Gold Valuation Framework

Table of contents :
Market equilibrium methodology
1. Determine drivers of demand and supply
2. Estimate expected demand and supply
3. Calculate implied equilibrium price
Segmenting demand and supply
Gold Valuation Framework in practice
Appendix I: Accuracy of Model v 1.0
Explanatory power (goodness of fit)
Directional accuracy
Appendix II: Model v 1.0 in detail
Jewellery
Technology
Identifiable investment
Central banks
Implied investment
Mine production
Producer hedging
Recycling

Citation preview

Gold Valuation Framework Your gateway to understanding gold performance January 2020

Gold does not fit within traditional valuation models. Its demand can be boosted by economic growth as well as by uncertainty and, since gold is a global asset, demand tailwinds from one region may counteract headwinds from another. These seeming contradictions pose their own challenges but also give gold its core characteristics as a unique investment. Our research shows that valuing gold is intuitive. In essence, its equilibrium price is determined by the intersection of demand and supply. And understanding the underlying drivers and interactions of gold demand and supply gives investors a robust framework upon which to determine gold’s performance.

Market equilibrium methodology Gold does not conform to common valuation frameworks used for stocks or bonds. Without a coupon or dividend, typical discounted cash flow models fail. And there are no expected earnings or book-to-value ratios either. But there is a good reason why gold does not pay a coupon or a dividend: as a hard currency, it carries no credit risk. Instead, our research shows that gold’s performance can be explained by market equilibrium. In other words, the historical performance of gold is a function of annual demand and supply.

1

Equations (2) and (3) depict changes in demand, supply and price – denoted by lower case d, s and p – instead of levels of demand, supply and price – or D, S and P – as in equation (1) because these are usually non-stationary.

www.gold.org

In mathematical form, this can be expressed as: 𝐷𝐷(𝑃𝑃∗ ) = 𝑆𝑆(𝑃𝑃∗ ),

(1)

where P* is the equilibrium price, D represents the demand function and S the supply. Based on this principle, the implied equilibrium price of gold can be determined in three simple steps.

1. Determine drivers of demand and supply In order to estimate the implied equilibrium price of gold, we first need to determine the relationship that demand and supply have with price (ie price elasticities) and their connection to any other relevant macroeconomic drivers. In particular, we express demand as a linear function of both price and other relevant macroeconomic variables. Namely,

and

𝑑𝑑𝑡𝑡 (𝑝𝑝) ∶ 𝑑𝑑𝑡𝑡 = 𝛼𝛼 + 𝛽𝛽𝑝𝑝𝑡𝑡 + 𝛾𝛾𝑋𝑋𝑡𝑡 + 𝜀𝜀𝑡𝑡 ,

𝑠𝑠𝑡𝑡 (𝑝𝑝) ∶ 𝑠𝑠𝑡𝑡 = 𝛼𝛼′ + 𝛽𝛽′𝑝𝑝𝑡𝑡 + 𝛾𝛾′𝑋𝑋′𝑡𝑡 + 𝜀𝜀′𝑡𝑡 ,

(2) (3)

where 𝑑𝑑𝑡𝑡 and 𝑠𝑠𝑡𝑡 represent changes in demand, 𝑝𝑝𝑡𝑡 changes in price, while 𝑋𝑋𝑡𝑡 and 𝑋𝑋′𝑡𝑡 represent a set of relevant explanatory variables, usually macroeconomic in nature, such changes in GDP, credit spreads or interest rates. 1,2

2. Estimate expected demand and supply The second step determines expected demand and supply based solely on information about relevant macroeconomic variables, assuming there is no change in the price of gold. In other words, to estimate changes in demand and supply at a given time T, assuming 𝑝𝑝𝑇𝑇 = 0, we solve: and

𝑑𝑑̂𝑇𝑇 (0) ∶ 𝑑𝑑̂𝑇𝑇 = 𝛼𝛼� + 𝛾𝛾�𝑋𝑋𝑇𝑇 ,

� + 𝛾𝛾′ � 𝑋𝑋′ 𝑇𝑇 , 𝑠𝑠̂ 𝑇𝑇 (0) ∶ 𝑠𝑠̂ 𝑇𝑇 = 𝛼𝛼′

(4) (5)

However, vector error-correction model (VECM) specifications could be used to model levels and changes jointly. 2

Note that Xt and X’t may not represent the same set of variables. 01

where •� represents estimated variables, including demand and supply, as well as any relevant coefficients determined by equations (2) and (3). For example, for any given year – past, present or future – this step estimates the demand and supply using information about the macroeconomy, but without taking into account the impact that changes in price may have on the behaviour of buyers (eg, consumers) and sellers (eg, producers).

3. Calculate implied equilibrium price Of course, the behaviour of buyers and sellers is influenced by price. And, generally, estimating demand and supply solely from macroeconomic variables will result in a market imbalance, where demand will be greater than the available supply or vice versa. The third and final step calculates the implied change in price necessary to bring the market back to equilibrium. In particular, it requires to find 𝑃𝑃� (or, equivalently, 𝑝𝑝̂ ) such that the level of demand equals the level of supply, in a given period. Namely, �𝑇𝑇 �𝑃𝑃�� − 𝑆𝑆̂𝑇𝑇 �𝑃𝑃�� = 0, 𝐷𝐷

(6)

�𝑇𝑇 �𝑃𝑃�� = 𝐷𝐷𝑇𝑇−1 (𝑃𝑃) + 𝑑𝑑̂ 𝑇𝑇 (𝑝𝑝̂ ), 𝑆𝑆̂𝑇𝑇 �𝑃𝑃�� = 𝑆𝑆𝑇𝑇−1 (𝑃𝑃) + where 𝐷𝐷 𝑠𝑠̂ 𝑇𝑇 (𝑝𝑝̂ ), and 𝑃𝑃�𝑇𝑇 = 𝑃𝑃𝑇𝑇−1 + 𝑝𝑝̂ 𝑇𝑇 . 3 Note that 𝑃𝑃� can be calculated using iterative numerical methods such as NewtonRaphson or other robust numerical alternatives. Segmenting demand and supply In many instances, segmentation may be required to appropriately capture the impact of diverse market participants. For example, in the case of gold, the behaviour of consumer and investors may be driven by different factors. Similarly, drivers of mine production may differ from those of gold recycling. In this context, equations (2) and (3) can be further split into a set of multiple sub-equations, each representing a distinct segment of demand and supply. Namely,

and

𝑑𝑑𝑖𝑖,𝑡𝑡 = 𝛼𝛼𝑖𝑖 + 𝛽𝛽𝑖𝑖 𝑝𝑝𝑖𝑖,𝑡𝑡 + 𝛾𝛾𝑖𝑖 𝑋𝑋𝑖𝑖,𝑡𝑡 + 𝜀𝜀𝑖𝑖,𝑡𝑡 ,

𝑠𝑠𝑗𝑗,𝑡𝑡 = 𝛼𝛼′𝑗𝑗 + 𝛽𝛽′𝑗𝑗 𝑝𝑝𝑗𝑗,𝑡𝑡 + 𝛾𝛾′𝑗𝑗 𝑋𝑋′𝑗𝑗,𝑡𝑡 + 𝜀𝜀′𝑗𝑗,𝑡𝑡 ,

(7) (8)

for 𝑖𝑖 = 1, … , 𝑚𝑚 and 𝑗𝑗 = 1, … , 𝑚𝑚′, where i and j represent each relevant sector of demand and supply respectively.

Gold Valuation Framework in practice It is often said that positive economic growth is bad for gold. But evidence suggests the opposite. The combined share of world gold demand from India and China grew from 25% in the early 1990s to more than 50% by 2015. Gold investment demand can, however, over the short and medium term, exert strong pressure on prices. This type of demand from physical markets, exchange-traded securities and over-the-counter (OTC) products increases during periods of economic and political uncertainty and falls as investor confidence grows. Central bank demand can influence the market too. And both economic expansion and uncertainty can be catalysts for central banks to add gold to their reserves. In addition, gold is scarce. Its availability is influenced by two factors: whether mines have the capacity to produce new gold and whether investors or consumers are willing to sell their gold holdings. We apply the proposed Gold Valuation Framework, based on market equilibrium, to key segments that appropriately capture gold’s dual nature as a consumer good and as an investment, as well as those that explain the different elements of gold supply. While demand through OTC markets is one of those key sectors, no direct estimates of this data exist. As such, we use instead information on COMEX net long positioning as a proxy to estimate implied gold investment. In particular, we split demand into five sectors: jewellery, technology, identifiable investment (bar, coin and ETFs), implied investment, and central banks. And supply into three sectors: mine production, hedging and recycling. 4 Our model (version 1.0), uses annual data going back to 1980 due to data availability. 5 We summarise its accuracy in Appendix I, while Appendix II provides a detailed description. There, we list the variables used to explain each of the relevant segments of demand and supply, their respective coefficients (elasticities), statistical significance, and likely economic interpretation.

Consequently, equation (6) becomes: � 3 4

𝑚𝑚

𝑖𝑖=1

�𝑖𝑖,𝑇𝑇 �𝑃𝑃�� − � 𝐷𝐷

𝑚𝑚′

𝑆𝑆̂𝑗𝑗,𝑇𝑇 �𝑃𝑃�� = 0.

𝑗𝑗=1

(9)

Letters with a tilde (•�), represent estimated variables, and letters without a tilde represent observed (past) variables. This methodology can be applied equally to a different set of relevant sectors of demand and supply if, when combined, they represent actual or estimated values for the total amount of demand and supply.

Gold Valuation Framework | Your gateway to understanding gold performance

5

The Gold Valuation Framework can, in principle, be used to analyse quarterly demand and supply data. At present, only quarterly data going back to 2000 is available; this creates potential limitations which we may explore at a future date.

02

Appendix I: Accuracy of Model v 1.0 The accuracy of Model 1.0, as detailed in Appendix II, can be demonstrated in two ways. 6 One is to measure the percentage of the variation in the actual gold price, or gold returns, explained by the model. The second is to assess the proportion of observations where the model correctly estimates the direction of the gold price, or gold returns.

Explanatory power (goodness of fit) The percentage of the variation in the price of gold explained by the model is 97% 7 (Chart 1). The model also explains 73% of the variation in gold price changes (ie, returns) as show in Chart 2. Directional accuracy Another way to assess the explanatory power of the model is to look at how good it is at accurately explaining the direction of movements in the gold price (returns). Using this metric, the model estimates correctly the direction of returns 73% of between 1980 and 2018.

Chart 1: Historical gold price vs the price implied by Gold Valuation Framework based on Model 1.0*

Chart 2: Historical gold returns vs returns implied by Gold Valuation Framework based on Model 1.0*

USD/oz 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 0 1980

USD/oz, YoY 500 400 300 200 100 0 -100 -200 -300 -400

1985 1990 1995 2000 2005 Modelled gold price, USD/oz Actual gold price, USD/oz

2010

2015

*As of December 2018. Based on annual averages of the LBMA Gold Price PM. Source: Bloomberg, ICE Benchmark Administration, World Gold Council

1980

1985 1990 1995 2000 2005 Modelled gold price, USD/oz, YoY Actual gold price, USD/oz, YoY

2010

2015

*As of December 2018. Based on year-on-year changes of the annual average of the LBMA Gold Price PM. Source: Bloomberg, ICE Benchmark Administration, World Gold Council

6

We currently use annual data going back to 1980 due to data availability.

7

Measured as the Explained Sum of Squares (ESS) [∑(𝑌𝑌�𝑖𝑖 − 𝑌𝑌� )2] divided by the Total Sum of Squares (TSS), where the TSS = ESS + RSS (Residual Sum

Gold Valuation Framework | Your gateway to understanding gold performance

of Squares, ∑(𝑌𝑌𝑖𝑖 − 𝑌𝑌�𝑖𝑖 )2 ). This measure is equivalent to R2 in traditional regression models.

03

Appendix II: Model v 1.0 in detail

Technology We modelled technology demand using two variables: developed market (DM) import growth and relative emerging market (EM) versus DM GDP growth.

I. Demand

Demand for gold in technology has historically been driven almost exclusively by the business cycle, with electronics being the dominant category.

Jewellery We modelled jewellery demand as a function of concurrent and lagged gold price, plus world GDP growth. Our analysis indicates that, historically, higher prices have curtailed jewellery demand and vice versa. But a positive coefficient for the lagged price suggests consumers increased purchases when the price trend was positive. We believe that this coefficient may also reflect the ‘investment’ characteristics of high karat gold, for example, in China and India. Finally, the analysis suggests that gold jewellery demand behaves like a normal good, increasing as incomes grows, as given by changes in GDP. Table 1: Jewellery, % y-o-y Jewellery Adj. R-squared:

0.48

No. Observations:

37

Durbin-Watson:

DM import growth reflects the indirect demand for electronics developed in EM Asia. 8 EM relative GDP growth reflects the expansion of productive capacity in emerging economies relative to advanced economies, which has driven fabrication demand for gold. Table 2: Technology, % y-o-y Technology Adj. R-squared:

0.52

No. Observations:

36

Durbin-Watson:

1.77 Coefficient

Std. Error

T-stat

p-value

Constant

-6.3

1.4

-4.6

0.000

DM imports, % y-o-y

1.2

0.2

6.3

0.000

EM vs DM GDP growth (-1)

0.3

0.1

1.8

0.075

1.65 Coefficient

Std. Error

T-stat

p-value

Constant

4.1

1.4

2.9

0.007

Gold price, % y-o-y

-0.8

0.1

-6.0

0.000

Gold price, % y-o-y (-1)

0.5

0.1

3.8

0.001

World GDP, % y-o-y

0.7

0.2

2.9

0.006

Chart 3: Jewellery, % y-o-y %, YoY 50 40 30 20 10

Chart 4: Technology, % y-o-y %, YoY 25.0 20.0 15.0 10.0 5.0 0.0 -5.0 -10.0 -15.0 -20.0 -25.0 1980

1985 1990 Residuals

1995 2000 2005 2010 Actual Implied

2015

0 Based on annual changes between 1983 and 2018.

-10

Source: Bloomberg, ICE Benchmark Administration, World Gold Council

-20 -30 1980

1985 1990 Residuals

1995 2000 Actual

2005 2010 Implied

2015

Based on annual changes between 1982 and 2018. Source: Bloomberg, ICE Benchmark Administration, World Gold Council

8

Between 2015 and 2017 almost 45% of total imports from China into the US were electronic products.

Gold Valuation Framework | Your gateway to understanding gold performance

04

Identifiable investment We capture changes in identifiable investment demand, which include bars, coins and exchange-traded funds, by using concurrent and lagged gold returns, relative yields of high- and low-quality corporate bonds, and Chinese and Eurozone money supply growth. Our analysis suggests that identifiable gold investment demand has historically being affected by both long- and short-run dynamics. 9 For example, US corporate credit spreads are often viewed as bellwethers of economic and financial health. Higher spreads have historically corresponded with periods when investors allocate to safer assets, such as government bonds or gold. Historically, changes in identifiable investment demand have also shown a negative correlation to changes in the gold price, like that seen in jewellery demand. There is also a positive correlation to the lagged price suggests that a positive price trend is supportive of investment. Especially, since the coefficient for the lagged price return is larger than the coefficient for the concurrent price return. Our model also contains lagged money supply variables for the Eurozone and for China – two key demand centres. Healthy money supply growth is a reflection of solid economic expansion. However, money in excess of what the economy needs is often viewed as inflationary. Either scenario captures the motives that may drive investors towards gold: inflationary concerns in the short term, but income in the long term. Table 3: Identifiable investment, % y-o-y Identifiable investment Centered R-squared: 10 No. Observations: Durbin-Watson: China M1, % y-o-y (-1) Baa and Aaa credit yield ratio Eurozone M1, % y-o-y (-1)

0.40

%, YoY 100 80 60 40 20 0 -20 -40 -60 -80 -100 1980

1985 1990 Residuals

1995 2000 Actual

2015

Based on annual changes between 1983 and 2018. Source: Bloomberg, ICE Benchmark Administration, World Gold Council

Central banks Our analysis suggests that changes in global savings have historically significantly explained changes in central bank (CB) demand, despite a low R-squared from a rather parsimonious model. High savings rates are a characteristic shared by all major reserve accumulators. 11 This is particularly true of China and India. Higher savings rates and reserve accumulation are partly a result of export-led growth, protection against financial crises (especially following the 1998-1999 Asian Financial Crisis). Equally, dis-savings in DM where consumer spending has been financed by borrowing has been a mainstay feature of this phenomenon.

3.1

Table 4: Central banks, tonnes y-o-y

Coefficient

Std. Error

T-stat

p-value

1.6

0.7

2.3

0.028

Central banks

369.1

101.1

3.7

0.001

Centered R-squared:

0.005

No. Observations:

37

Durbin-Watson:

1.6

6.2

2005 2010 Implied

All variables are lagged consistent with the cautious and non-reactive nature of CB accumulation

36

2.0

3.1

Gold price, % y-o-y

-0.6

0.4

-1.4

0.161

Gold price, % y-o-y (-1)

1.2

0.4

2.8

0.009

9

Chart 5: Identifiable investment, % y-o-y

Financial inclusion and long-term savings are two factors our model doesn’t explicitly include. In the long-run and in aggregate, investors in physical gold allocate some of their income and/or wealth to gold. How much depends on where: EM investors invest more per capita than AE investors. Level of income/wealth: AE investors have greater average income/wealth with which to invest. Opportunity cost: Are there better alternatives to direct investors’ savings? And risk aversion: The higher the level of risk aversion the higher the level of gold allocation.

Gold Valuation Framework | Your gateway to understanding gold performance

0.23

Coefficient

Std. Error

T-stat

p-value

Central banks, y-o-y (-1)

-0.2

0.2

-1.2

0.230

EM savings (-1)

94.6

32.5

2.9

0.006

DM savings (-1)

-66.2

46.8

-1.4

0.167

10 We use a centered R-squared instead of an adjusted R-square for models that do not include a constant as it has been shown to more accurately reflect goodness of fit. See Woolridge, Introductory Econometrics: A Modern Approach, 2009. 11 https://www.ecb.europa.eu/pub/pdf/scpops/ecbocp43.pdf

05

Table 5: Implied investment long proxy, % y-o-y OTC long proxy

Chart 6: Central banks, tonnes y-o-y

Adj. R-squared:

Tonnes, YoY 600 400 200

0.37

No. Observations:

31

Durbin-Watson:

2.2 Coefficient

Std. Error

T-stat

p-value

149.9

61.6

2.4

0.020

Gold price, % y-o-y

0.6

0.7

0.8

0.430

Gold price, % y-o-y (-1)

-0.9

0.7

-1.3

0.201

OTC long proxy, % y-o-y (-1)

-0.5

0.2

-3.6

0.002

-145.3

62.3

-2.3

0.028

Constant

0 -200 -400

US 10y yield, %

-600 -800 1980

1985 1990 Residuals

1995 2000 Actual

2005 2010 Implied

2015

Based on annual changes between 1982 and 2018. Source: Bloomberg, ICE Benchmark Administration, World Gold Council

Chart 7: Implied investment long proxy, % y-o-y %, YoY 150.0 100.0 50.0

Implied investment We modelled implied investment, reflecting demand through OTC transactions, by using data on positioning in the COMEX futures market as a proxy. The rationale here is that we believe the motives behind OTC investment demand are very similar to those behind gold futures. In many cases they may be simply different vehicles with which investors can express a view about gold; both are subject to potential leverage as well as roll-costs from inherent curve structures. We also split the COMEX positioning data in two: long positioning and short positioning. Our historical analysis suggests that changes in OTC, based on our long proxy, can be explained by concurrent and lagged returns, lagged changes to OTC longs, as well as the level of the US 10-year government bond yield. The 10year yield captures both opportunity costs among investors and their expectations for growth and inflation. The OTC short proxy has historically responded to concurrent and lagged gold returns, lagged changes in OTC shorts and lagged changes to credit spreads. For example, a narrowing of credit spreads in a previous period historically reduced short flows in the concurrent period.

Gold Valuation Framework | Your gateway to understanding gold performance

0.0 -50.0 -100.0 -150.0 1980

1985 1990 Residuals

1995 2000 2005 2010 Actual Implied

2015

Based on annual changes between 1988 and 2018. Source: Bloomberg, ICE Benchmark Administration, World Gold Council

Table 6: Implied investment short proxy, % y-o-y OTC short proxy Centered R-squared:*

0.39

No. Observations:

31

Durbin-Watson:

1.9 Coefficient

Std. Error

T-stat

p-value

Gold price, % y-o-y

-1.0

0.5

-2.1

0.042

Gold price, % y-o-y (-1)

1.1

0.5

2.4

0.025

OTC short proxy, % y-o-y (-1)

-0.4

0.1

-2.8

0.008

Baa/Aaa spread (-1)

-0.4

0.2

-2.2

0.037

06

Chart 8: Implied investment short proxy, % y-o-y %, YoY 80 60

Credit spreads have captured a potentially higher credit risk environment. And pressures on medium-term costs can be explained by lagged credit spreads and DM GDP growth. Table 7: Mine production, % y-o-y

40

Mine production

20

Adj. R-squared:

0 -20

0.38

No. Observations:

37

Durbin-Watson:

2.0

-40

Coefficient

Std. Error

T-stat

p-value

-60

Constant

-2.9

1.2

-2.4

0.023

-80

Baa/Aaa spread (-1)

3.8

1.0

3.8

0.001

4y mining trend

0.6

0.2

3.0

0.005

DM GDP, % y-o-y

-0.1

0.1

-2.0

0.057

1980

1985 1990 Residuals

1995 2000 Actual

2005 2010 Implied

2015

Based on annual changes between 1988 and 2018. Source: Bloomberg, ICE Benchmark Administration, World Gold Council

Chart 9: Mine production, % y-o-y

II. Supply Mine production We modelled annual changes in mine production as a function of lagged US credit spreads, DM GDP growth and a mine production trend. Our analysis indicates that changes in mine production have historically been explained by an interaction of longterm trends and expectations with some medium-term manoeuvrability. Long term factors determining output are: • new mine development cycles which are both long (three

to seven years) and costly • historically falling average ore grades, raising marginal

costs • a low number of new discoveries limiting expansion • increasing regulatory hurdles and that sustain price

increases and/or longer-term expectations of higher prices. The analysis suggests that these long-term considerations can be historically captured implicitly by a four-year lagged mine production trend. In the medium term large capital-intensive mines may have the ability to make tactical shifts output. These shifts have historically occurred in response to pressures to production costs: higher labour/energy/financing costs and/or lower fabrication demand.

Gold Valuation Framework | Your gateway to understanding gold performance

% YoY 12.0 10.0 8.0 6.0 4.0 2.0 0.0 -2.0 -4.0 -6.0 -8.0 1980

1985 1990 Residuals

1995 2000 Actual

2005 2010 Implied

2015

Based on annual changes between 1982 and 2018. Source: Bloomberg, ICE Benchmark Administration, World Gold Council

Producer hedging We modelled hedging as a function of a lagged US yield curve term and lagged hedging. Our analysis suggests that, historically, the yield curve variable represents a number of channels through which hedging decisions are impacted: • A steeper yield curve is consistent with economic

expansions, and a weaker gold price via a higher opportunity cost. This impacts miners’ profit margins and thereby a propensity to hedge. • A steeper yield curve reflects funding costs for gold miners, making gold loans - a component of hedging more attractive • A steeper yield curve is often associated with a flatter gold price curve. A flatter curve suggests at times that spot prices are ‘rich’ relative to futures prices and attractive to ‘lock in’ through forward sales, swaps or loans.

07

purchases (for example, the coefficients for the models of jewellery and recycling are similar but their tonnage effect is different).

Table 8: Producer hedging, tonnes y-o-y Producer hedging Adj. R-squared:

0.51

No. Observations:

34

Durbin-Watson:

1.6 Coefficient

Std. Error

T-stat

p-value

-173.2

55.8

-3.1

0.004

Hedging, y-o-y (-1)

-0.7

0.1

-5.6

0.000

Recycling

US yield curve (-1)

91.4

26.4

3.5

0.002

Centered R-squared:*

Constant

Table 9: Recycling, % y-o-y 0.47

No. Observations:

37

Durbin-Watson:

2.4 Coefficient

Std. Error

T-stat

p-value

Gold price, y-o-y

1.0

0.2

4.4

0.000

Gold price, y-o-y (-1)

-0.7

0.2

-3.3

0.002

EDA GDP growth

-1.6

0.4

-4.4

0.000

Chart 10: Producer hedging, tonnes y-o-y Tonnes, YoY 600 400

Chart 11: Recycling, % y-o-y

200

%, YoY 80.0 60.0

0

40.0

-200

20.0

-400

0.0 -20.0

-600 1980

1985 1990 Residuals

1995 2000 Actual

2005 2010 Implied

2015

Based on annual changes between 1985 and 2018. Source: Bloomberg, ICE Benchmark Administration, World Gold Council

Recycling We modelled annual changes in recycling as a function of concurrent and lagged gold returns as well as GDP growth for emerging and developing Asia.

-40.0 -60.0 -80.0 1980

1985 1990 Residuals

1995 2000 2005 2010 Actual Implied

2015

Based on annual changes between 1982 and 2018. Source: Bloomberg, ICE Benchmark Administration, World Gold Council

Our analysis suggests that concurrent change to the gold price have historically had a positive correlation to recycling while lagged changes to price have had a negative one. We believe that the rationale for this is that: • recycling levels adjust to prices as near-market supplies

get exhausted • recycling may be a function of the trend in the gold price.

We also find that recycling and jewellery demand are near opposites in terms of their drivers. But there are a few key differences we can identify: • recycling is not always recorded in the country where

demand occurred, and it may contain non-jewellery items • recycling from fabricators or consumers may reflects

different motivations • there seems to be an asymmetry between the

motivations to recycle relative to those for jewellery Gold Valuation Framework | Your gateway to understanding gold performance

08

About the World Gold Council

For more information

The World Gold Council is the market development organisation for the gold industry. Our purpose is to stimulate and sustain demand for gold, provide industry leadership, and be the global authority on the gold market.

Market Intelligence and Research:

We develop gold-backed solutions, services and products, based on authoritative market insight, and we work with a range of partners to put our ideas into action. As a result, we create structural shifts in demand for gold across key market sectors. We provide insights into the international gold markets, helping people to understand the wealth preservation qualities of gold and its role in meeting the social and environmental needs of society. Based in the UK, with operations in India, the Far East and the US, the World Gold Council is an association whose members comprise the world’s leading gold mining companies. World Gold Council 10 Old Bailey, London EC4M 7NG United Kingdom T +44 20 7826 4700 F +44 20 7826 4799 W www.gold.org

Juan Carlos Artigas Director, Investment Research [email protected] +1 212 317 3826 Alistair Hewitt Director, Market Intelligence [email protected] +44 20 7826 4741 John Reade Chief Market Strategist [email protected] +44 20 7826 4760 Distribution and Investment: Jaspar Crawley Director, EMEA [email protected] +44 20 7826 4787 Matthew Mark Director, North America [email protected] +1 212 317 3834 Fred Yang Director, China [email protected] +86 21 2226 1109

Copyright and other rights © 2020 World Gold Council. All rights reserved. World Gold Council and the Circle device are trademarks of the World Gold Council or its affiliates. Any references to LBMA Gold Price are used with the permission of ICE Benchmark Administration Limited and have been provided for informational purposes only. ICE Benchmark Administration Limited accepts no liability or responsibility for the accuracy of the prices or the underlying product to which the prices may be referenced. All third-party content is the intellectual property of the respective third party and all rights are reserved to such party. Reproduction or redistribution of any of this information is expressly prohibited without the prior written consent of World Gold Council or the appropriate intellectual property owners, except as specifically provided below. Use of any statistics in this information is permitted for the purposes of review and commentary in line with fair industry practice, subject to the following preconditions: (i) only limited extracts may be used; and (ii) any use must be accompanied by a citation to World Gold Council and, where appropriate, to Metals Focus, Refinitiv GFMS, or other identified third party, as their source. World Gold Council does not guarantee the accuracy or completeness of any information and does not accept responsibility for any losses or damages arising directly or indirectly from the use of this information.

Information regarding Qaurum and the Gold Valuation Framework. Note that the resulting performance of various investment outcomes that can generated through use of Qaurum, the Gold Valuation Framework tool, and other information are hypothetical in nature, may not reflect actual investment results and are not guarantees of future results. Diversification does not guarantee investment returns and does not eliminate the risk of loss. World Gold Council and its affiliates and subsidiaries (collectively, “WGC”) provide no warranty or guarantee regarding the functionality of the tool, including without limitation any projections, estimates or calculations. WGC does not accept responsibility for any losses or damages arising directly or indirectly from the use of the tool or any information provided herein. WGC does not guarantee the accuracy or completeness of any information provided herein. The sole purpose of this information, including the tool, is educational in nature. By receiving this information, you agree with this intended purpose. This information does not take into account any investment objectives, financial situation, or particular needs of any person. Nothing contained herein is, or is intended to constitute, any recommendation, investment advice or offer for the purchase or sale or acquisition of or solicitation of investment in gold, any goldrelated products or services or any other products, services, securities or financial instruments.

This information is not a recommendation or an offer for the purchase or sale of gold or any products, services, or securities. This information contains forward-looking statements which are based on current expectations and are subject to change. Forward-looking statements involve a number of risks and uncertainties. There is no assurance that any forward-looking statements will be achieved.

Gold Valuation Framework | Your gateway to understanding gold performance

09