“Gold, though of little use compared with air or water, will exchange for a great quantity of other goods” – David Ricardo
Gold has been among the best performing asset classes of the decade, with a compound annual growth rate (CAGR) of 13.2% from 2000 to 2008 and a noteworthy 25.5% in the last three years. Furthermore, gold has consistently outperformed other traditional asset classes such as developed and emerging market (EM) equities and US bonds throughout the same time period.
The pace of nominal price appreciation inevitably raises questions about how long the rally can last. Gold experienced a violent correction in September 2011, falling over 15% from $1900/t oz to $1600/t oz, sparking concerns that the historic gold rally may be ending. Gold prices have already recently surpassed their previous inflation-adjusted highs in 1980 of $1627/ t oz (Figure 4).
The unique challenge with gold price forecasting is that, due to its peculiar chemical and historical characteristics, gold is neither perishable nor consumed like most other commodities. This has created a large overhang of readily accessible inventory above ground, mostly in the form of jewellery, physical investment holdings, and government bullion stockpiles (Figure 5).
The scale of this inventory dwarfs that of most other commodities. For example, GFMS estimates the total amount of above-ground stocks of gold at roughly 166,600 tons in 2010. At current levels of jewellery and industrial fabrication demand, this equates to about 73 years worth of consumption. Even including demand from investment, current holdings can satisfy 45 years at current rates. And even this concept of inventory overhang is slightly misleading as all above-ground stock, whether in the form of jewellery sales, electronics scrap, or investment liquidation, is not consumed but recycled into circulation.
Hence, gold prices have at best a tenuous relationship with traditional demand and supply factors in the short term. Familiar concepts to guide equilibrium price forecasting, such as demand and supply balances and replacement costs of production become less useful. Figure 6 shows total supply from mine production, producer hedging, central bank sales, and gold scrap, balanced against total demand from fabrication and investment purposes. Figure 7 shows the weakness of the relationship between the derived excess demand balances and prices.
Already, gold prices have far overshot most physical measures of replacement costs. The marginal producing basin is South Africa at $750-800 in operating costs, but new mines have opened to replace falling South African production with operating costs at $500 and all-in costs at $650.
Figure 1. Citi Gold Price Forecast
|Figure 2. Nominal Gold Prices from 2000 to Present
|| Figure 3. Gold Returns Versus Other Asset Classes, 2011 to Present
Instead, gold (and to a lesser extent, other precious metals and gemstones) possess a special status as a universally accepted medium-of-exchange, despite their relative physical uselessness. This has made the metal’s market price beholden to the same macro-financial factors such as inflation, currency movements, and interest rates that drive other traditional asset classes. Indeed, the word “precious” derives from the Latin word pretium, meaning “price.”
Until the 1971 collapse of the Bretton Woods agreement, the world financial system was based upon the gold standard, with the value of paper currency backed by guaranteed exchangeability with bullion. Today, in reverse fashion, we should consider the price of gold as a proxy for the purchasing power and desirability of paper currency and other financial stores of value against real goods.
Figure 4. Long-Term Real Gold Prices from 1850 to 2010
Figure 5: Breakdown of 2010 Above-Ground Gold Holdings by Source
Inflation and other Denomination Effects
As nominal gold prices are quoted in US dollars, we must first consider changes in nominal prices that are due purely from changes in the value of the underlying denominator.
One of the most common measure of the changing purchasing power of the US dollar against real goods is nominal price inflation against a consumption basket. Real gold prices after adjusting for inflation have reached all-time price highs, surpassing previous highs in 1980 (Figure 4). But even after adjusting for inflation 1-for-1, inflation can still drive real gold price returns, particularly in periods of high inflation such as during the 1970s (Figure 8).
Another common measure of changes in the purchasing power of the US dollar are foreign exchange movements against baskets of other currencies. In fact, due to the law of Purchasing Power Parity, inflation and currency depreciations should track each other over the longer run. But assessing the impact of currency movements on nominal gold returns depends on the choice of benchmark.
|Figure 6. World Gold Demand and Supply Balances
||Figure 7. World Excess Demand Against YoY Returns
For example, the US dollar has actually appreciated relative to a broad basket of currencies of its major trading partners since the 1980s, and hence the rise in nominal gold prices is actually even more striking in terms of a broad currency basket. However, at the same time, the US dollar has depreciated against a smaller subset of major currencies such as the euro and the yen in the same time period. Figure 9 shows how at least some of the nominal appreciation in gold prices may be due to this depreciation against major currencies.
It is likely that the dollar would continue to appreciate against a broad basket given the higher rate of inflation in the US’ major trading partners. However, the dollar’s movement against major currencies such as the euro remain more challenging. Our forecast is for an appreciation of the dollar against the euro to 1.25 by end of 2012. But an exit, whether temporary or permanent, by one or more of the troubled periphery countries such as Greece, may cause the euro to violently appreciate as the currency zone offloads the sovereign risk.
What is the outlook for inflation in the future? According to the standard “New Keynesian” inflation paradigm, future inflation is determined by a combination of monetary expansion, the output gap, and inflation expectations. We will discuss monetary policy and real interest rates in the next section.
|Figure 8. Real Gold Returns and Nominal Inflation from 1971 to Present
||Figure 9. Gold Prices Adjusted Against Currency Baskets
Measures of current output gap, whether of capacity utilization or the unemployment gap, remain historically large, providing deflationary pressure onto the US and global economy (Figure 10). Furthermore, measures of inflation expectations, whether from surveys or the inflation-linked securities market, remain stable (Figure 9). On the other hand, should inflation expectations un-anchor due to the unprecedented amount of monetary expansion, nominal gold prices may react stronger than one-of-one as was seen in the 1970s.
Real Interest Rates
Having adjusted for denomination effects, what else drives real gold price returns? Historically, a reasonable proxy has been real interest rates (Figure 10). The logic is as follows. If real interest rates are high, that means the cost of borrowing capital (or the opportunity cost of not investing in bonds and other real assets) is high, weakening investment demand for gold. Conversely, if real interest rates are low, capital is cheap and investors seek non-traditional methods of gaining returns on their portfolios, including investing in gold.
Furthermore, real interest rates capture through forward-looking nominal interest and inflation rates the impact of central bank money supply expansion and the resulting loss of purchasing power of the monetary base.
|Figure 10. Measures of US Output Gap, 2000 to Present
||Figure 11. Measures of Inflation Expectations, 2000 to Present
|Figure 12. Real Gold Returns and Real Interest Rates, 1969 to Present
||Figure 13. Central Bank Net Gold Purchases
Currently, real interest rates are at all time lows, due to the after-effects of the Great Recession of 2008. Not only have most of the large economy central banks pushed their nominal (and real) interest rate targets to near bottom, but many, confronted with the need for further monetary expansion in the face of the zero-interest lower bound, have resorted to unconventional balance sheet expansions, loosely known as quantitative easing.
Though the US Federal Reserve has recently chosen not to undertake a third round of balance sheet expansion in favour of a less dramatic “Operation Twist,” Fed Chairman Bernanke has left open the option for more QE. Meanwhile, the Bank of England has decided to expand its current QE program by another £75bn, and the European Central Bank is weighing another cut to its main refinancing rate before year-end.
High levels of sovereign indebtedness and political deadlock between calls for fiscal easing versus austerity have handicapped fiscal policy. This has placed an undue amount of pressure on monetary authorities to provide a commensurate monetary response. With economic growth in the developed world struggling to gain momentum in the face of high unemployment and consumer pessimism, it is challenging for us to imagine any kind of interest rate tightening from the developed world central banks for the foreseeable future.
|Figure 14. Gold ETF Flows and the VIX Index, Jan-06 to Present
||Figure 15. Realized and Option-Implied 1-month Gold Volatility
Lastly, on top of the generally poor environment for economic growth and returns, global markets have been shaken by new systematic contagion fears stemming from the near-insolvency of several euro zone periphery and core nations.
At the same time, central banks, already among the largest holders of physical gold reserves, have become net buyers of gold for the first time in two decades (Figure 11). Notably, purchases include 25 tons (or $1.24bn) of gold by the Bank of Korea in July, with central banks from India, Thailand, Mexico and Russia also revealing large gold purchases this year.
Furthermore, the financial industry has introduced new ways for investors to seek investment exposure to gold, notably through Exchange-Traded Funds or ETFs, which provide an equity-like financial structure linked to physical gold. The largest ETFs include the SPDR Gold Trust and the ETFS Physical Gold fund, and collectively account for about 2,000 tons of physical gold.
As financial fear (as seen in the VIX index) spiked both in 2008-09 and again more recently, gold ETF investment has also seen substantial inflows, with an estimated 876 tons of gold equivalent purchased by investors from 2008 to 2009 and an additional 343 tons since 2010.
Despite some sales in recent months, net inflows into gold ETFs should remain strongly positive as long as macro-financial conditions remain unsettled. On the other hand, the usefulness of gold ETFs as safe havens and a source of collateral also means that they are exposed to sudden sell-offs during a liquidity squeeze, such as that in September 2011.
Figure 16. Citi Gold Price Forecast Trajectory
We believe the political intricacies surrounding a euro zone bailout preclude a rapid resolution of the ongoing crisis, while the weakening economic environment may spark other financial and sovereign difficulties elsewhere in the globe. In such conditions, financial demand for gold as a safe-haven asset class should remain as robust as ever but subject to high volatility.
When Do the Eggs Stop?
When gold is affected by such extreme macro-financial pressures, there is no “right” price for gold and making an actual price forecast is a somewhat futile gesture, in our view. Forecasting nominal prices to continue rising at the same average rate as occurred since 2009 seems as reasonable an assumption to us as any. However, we expect volatility, both realized and implied, to rise (Figure 13), with gold prices oscillating wildly above and below the $2000/t oz threshold throughout 2012.
Once macroeconomic conditions stabilize and real interest rates begin rising, gold would likely correct violently closer to its historical mean at $450/t oz. But timing the popping of the gold bubble is, almost by definition, subject to psychology and market conditions, making it a nigh impossible task.
In its September 2011 Global Financial Stability Report, the IMF has warned that the low policy rates, though necessary to stimulate growth, may threaten longer-term risks to financial stability by forcing investors searching for higher yield to leverage up certain sectors of the global economy. It warns that both advanced and particularly emerging market policymakers need to step up macro-prudential policies to contain excesses.
A strong economic recovery in 2012 may bring with it improved resilience of the global financial system and a sooner resolution of outstanding sovereign and banking risks, but history suggests that the aftermath of an asset price collapse casts a long shadow on potential future growth prospects.
In times of macro economic stress, gold can unlink from traditional demand and supply equilibrium and instead move according to macro-financial drivers such as real interest rates, inflation/currency denomination effects, and safe-haven financial demand. A confluence of weak economic growth, unprecedented monetary expansion, and heightened fears in the global financial system has driven historic nominal and real gold returns. We believe this confluence of factors will remain in place for the foreseeable future. Assuming nominal gold prices continue to rise at the same rate as it did for the previous three years, our price forecast is for nominal gold prices to average $1575/t oz for 2011 and $1950/t oz for 2012.