1.3 – 0.9 – 1.1 – 0.5

– 6.9 – 8.1 – 7.8 – 12.5

– 4.9 1.3

0.6 – 1.0

0.9 1.1

6.6 – 0.2

Ethnicity and ability to speak Estonian Non-Estonian 0.2 – 5.8 Ability to speak Estonian, if NonEstonian – 2.8 2.7 Age (age 20 to 25 excluded) Under 20 – 6.1 25 to 30 3.8 30 to 35 4.3 35 to 40 6.4 40 to 45 12.1 45 to 50 11.9 50 to 55 12.9 55 to 60 11.3 60 to 65 12.2 Over 65 20.3

Appointment type Non-regular (fixed-term) Part-time

3.7 – 1.8

– 9.2 3.6

Worker reallocation in Estonia

– – – – – – – – –

3.8 1.2 2.4 2.7 3.0 2.5 0.9 0.2 – 0.8 1.7

– – – – – – – – –

5.7 3.9 4.3 5.9 7.3 7.3 6.3 3.4 1.0 0.1

– – – –

– – – –

0.7 2.0 2.3 3.2

– – – – – –

0.4 1.3 1.7 1.4

0.3 – 0.7

2.1 – 2.3

Notes: P = Pre-transition; T = Transition; a Number of observations for 1990: 6110, Log Likelihood = – 2,727.7; Number of observations for 1994: 5510, Log Likelihood = – 4029.3. a Pre-transition refers to the year 1990, and transition to the year 1994. Exit is defined as the first transition occurring after January of the year, but not later than December of the same year. Probabilities associated with coefficients which are significant at the 10 percent level are reported in bold (probabilities of staying in the same employment are derived from other transitions, so significance levels are not reported); b Baseline probability applies for individuals with the following characteristics: Estonian males, age 20 to 25, with primary education (with up to four schooling years), with three to five years of work experience, three Table V. to five years of tenure, not married (single, divorced, or widow/widower), with no children, Probability of exit from with low wages (with wages in the lower third of wage distribution), working in regular, fullemployment, pretime employment in the state sector, domestically owned, large manufacturing firm in the transition, and capital area transition (in percent)a

Education. Perhaps the most consistent and persuasive result of the transition probability analysis is the advantage the reforms brought to the more educated. Before the transition, investment in education not only did not pay

International Journal of Manpower 23,1 90

Probability of staying in unemployment Baseline probabilityb

52.5

Difference in probability associated with: Gender Female – 3.8 Ethnicity Non-Estonian 3.8 Ability to speak Estonian, – 7.7 if Non-Estonian Age Under 20 – 17.8 25 to 30 19.5 30 to 35 1.8 35 to 40 – 5.3 40 to 45 6.3 45 to 50 7.1 50 to 55 5.2 55 to 60 20.6 60 to 65 43.3 Education Elementary ed. 2.7 Secondary ed. 0.4 Special secondary – 6.1 University ed. – 27.7

Probability of exit to employment

Probability of exit to inactivity

36.5

11.0

– 0.3

4.0

– 6.2 7.3

2.4 0.4

8.5 – 15.5 – 1.4 – 8.4 – 12.6 – 17.0 – 23.4 – 24.0 – 32.3

9.3 – 4.1 – 0.4 13.7 6.2 9.9 18.2 3.4 –

2.7 5.7 12.2 34.7

– – – –

5.4 6.1 6.0 7.0

Notes: Number of observations: 534, Log Likelihood = – 475.7. a Pre-transition refers to the year 1990, and transition to the year 1994. Exit is defined as the first transition in the year. Probabilities associated with coefficients which are significant at 10 percent level are reported in bold (probabilities of staying in unemployment are derived from other Table VI. b Probability of exit from transitions, so significance levels are not reported). Baseline probability applies for individuals with the following characteristics: Estonian males, age 20 to 25, with primary unemployment, 1994 education, three to five years of work experience, not married, having no children (in percent)a

high returns, but it also did not bring advantage in terms of easier access to jobs (the exception was a higher probability of the more educated to exit from inactivity to work, see Table VII). During the transition, returns to education greatly increased, and education also brought additional benefits in terms of higher ability to switch from one job to another; to prevent unemployment; and to exit from both unemployment and inactivity to employment. These results nicely complement the results on the increase of returns to education during Estonian transition (see Noorkoiv et al., 1998; Smith, 1999). Type of appointment. Workers with fixed-term appointments are more prone to exit from employment to unemployment, as well as to inactivity (Table V). This indicates that employers make use of fixed-term appointments to screen the applicants. Moreover, those employed part-time show no higher probability of exiting from employment to another job or to unemployment. Interestingly,

Probability of staying Probability of exit to Probability of exit to in inactivity employment unemployment PrePrePretransition Transition transition Transition transition Transition Baseline probabilityb

84.9

72.7

Difference in probability associated with: Gender Female – 9.1 8.9

14.7

22.4

0.5

4.9

91 5.5

– 8.0

3.6

– 0.9

13.1

– 3.7

5.9

1.5

4.4

11.2

5.0

– 0.3

6.6 4.3 – 0.9 9.7 19.5 25.6 27.0 27.2 27.3

4.3 4.0 4.0 – 8.9 – 5.7 – 10.5 – 9.0 – 12.2 – 11.9

– 5.6 – 5.3 1.0 – 8.4 – 17.1 – 22.2 – 22.1 – 22.3 – 22.4

2.4 6.4 – 0.5 40.0 46.4 33.4 – 0.5 3.8 – 0.5

– 1.0 1.0 0.0 – 1.3 – 2.5 – 3.4 – 4.9 – 4.9 – 4.9

Education (primary education excluded) Elementary ed. – 10.7 – 6.7 Secondary ed. – 49.9 – 36.2 Special secondary – 44.4 – 45.7 University ed. – 37.5 – 53.3

7.0 27.3 23.6 26.3

3.5 34.4 44.5 56.5

3.7 22.6 20.8 11.1

3.2 1.9 1.2 – 3.1

Ethnicity and ability to speak Estonian Non-Estonian – 19.0 2.3 Ability to speak Estonian, if NonEstonian – 9.4 – 10.9 Age (age 20 to 25 excluded) Under 20 – 6.7 25 to 30 – 10.4 30 to 35 – 3.5 35 to 40 – 31.1 40 to 45 – 40.7 45 to 50 – 22.9 50 to 55 9.5 55 to 60 8.3 Over 60 12.4

Worker reallocation in Estonia

Notes: Number of observations for 1990: 2091, Log Likelihood = – 775.7; Number of observations for 1994: 2554, Log Likelihood = – 1,274.4. a Pre-transition refers to the year 1990, and transition to the year 1994. Exit is defined as the first transition in the year. Probabilities associated with coefficients which are significant at the 10 percent level are Table VII. reported in bold (probabilities of staying in unemployment are derived from other Probability of exit from transitions, so significance levels are not reported); b Baseline probability applies for inactivity, preindividuals with the following characteristics: Estonian males, age 20 to 25, with primary transition, and education, three to five years of work experience, not married, having no children transition (in percent)a

their propensity to exit to inactivity is lower than for those under full-time appointments, suggesting that their workforce attachment is strong. 4.2 International comparison of determinants of transitions How do the above results on determinants of labor market transition compare with the evidence from other economies? Because of demanding data requirements, there is less evidence on determinants of labor market transitions than on determinants of the wage structure – nonetheless, studies on the Czech

International Journal of Manpower 23,1 92

Republic (Sorm and Terrell, 1999), and Slovenia (Abraham and Vodopivec, 1993) allow the comparison. In many aspects, the Estonian evidence confirms the results from other studies. For example, both the Czech and the Slovenian study also find that more educated workers have a strong advantage as far as avoiding exits from employment to both unemployment and inactivity, and in accessing jobs if a person is inactive or unemployed. Similarly, the present study confirms the results of the above mentioned studies that young workers have better chances to find a job if unemployed than old workers (the Czech study finds that the young are also more likely to move directly from one job to another, and that they face the higher probability the exiting from employment to unemployment, the findings which are not shared by the Slovenian study). As ours, both the Czech and the Slovenian studies also find that women were less mobile than men – they are less likely to switch jobs, less likely to exit from employment to unemployment (in contrast to Estonia, the other studies do not find that women face the higher probability of exit from employment to inactivity). Moreover, the Slovenian study also finds that ethnic minorities witnessed lower chances of finding a job if unemployed, and higher chances to exit from employment to unemployment. The comparison of labor market outcomes of different population groups in Estonia and Slovenia also allows one to examine the prediction of the abovedescribed Blanchard’s model about the labor market outcomes for marginal groups. For some groups of workers – notably, more educated, ethnic minorities, and older workers – the labor market outcome results in the two economies are quite similar, suggesting that common factors were dominant. But outcomes for two marginal groups, young workers and workers under fixed-term contracts, are quite different and thus support the predictions of Blanchard’s (1998) model. Young workers have fared better in Estonia than in Slovenia: the share of those under 20 years in employment increased from 3.1 percent in 1989 to 3.5 percent in 1993, while in Slovenia it fell from 4.2 percent in 1988 to 1.5 percent in 1992[10]. Although the share of the young in Estonia declined after 1993, the overall reduction of the share was much smaller than in Slovenia. Consistent with the above, the share of young workers among the unemployed was smaller in Estonia than in Slovenia. It also seems that a richer bundle of workers rights in Slovenia stimulated the emergence of a ‘‘dual’’ labor market: regular contracts were offered to more valuable workers, and fixed-term contracts to less productive, marginal groups of workers. Indeed, in Slovenia the share of workers employed under fixed-term contracts dramatically increased in the 1990s, covering one-fifth of employment by 1998. In Estonia, in contrast, the share of fixed-term appointments culminated at 5.1 percent in 1995 (Table VIII). Duality of the labor market is underscored also by the fact that workers with fixed-term contracts face a much larger probability of exiting from employment to unemployment in Slovenia than in Estonia. In comparison to a selected baseline individual,

Slovenian fixed-term workers are more than twice as likely to exit to unemployment, and in Estonia, only 43 percent more likely. The above international comparison of employment outcomes of young workers and workers with fixed-term contracts thus supports the hypothesis that the dynamic labor market adjustment of Estonia contributed to more favorable labor market outcomes for marginal groups of workers.

Worker reallocation in Estonia 93

5. Conclusions and policy issues Estonia’s intense and fundamental free market reforms profoundly affected the working of the labor market. The above analysis showed that during Estonia’s transition: . Employment was strongly reduced, and the number of unemployed and inactive individuals increased. The unemployment rate increased from less than 1 percent in 1989 to 10 percent in 1998, and labor force participation was reduced from 69 percent in 1989 to 61 percent in 1998. . Many worker flows considerably increased, particularly out of employment, and flows to and from unemployment emerged that did not exist before the transition. In spite of a worsening of the conditions of the labor market, accessibility of employment as measured by the employment accession rate increased. . Direct job-to-job transitions strongly increased, more than doubling their pre-transition rate during the period of the most intensive restructuring of the economy. Underscoring a strong job creation capacity of the Estonian economy, throughout the transition about half of the workers who lost jobs have been able to transfer to a new job without an intervening period of unemployment. . The more educated emerged as the winners of transition reforms, increasing their advantage over the less educated in accessing jobs and in relative pay, and the Non-Estonians emerged as the group which was most adversely affected by the transition. The relative fate of women and the young as well as old workers was uneven – they improved some of their labor market outcomes and worsened others. . The intensity of labor turnover during Estonian transition has been only slightly below the long-term intensity in mature market economies, and during its most intense period of restructuring, employment accession rates exceeded those in other transition economies.

Estonia Slovenia

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1.7 n.a.

2.0 4.8

2.5 6.5

2.9 7.5

4.0 10.2

5.1 13.6

2.3 15.2

2.5 16.6

2.8 17.5

1.8 19.7

Sources: Same as for Table I; Slovenia: Statistical Office of Slovenia, internal material

Table VIII. Fixed-term employment in Estonia and Slovenia, 1989-98

International Journal of Manpower 23,1 94

.

In comparison with a more ‘‘sclerotic’’ Slovenian labor market, a more fluid Estonian one produced a much more intense labor turnover, larger job destruction and job creation rates, and higher probabilities of exit from unemployment. These differences can be traced, among others, to much more liberal employment protection legislation in Estonia. For the same reasons, the Estonian labor market also produced less segmentation along the permanent vs. fixed-term appointment divide, as well as enabling better access to jobs for young workers, thus allowing the burden of adjustment to be shared more equally by different groups.

The above results show that Estonian reforms generated large worker flows that have facilitated intense labor reallocation across sectors as well as the creation of many productive jobs while simultaneously allowing the destruction of unproductive jobs. They thus confirm the hypothesis that Estonia’s liberal policies in labor and product markets – above all, low layoff costs, relatively low payroll taxation, low minimum wage, and strong encouragement of foreign trade and investment – produced a more fluid and dynamic labor market and contributed to faster reallocation of labor than was the reallocation in other transition economies. These results speak in favor of the radical as opposed to the gradualist approach. Not only have Estonian reforms contributed to a fast reallocation of labor, a more fluid Estonian labor market contributed also to more favorable outcomes from a distributive viewpoint. It produced less segmentation than a more ‘‘sclerotic’’ Slovenian labor market and thus enabled better labor market outcomes for marginal groups (young workers and those employed under the fixed-term appointment). Notes 1. See Commander and Coricelli (1995) for a broad overview. Specific country studies include Jones and Kato (1993) for Bulgaria, Sorm and Terrell (1999) for the Czech Republic, Krueger and Pischke (1995) for East Germany, and Orazem and Vodopivec (2000) for Slovenia. 2. In 5.4 percent of the cases, the recall data indicated labor force participation when the census indicated inactivity. The opposite disagreement occurred in 3.2 percent of the cases. The former cases were concentrated among women in their twenties, and such mismatches are attributable to a change in labor force definition (see Noorkoiv et al., 1998, for details). 3. See the discussion of the macroeconomic performance and institutional background of Estonia during its early transition stage in Haltiwanger and Vodopivec (2002). 4. The analysis complements the studies on job creation/destruction, and on evolution of wage structure and returns to skills presented elsewhere (based on the same data sources, Haltiwanger and Vodopivec (2002) analyze the Estonian job creation and destruction process, and Noorkoiv et al. (1998) analyze changes in wage structure). 5. Note that the scale of employment reduction and/or unemployment increase is not a good proxy for the scale of worker and job reallocation. For a given reduction in employment or increase of unemployment, a flexible labor market is likely to produce a high rate of worker and job reallocation, while relatively inflexible labor market institutions may accommodate only a modest rate of worker and job reallocation. A similar remark applies to the unemployment rate.

6. Comparisons with other transition economies also suggest that Estonian labor market flows have been very intense. In the Czech Republic, for example, probabilities of exit from employment have been much lower than in Estonia, including job-to-job transition (see Sorm and Terrell, 1999). Moreover, unlike the usual procyclical behavior of quits in early Polish transition discussed by Blanchard (1997), quits in Estonia exhibit countercyclical behavior, increasing during the output contraction (their share in separations, however, is procyclical). This points to a strong job creation capacity of the Estonian economy, allowing workers to quit even in worsening labor market conditions (job creation capacity is underscored also by the fact that as many as 45 percent of job-to-job transitions were employer initiated, that is, involved job destructions). 7. It is interesting to note that job flow rates for Estonia are also much higher than those reported for some other non-radical reformers: Bulgaria, Hungary, and Romania (see Bilsen and Konings, 1998). 8. Note that the changes of transition probabilities of a ‘‘baseline’’ person confirm conclusions of aggregate analysis. For him, the probability of exit to unemployment and to inactivity strongly increased, and the probability of staying employed and even inactive markedly decreased (see Tables V and VII). Interestingly, the job-to-job probability also increased. 9. Estonia is one of the rare transition economies where enrolment in secondary education increased during the transition (see UNICEF, 1998). 10. This is consistent with the OECD (1999) finding of a strong link between stricter employment protection legislation and lower employment rates for young workers. References Abraham, K. and Vodopivec, M. (1993), ‘‘Slovenia: a study of labor market transitions’’, The World Bank, Washington, DC. Bilsen, V. and Konings, J. (1998), ‘‘Job creation, job destruction growth of newly established, privatized and state-owned enterprises in transition economies: survey evidence from Bulgaria, Hungary and Romania’’, Journal of Comparative Economics, Vol. 26, pp. 429-45. Blanchard, O. (1997), The Economics of Post-Communist Transition, Oxford University Press, Oxford. Blanchard, O. (1998), ‘‘Employment protection and unemployment’’, paper available at: web.mit.edu/blanchar/www/ Boeri, T. (1998), ‘‘Labor market flows in the midst of structural change’’, in Commander, S. (Ed.), Enterprise Restructuring and Unemployment Models of Transition, EDI, The World Bank, Washington, DC. Bojnec, S. and Konings, J. (1998), ‘‘Job creation, job destruction and labour demand in Slovenia’’, Working Paper 74/1998, Katholieke Universiteit Leuven, Leuven Institute for Central and East European Studies, Leuven. Caballero, R.J. and Hammour, M.L. (2000), Institutions, Restructuring, and Macroeconomic Performance, MIT, Cambridge, MA, p. 39. Commander, S. and Coricelli, F. (1995), Unemployment, Restructuring, and the Labor Market in Eastern Europe and Russia, The World Bank, Washington, DC. Haltiwanger, J.C. and Vodopivec, M. (2002), ‘‘Gross worker and job flows in a transition economy: an analysis of Estonia’’, Labour Economics (forthcoming). Jones, D. and Kato, T. (1993), ‘‘The nature and determinants of labor market transitions in former socialist economies: evidence from Bulgaria’’, Working Paper No. 93/5, Hamilton College. Krueger, A.B. and Pischke, J-S. (1995), ‘‘A comparative analysis of East and West German labor markets: before and after unification’’, in Foseman, R.B. and Katz, L.F. (Eds), Differences and Changes in Wage Structures, University of Chicago Press, Chicago, IL.

Worker reallocation in Estonia 95

International Journal of Manpower 23,1 96

Noorkoiv, R., Orazem, P.F., Puur, A. and Vodopivec, M. (1998), ‘‘Employment and wage dynamics in Estonia, 1989-1995’’, Economics of Transition, Vol. 6 No. 2, pp. 481-503. OECD (1994), The OECD Jobs Study, Paris. OECD (1999), Employment Outlook, Paris. Orazem, P.F. and Vodopivec, M. (2000), ‘‘Male – female differences in labor market outcomes during the early transition to market: the cases of Estonia and Slovenia’’, Journal of Population Economics, Vol. 13, pp. 283-303. Smith, K. (1999), Age/earnings Profiles in Transition Economies: the Estonian Case, Center for European Integration Studies, January. Sorm, V. and Terrell, K. (1999), A Comparative Look at Labor Mobility in the Czech Republic: Where Have All the Workers Gone?, The Davidson Institute Working Paper Series, No. 140, May. Statistical Office of Estonia (1998), Methodological Report on 1997 Estonian Labour Force Survey, Tallinn. UNICEF (1998), Education for All, Regional Monitoring Report No. 5. Appendix: Comparison of transition reforms in Estonia and Slovenia Estonia and Slovenia took sharply differing labor market policy approaches in their initial years of transition. Slovenia took an interventionist approach, with significant barriers to job dislocation, generous support for unemployed workers and pensioners, and efforts to prevent reductions in real wages below a base level of consumption. Estonia took a very liberal approach, with few barriers to labor market dislocations or new job creation, meager support of the unemployed, and no effective wage floor. One major difference between the two countries is in the treatment of workers who became redundant in the transition. In Estonia, there have been modest restrictions on layoffs from the beginning. Layoffs require a two month advance notice and a severance package equal to two to four months of wages, depending on the length of service with the employer. The firm is not liable for other mandated benefits for its fired workers such as job placement or retraining. In fact, during the period studied, unemployment benefits were paid out of general tax revenues rather than experience-rated insurance premiums. All of these policies implied that Estonian firms faced an unusually low marginal cost of layoffs. Layoffs were allowed in Slovenia, but at a large expense to firms. For each dismissal, firms must provide six months advance notice (even 24 months before 1991), and are liable for reassignment, retraining, or early retirement of the fired worker. If none of these options is available, workers are entitled to severance pay of one monthly wage for each year of services with the firm. Clearly these costs serve to reduce firm incentives to initiate layoffs. Once unemployed, Estonia’s policies have encouraged reemployment. The unemployment benefit is very low with benefits averaging about one-tenth of average monthly earnings. Benefits last six months, with an additional three months of benefits possible only if the individual has at least three children and has an income below a poverty threshold. To retain benefits, the unemployed must report every two weeks to the local employment office, they must accept public works jobs, and they may not refuse more than one job offer. Workers for whom no suitable jobs exist are eligible for up to six months of free training, and almost 40 percent undergo such training. In contrast, unemployment insurance in Slovenia is much more generous. Unemployment benefits replace up to 70 percent of previous earnings in Slovenia and benefits can last up to 24 months. Thereafter, unemployed individuals may qualify for means tested unemployment assistance. Individuals may lose benefits if they refuse a job offer or training, but there is no requirement of active job search. The lack of a job search requirement and lax enforcement of the provisions for continuation of unemployment benefits have resulted in relatively little incentive to exit unemployment. For those who were employed in Estonia, there were few distortions in setting wages or the number employed by sector. Minimum wages were imposed, but were so low as to be almost

irrelevant–less than 1 percent of the labor force were paid the minimum wage in 1995. There was no program to subsidize failing firms or to use trade protection to preserve jobs, so growing sectors were not taxed to shore up shrinking sectors. Pensions were very low (the average pension was about one third of the average wage), so the tax burden for funding pensions represented only 5 percent of Estonia’s GDP. In Slovenia, minimum wages were much higher, and the minimum was indexed to inflation at least twice yearly. Consequently, about 10 percent of the employed were at the minimum, suggesting that there was a binding wage floor to hire the least skilled. Pensions were indexed to the growth of average wages on a monthly basis, with the average pension amounting to about 75 percent of the average wage. For many workers, particularly low-skilled workers whose wage increases were below average, retirement was an attractive option. The implicit tax burden for funding the pensions’ 15 percent share of GDP is a serious drain on the Slovenian economy. An additional implicit tax on growing sectors of the economy was the use of subsidies and tariffs to maintain employment in failing sectors. These subsidies represented nearly 1 percent of GDP. Both countries faced sharply changing trade patterns in transition. In Slovenia, the war among the republics of the former Yugoslavia disrupted trading patterns for many sectors. In Estonia, the disruption of trade with the former Soviet Union also created large shifts in the composition of final demand for sectoral outputs. As a consequence, in both countries there were sectors which faced large disruptions in labor demand. At the same time, sectors that were underdeveloped under central planning such as financial services and retail trade might be expected to expand, partially mitigating the adverse effects of the employment problems elsewhere. The extent to which these underdeveloped sectors grew depended upon the existence of capital to finance their expansion, and the potential for profit after taxes. In addition to greater tax burden, Slovenian firms also faced more restricted access to capital. In Estonia, there were no restrictions placed on foreign ownership of former state enterprises or on new foreign investment. As a consequence, there was a tremendous flow of foreign capital into Estonia. By 1995, 9.1 percent of employed Estonians worked for foreign-owned firms and cumulative foreign direct investment was over 5 percent of GDP. In Slovenia, there were large barriers to foreign investment initially, and there are still restrictions on foreign ownership of land and equity. As a result, foreign direct investment in Slovenia lagged behind Estonia, even though the Slovenia transition began two years earlier and the per capita income in Slovenia was much higher than in Estonia. Five years into the Slovenian transition, no sector had greater employment than before the transition began.

Worker reallocation in Estonia 97

International Journal of Manpower 23,1 98

About the authors Lisa Giddings Lisa Giddings is Assistant Professor of Economics at the University of Wisconsin-La Crosse. She earned a PhD from American University in Washington, DC. Prior to joining the Wisconsin system, she taught at Trinity College in DC, and was a research associate at Nathan Associates, Inc. in Arlington, Virginia. Her dissertation focused on both male – female and ethnic earnings differentials in Bulgaria as well as changes in occupational segregation and labor force participation in the early transition. Her current research focuses on the Turkish minority in Germany and labor market effects of East-European immigration to Germany. Susan J. Linz Susan J. Linz is Professor of Economics, Michigan State University; Research Fellow, William Davidson Institute, University of Michigan; and Distinguished University Professor, Taganrog State University for Radio Engineering. She is the author of over three dozen articles and book chapters and editor/contributor to a further three books. Her research is supported by the National Science Foundation, International Research and Exchanges Board, Ford Foundation, and the American Association for the Advancement of Slavic Studies. Her current research focuses on the barriers to investment among Russian firms, as well as the factors the motivate Russians to work. Aleksandra Rogut Aleksandra Rogut is a PhD student at the Institute of Economics, University of L l o´dz´, Poland. She received her MA degree from the University of L l o´dz´, Faculty of Economics in 2000. Her major research interests are the theoretical and empirical analyses of labor markets and the Polish labor market. Tomasz Tokarski Tomasz Tokarski has been Assistant Professor in the Institute of Economics, University of L l o´dz´, Poland since 1991. He received his MA degree from the University of L l o´dz´ in 1991 and his PhD degree in economics from the University of Lodz in 1997. His principal research interests are theoretical aspects of neoclassical and endogenous models of economic growth, technical change and economic growth, and econometric analysis of the Polish labor market.

International Journal of Manpower, Vol. 23 No. 1, 2002, p. 98. # MCB UP Limited, 0143-7720

Milan Vodopivec Milan Vodopivec obtained his PhD in economics from the University of Maryland, College Park. He joined the World Bank in 1989 and is now a Senior Economist. He also served as a State Undersecretary at the Ministry of Labor of Slovenia and was a teacher and dean of a private undergraduate school in Slovenia. He specializes in the analysis of labor markets and cash benefit systems.