Marcus Grubb

Managing Director,

Investment Strategy

In the years that followed the 2008-2009 financial crisis, many investors turned to gold as a store of wealth amid challenging economic conditions and heightened uncertainty. Today, particularly in the US, the economic recovery is gaining momentum and common wisdom suggests that a better economic environment would be bad for gold. Higher interest rates increase the opportunity cost of holding gold and drive away investors. However, this oversimplification misses four key points:

•

The long-term performance of the gold market is not solely linked to investment demand in developed markets or the US alone

•

There is a strong correlation between economic growth and gold consumption (as a luxury good and as integral component in high-end electronics)

•

Maintaining a proper portfolio risk management strategy is as important in good times as it is in bad

•

The gold market has undergone several structural shifts that have increased the demand base.

In Gold Investor, Volume 7, we explore these issues in greater detail.

Juan Carlos Artigas

Director,

Investment Research

Not all gold demand is counter-cyclical

First, we explore the connection between economic growth and gold through our analysis and third-party research in The growth dividend: how rising GDP lifts gold consumer demand.

Investors understand that in times of economic duress high-quality, liquid assets such as gold are in high demand. What is less understood is that the majority of gold demand is linked to consumption and long-term savings and not to speculative investment. As an economy expands, incomes grow and gold demand increases – counter short-term investment flows. In turn, this pro-cyclical nature of consumer demand and counter-cyclical nature of investment demand make gold an effective diversifier and valuable portfolio component.

Tail-risk hedging should be a multi-environment affair

In A practical hedge: less exotic, multipurpose, lower cost, we make the case for gold as a valuable component of any portfolio risk-hedging strategy.

The current low volatility environment has made many portfolio hedges “cheap” to buy, but that doesn’t mean that they are inexpensive to hold. We analyse and compare multiple investment vehicles commonly used as tail-risk hedges. Our research finds a compelling case for gold as a valuable hedge. It is cost-effective, easy to implement, and has a wide range of applications – even if its return during tail events is often lower than bespoke (and often derivative-based) tail risk-mitigating strategies.

The market has changed and more investors can gain access to gold more easily

Finally, we look at one of the most revolutionary structural shifts the gold market has experienced in Ten years of gold ETFs: a wider and more efficient market.

Over the past decade, gold-backed ETFs have become a US$70 billion dollar market – with holders of all sizes and types spread across various geographies. Yet, the advent of ETFs has benefitted other gold investors as well: they have increased demand, democratised access, increased competition for new and existing gold vehicles, and contrary to popular belief they have responded to rather than driven gold volatility higher.

We hope you enjoy this edition of Gold Investor as much as we enjoyed writing it. If you have comments or questions on any of the articles you may reach us at investmentresearch@gold.org.

Consumer demand for gold closely tracks growth

The gold market has three main market segments: consumers (jewellery and technology), investors (bars, coins and gold-backed financial products) and central banks. The issues of their relative weightings, effects and influence on gold are complex and will, we expect, be subject to discussion for a long time.

A lot has been said about the link between recessions and gold, generally that investors look for a store of wealth and thereby lift demand for gold. Much less has been written about the effect that positive economic growth has on consumer demand, despite the fact that this makes up the lion’s share of physical gold demand (Chart 1).

We have turned to our own research as well as a third-party econometric study to form our view on this topic. The research has provided a striking reminder of what we see as gold’s positive duality: its ability to benefit from both the contraction and expansion phases of the business cycle.

Chart 1: Consumer demand provides long-term support to the gold market

Five-year average consumer and investment demand ending 2013

Reference notes are listed at the end of this article.

Source: GFMS-Thomson Reuters, respective ETF and ETP providers, World Gold Council

The health of the economy has come a long way since the financial crisis – but the outlook across regions varies

Europe’s outlook is less rosy

Emerging markets still have plenty of room to grow

Our research and analysis shows that gold is:

A recap of gold’s consumer, investment, and geographic demand factors

As we see it, gold is a good diversifier as a by-product of its sources of demand and supply, their interactions, and the factors that influence them. In fact, while investment demand often gets a lot of attention – especially in the West – consumer demand is greater by some margin. In the last five years average demand breaks down as follows (Chart 3a):

•

Gold demand is also geographically diverse (Chart 3b). This geographic spread, along with the oppositely correlated demand streams, helps to smooth out gold’s demand curve. For example, in 2010-2011 investment demand was quite strong in the West while consumer demand was flat; in contrast, consumer demand in Asia outstripped investment demand as the region quickly recovered after the 2008-2009 financial crisis.

In addition, Asian economies typically devote more money to savings (the real average saving rate is around 30% and can reach levels greater than 50% of disposable income), supporting a more robust long-term demand trend.⁴ Indeed, gold is a natural vehicle for savings in many Asian economies: from direct gold purchases to innovative banking solutions such as gold-accumulation plans.⁵

Reference notes are listed at the end of this article.

Source: Avinash Persaud, World Gold Council

References

Chart 1: Consumer demand provides long-term support to the gold market

Average demand by region; it includes jewellery, technology, bars, coins and ETFs and excludes over-the-counter and central bank purchases. ETF demand assigned to the country of where the fund is listed. See reference notes for Chart 2 for a complete list of ETFs used for estimating full tonnage.

Chart 2: Investment demand through ETFs and similar products increased during the financial crisis and came down in 2013

Cumulative demand since 2003 from gold-backed ETFs, gold ETNs, and similar products that we periodically monitor. Namely: SPDR Gold Shares, iShares Gold Trust, Central Fund of Canada Ltd, Sprott Physical Gold Trust, ETFS Gold Trust, Central GoldTrust, Royal Canadian Mint - Canadian Gold Reserves, iShares Gold Bullion ETF, ETFS Physical Precious Metal Basket Shares, Merk Gold Trust, ETFS Asian Gold Trust, ZKB Gold ETF, ETFS Physical Gold/Jersey, Gold Bullion Securities Ltd, Julius Baer Precious Metals Fund - JB Physical Gold Fund, Xetra-Gold, Source Physical Gold P-ETC, iShares Gold CH, db Physical Gold Euro Hedged ETC, UBS ETF CH-Gold CHF hedged CHF, db Physical Gold ETC, db Physical Gold ETC EUR, Pictet CH Precious Metals Fund - Physical Gold, iShares Gold CHF Hedged CH, ETFS Physical Swiss Gold, iShares Physical Gold ETC, UBS ETF CH-Gold EUR hedged EUR A-dis, EUWAX Gold, Credit Suisse Institutional Fund II - Gold Blue, Raiffeisen ETF - Solid Gold Ounces - AC, iShares Gold EUR Hedged CH, ETFS Physical PM Basket, Raiffeisen ETF - Solid Gold, db Physical Gold GBP Hedged ETC, ETFS EUR Daily Hedged Physical Gold, FinEx Physically Held Gold ETF, ETFS GBP Daily Hedged Physical Gold, Goldman Sachs Gold Exchange Traded Scheme - GS Gold BeES, R*Shares Gold Exchange Traded Fund, Mitsubishi UFJ Japan Physical Gold ETF, SBI Gold Exchange Traded Scheme, Kotak Gold ETF, HDFC Gold Exchange Traded Fund, ValueGold ETF, UTI-Gold Exchange Traded Fund, AXIS GOLD ETF, Canara Robeco Gold Exchange Traded Fund, ICICI Prudential Gold ETF, IDBI Gold ETF, Birla Sun Life Gold ETF, Hang Seng RMB Gold ETF, Bualuang CHAY Gold ETF, K Gold ETF, Quantum Gold Fund, Motilal Oswal MOSt Shares Gold ETF, Religare Invesco Gold Exchange Traded Fund, Thanachart Gold ETF, ThaiDEX GOLD ETF, ETFS Physical Gold ETF, NewGold Issuer Ltd, ETFS Metal Securities Australia Ltd - ETFS Physical Gold, Istanbul Gold Exchange Traded Fund, BetaShares Gold Bullion ETF Currency Hedged, Kuveyt Turk Katilim Bankasi A.S.B Tipi Altin Borsa Yatirim Fonu, AfricaGoldETF.

For the few ETFs and ETPs in this list containing a portion of their assets in derivatives, only the total amount of actual physical gold was estimated and included in the computations.

Chart 3: Jewellery and technology make up nearly 60% of demand as emerging markets buy nearly 70% of all gold

(a)    Average demand by source. It excludes over-the-counter gold demand.

(b)    Average demand by region. It excludes over-the-counter and central bank gold demand.

Chart 4: China and India’s appetite for gold is closely linked to their rising incomes

(a)    Gold expenditure per capita as a function of gross national income per capita. Both in Indian rupee terms.

(b)    Gold expenditure per capita as a function of gross national income per capita. Both in Chinese yuan terms.

Chart 5: Rises in economic growth explain increases in jewellery demand

Changes in global gold jewellery demand regressed against changes in real GDP and the price of gold. All input variables in log scale. The original analysis performed using annual data between 1993 and 2012. The chart incorporates 2013 data.

Table 1: Jewellery demand in the top gold-consuming countries responds consistently to economic growth, income and price

Changes in regional gold jewellery demand regressed against changes in local real GDP, local inflation the price of gold in local currency. All input variables except inflation in log scale. The original analysis included data up to 2012, but due to data availability, starting points change: world (1993); India (1989); China (1999); US (1981); Russia (2001); Turkey (1984).

Chart 6: Electronics’ demand for gold is highly correlated to economic growth

Changes in global gold electronics demand regressed against changes in world real GDP, the previous year’s real GDP and a dummy (0/1) variable for 2001. Demand and GDP in log scale. The original analysis performed using annual data between 1985 and 2012. The chart shows the fitted model between 1993 and 2013 for consistency with Chart 5.

Chart 7: Geographic composition and source of demand is more diverse today than it was during the 1970s

The 1970s period is computed using the annual demand date from 1973 to 1979. The 2000s period is computing using annual demand data from 2004 to 2012. Negative values of demand was considered part of supply and not included in the calculations. Many of these figures have been subsequently revised as a result of changes in methodology.

•     

•     

•     

•     

Tail-risk hedging is not just a defensive strategy

A tail risk hedge is designed to gives investors positive cash-flow during market turbulence, thereby providing them with the opportunity to take advantage of oversold and mispriced assets.

Research from bond giant PIMCO argues that an aggressive portfolio with tail-risk hedges in place has more upside potential than a less risky one but similar levels of downside risk.⁶ This can give investors the confidence to use more aggressive investment strategies and ‘harvest more equity premium’ in the long term.⁷ As the PIMCO report comments, ‘Tail-risk hedging may allow certain investors to maintain an allocation to risk assets where they might otherwise deem the position to be too risky. It can also help stabilise portfolios on a mark-to-market basis.⁸

Tail risk hedges therefore provide not just the short-term buffer when markets fall but also confidence to invest higher up the risk scale over the long-term - and the opportunity to invest when markets are distressed.

Hedge strategies have to balance performance and cost

Highly levered strategies such as put option strategies or volatility futures require little initial funding and respond well but are costly over time. This partly explains why traditional strategies using cash or treasuries are still popular candidates as all-weather hedges – though the current low-interest rate environment has put a mathematical cap on their absolute effectiveness.⁹

Tail events are rare and almost impossible to accurately predict. There are plenty of natural buyers, but few sellers – who face large losses when tail events occur. Hence there is a high premium for this insurance and the higher the expected quality of protection, the more expensive the hedge. For example, a put option position could cost a standard portfolio about 2% per year in performance.¹¹ This can be hard to sell to investors when seas look flat and calm.

This is where we believe gold offers investors a useful alternative.

Although not immune to price drops, gold’s long term annual real return is positive. Gold also tends to negatively correlate to equities at times of large selloffs – a precondition of tail-risk hedges yet is often positively correlated to equities when they do well (Chart 2). Most tail-risk hedges, while more negatively correlated during tail events, do not enjoy that additional benefit.

Our dataset includes 11 two-sigma falls (and seven two-sigma rises). This is well in excess of the five events expected under a bell curve and reminds us that asset prices frequently behave in a non-normal way.

Volatility-based:

•

VIX short-term future: The S&P 500 VIX short-term futures index, an index of rolled front month VIX futures (rolls between 1^st and 2^nd future). The shorter maturity future has a greater response to volatility but is more mean reverting and has a higher roll cost than longer dated futures.

CDS IG: A liquid investment-grade credit default swap (CDS) on the bonds of 125 of the most liquid North American investment-grade entities. A systemic event is likely to instantly affect the market’s perception of credit ratings and default probability. CDS indices have become popular hedge instruments following the 2008-2009 financial crisis.

A CDS IG ‘on-the-run’ series is available through Markit; however, a listed vehicle available that incorporates the costs of rolling an ‘off-the-run’ into a new ‘on-the-run’ series is not currently available. Investors need to actively manage the timing of the roll when a new ‘on-the-run’ series when it becomes available. For this analysis, we constructed a rolling ‘on-the-run’ index that rolls at expiry.

*  CDS IG represents a theoretical construct based on available ‘on-the-run’ tradable indices.

Additional reference notes are listed at the end of this article.

Source: World Gold Council

*  CDS IG represents a theoretical construct based on available ‘on-the-run’ tradable indices.

Additional reference notes are listed at the end of this article.

Source: World Gold Council

*  CDS IG represents a theoretical construct based on available ‘on-the-run’ tradable indices.

Additional reference notes are listed at the end of this article.

Source: World Gold Council

*  CDS IG represents a theoretical construct based on available ‘on-the-run’ tradable indices.

Additional reference notes are listed at the end of this article.

Source: World Gold Council

The unique nature of each tail event is often overlooked in the literature. All tail events are likely to deliver higher asset correlations and lower asset prices but the relative collapse of assets will not be the same. And different triggers may have different consequences. So it is likely that the more prescriptive a hedge is, the less likely it is to react well to all possible triggers.

*  CDS IG represents a theoretical construct based on available ‘on-the-run’ tradable indices.

Additional reference notes are listed at the end of this article.

Source: World Gold Council

Gold has a role to play

Excluding the high costs of most hedges, gold has not – at least over recent and very short windows – been a particularly effective tail risk hedge. Our previous research has shown that in the longer term and over longer windows, gold’s attributes are far more certain. But the return profile of gold is compelling. The asymmetric correlation means it straddles the ground between a dedicated tail risk hedge and a risk asset. It contributes to portfolio performance during normal environments but still displays the properties of a tail-risk hedge. This suggests that some combination of hedges – following the old adage of not putting all of one’s eggs in one – should perform better than any single hedge. Gold’s unique profile further suggests that it should typically form a part of this combination.

References

Chart 1: When the left tail bites: market crises prompt big drawdowns

The 60/40 portfolio is a combination of MSCI World Total Return Index and the BofA Merrill Lynch Global Government 7-10 year Total Return Index.

Chart 2: Gold’s equity correlation is positive in right-tail events and negative when left-tail events hit markets

Chart 2 shows weekly return correlation between equities, gold and commodities when equities move by more than two standard deviations; January 1987 to July 2010

Chart 3: The last nine years have been particularly volatile

The chart shows the CBOE Volatility (VIX) index and its average level over four sub-periods between January 1990 and July 2014. The periods were picked arbitrarily to highlight the existence of periods of extended market uncertainty like the one we have seen during the past 9 years.

Table 1: On the basis of conditional returns during tail events, gold ranks in the middle of the pack

Table 1 shows the absolute maximum, minimum and average performance of our selection of hedges during ‘tail event’ weeks when our 60/40 portfolio dropped by two standard deviations or more (-3.9% or more). Our weekly dataset covers the period between January 2005 and July 2014 and results in 11 observed ‘tail events’.

Ranking is performed as follows:. Each category, represented by a column is ranked from one to nine. An average is taken of each ranked column and this average is then ranked from one to nine – a rank of ranks.

Table 2: Gold’s reliability has been uncertain

Table 2 follows the same ranking process as detailed in Table 1, but does not include the average return which was already ranked in Table 1 under conditional performance. It is listed however because it is a component of the information ratio (average divided by standard deviation) which is ranked. The column positive percent shows the proportion of tail-event returns that were higher than that of the portfolio (excess returns).

Table 3: The opportunity cost and damage limitation of your hedge is key

Table 3 ranks four measures that represent the opportunity cost of holding a hedge: That which is given up or sacrificed by allocating to the hedge. The first column is the allocation required to reduce average portfolio losses during tail events by 20% (from -5.5% to -4.4%). Twenty percent was chosen based on previous research (see footnote 35). Using this allocation, portfolio minimum and maximum returns are then ranked (the average is a constant -4.4%) as a higher value for either represents less return given up during the tail events. Equally the standard deviations of the portfolio returns are ranked, as they represent the cost (gain) of risk reduction. Ranking is done as per Table 1.

Table 4: A passive allocation to many hedges eats away at portfolio performance

Table 4 measures and ranks the annual average, volatility, information ratio and maximum drawdown of a portfolio with an allocated hedge (as per table 3) over the whole analysis period. Ranking is done as per Table 1. For CDS We estimate that this cost would have been less than 0.2% per year over the analysed period. Though small in this environment, a prolonged period of benign markets would amplify that impact as the cost of rolling could be higher as rates increase. The available ‘on-the-run’ index we use omits the roll-cost between tranches. To illustrate the estimated additional impact of the roll cost assuming a roll at expiry of a tranche, we constructed an index that includes the cost of rolls. During the period analysed, this instrument would have lost an annualised 2.2% vs. a gain of 0.4% excluding the cost. Using a 6.8% allocation as discussed in Table 3, the portfolio impact of including the roll would be 6.8% * (2.2%- 0.4%%) per year or 0.17%.

Table 5: Across all measured criteria, gold ranks jointly at second among hedges

Table 5 adds all the ranks together from Tables 1 to 4, averages these ranks and ranks them.

Gold-backed ETFs have been good for investors – and for gold

Over the past decade, gold-backed exchange-traded funds (ETFs) have transformed the gold investment market. Yet, their rise has not been universally popular. They may have put pressure on gold vendors’ margins, and some investors believe that they have led to an increase in the volatility of the gold market.

As we see it, ETFs may not be an answer to every gold investor’s needs, but the gold market has benefited more broadly since their launch. Overall, we believe that ETFs have reduced total cost of ownership, increased efficiencies, provided liquidity and access, and brought new interest – and demand – into gold as a strategic investment.

In this article, we explore the following themes:

Reference notes are listed at the end of this article.

Source: Respective ETF/ETC providers, Bloomberg, World Gold Council

*  For bars and coins, the average was computed using data from 1980 to 2007. For ETFs, the average represents data from 2003 to 2007. The average exclude data post 2007 to remove the effect of the 2008-2009 financial crisis on investment demand. Additional reference notes are listed at the end of this article.

Source: GFMS-Thomson Reuters, World Gold Council

*  Data as of 1997. Reference notes are listed at the end of this article.

Source: Bloomberg, Certified Coin Exchange, LBMA, World Gold Council

Our analysis suggests that, unlike in other markets, gold ETFs are unlikely to have added volatility to gold prices. In recent years, there has been a lot of criticism and commentary about the increase of volatility and correlations in markets attributed to the increased usage of ETFs – but the evidence is mixed.¹³

Reference notes are listed at the end of this article.

Source: Bloomberg, CFTC, CME, World Gold Council

ETFs have opened the door to additional investment demand

The past decade for gold ETFs has left a prominent mark, with future potential growth and development. Yet, there is room for further innovation. Some ETFs are available in countries with less developed financial markets, making them less popular vehicles for providing access to the gold market. In developed markets, there is still unmet demand for physical bullion.²²

As new interest in gold has combined with the established global investment demand for gold as a diversifying asset, tail-risk hedge, and preserver of wealth, we believe there is potential growth for several years to come. New and differentiating investment products will likely come to market to meet investors’ needs. This will no doubt continue to shape gold investment and further develop the efficiency of the global gold market into the next decade.

References

Chart 1: Currently, physically backed ETFs collectively hold 1,700 tonnes of gold worldwide

Data as of 31 August 2014. Gold holdings are as reported by the ETF/ETC issuers. Where data is unavailable, holdings have been calculated using reported AUM numbers.

Chart 2: Gold ETFs provide plentiful liquidity for investors

Average daily traded value calculated as the trailing one year volume as of 31 August 2014. Volume of securities is the aggregate trade value across publicly available exchanges.

Chart 3: ETF demand stands largely in addition to bar and coin demand

Deviation from mean calculated as the difference from the three-year annual average minus the average annual demand from the respective start date through 2007. For bars and coins, we computed the average using data from 1980 to 2007. For ETFs, the average represents data from 2003 to 2007. The average exclude data post 2007 to remove the effect of the 2008-2009 financial crisis on investment demand.

Chart 4: While many speculative holders have recently exited the market, strategic investors remain steadfast in gold ETFs

Data as of 30 June 2014. Holdings were taken from quarterly SEC Form 13F filings for SPDR Gold Trust and iShares Gold Trust to provide a representation of US gold ETF investors. 13F Filings have historically represented 30-40% of total outstanding shareholders at quarter end since inception of both ETFs. The 13F form is a filing with the Securities and Exchange Commission, also known as the Information Required of Institutional Investment Managers Form. It is a quarterly filing required of institutional investment managers with over $100 million in qualifying assets. Companies may include insurance companies, banks, pension funds, investment advisors, hedge funds, broker-dealers, and private wealth investors.

Table 1: Bid-ask spreads have remained tight for gold

Most volatile days defined as those exceeding three standard deviations of the average daily return for SPDR Gold Trust since inception (18 November 2014). Calculations as of 15 August 2014. Average of the best bid and best ask for the NYSE ticks were used to calculate the intraday spreads.

Chart 5: Premiums on gold bullion coins have reduced since ETFs have come to market

For the American Eagle, Canadian Maple Leaf, and South African Krugerrand prices, data series begins in December 1991. For the Chinese Panda, Viennese Philharmonic, and Australian Kangaroo prices, data series begin in June 1997 due to data availability. Average premiums calculated on daily basis as difference of coin price over daily London Gold PM Fix price.

Chart 6: Trading volume of gold futures is several factors larger than primary and secondary market volume of gold ETFs

(a)    Futures volume based on average daily volumes per month of all gold contracts traded on CME COMEX through 31 July 2014. GLD primary market volume based on daily average per month of net creation/redemption activity by authorized participants of the SPDR Gold Trust through 31 July 2014. GLD secondary market volume based on average daily volume per month of SPDR Gold Trust shares traded in the NYSE through 31 July 2014.

(b)    OTC market volumes are not often published. Instead, we use the daily average per month of clearing statistics of the net (loco London) gold transfers settled between clearing members of the LBMA through 30 June 2014. The range shown in the picture multiplies net transfers by a factor of six (lower limit) and 10 (upper limit). This is consistent with previously published research by the World Gold Council. It is also consistent with an estimate of total loco London volumes based on a 2011 survey that the LBMA conducted among its gold trading members. Futures volume based on average daily volumes per month of all gold contracts traded on CME COMEX through 31 July 2014. GLD secondary market volume based on average daily volume per month of SPDR Gold Trust shares traded in NYSE through 31 July 2014.

Chart 7: Speculative net long positions vary much more than gold ETF

Standard deviation is calculated on a rolling three-year basis. For GLD the start date of data is March 2005 to prevent volatility due to inflows of fund of the ETF during the inception month of November 2004

The printer is accredited to

ISO14001 environmental standard

I093201409