US Dollar Bears and Gold Bulls

Curb Your Enthusiasm

This week’s AOFM 30-year bond syndication with a $36 billion book was well-received, not least by offshore investors, who secured a record 67% allocation. These foreign inflows into AUD debt and associated AUD strength are symptomatic of the RBA’s inaction relative to other, more activist, central banks in recent months. As WIB wrote:

At the other end of the spectrum, the RBA and the Riksbank remain among the least “generous”, neither central bank expanding their balance sheet at all for several months. Is it purely coincidental that AUD and SEK have also been among the strongest G10 currencies in the last 3 months?

We should nonetheless concede that the USD has been weaker across the board, not just against the AUD, which has brought out the ‘end of the dollar’ crowd in predictable fashion, although one would normally expect a bit better than this effort from Goldman Sachs.

The weakness in the US dollar is only relative to its surge in March and April on safe-haven flows that have since moderated along with the pandemic, at least outside the US itself. So recent price action is perfectly consistent with the US dollar’s safe asset status. Moreover, as I pointed out in my report on the future of the US dollar for USSC last year, fluctuations in the US dollar exchange rate are perfectly consistent with the dollar’s international role and make the US economy more resilient to shocks, not less.

If there is anything more annoying that US dollar perma-bears, it’s gold price perma-bulls, whose joint appearance is almost guaranteed by the straightforward valuation effect the US dollar exerts on the US dollar gold price. Using strength in the US dollar gold price as a benchmark of the US dollar exchange rate’s fortunes is somewhat tautological.

So here’s my take on gold, but here’s the kicker. I wrote most of this in 20006 with an update in 2016. Apart from a few stylistic changes, I haven’t changed anything of substance. It hasn’t appeared anywhere else as it was part of an unpublished book draft, so you will just have to take my word for it.

Gold is often portrayed as an inflation hedge or protection against other adverse macroeconomic outcomes, but in inflation-adjusted terms, the long-term returns to gold have been risible. Between 1836 and 2011, the average real rate of change in the US dollar gold price was 1.1% per year.  Even at its previous peak in 2011, the inflation-adjusted gold price was below its previous peak in 1980, implying no real returns to gold from a buy and hold strategy over a period over more than 30 years. While there may be some diversification value and an unobserved implied yield from the ‘services’ gold provides to those who hold it, the investment case for gold is not compelling.

Since 1975, there has been a positive, but not statistically significant, covariance between real gold returns and US inflation. The real gold price also shows little covariance with consumption and economic growth. While this implies that gold is not a risky asset, it also means that the returns to gold will be similar to the low ‘risk-free’ rate of return typically associated with short-term, government-issued debt instruments. As an investment, gold is historically little better, although certainly not worse than, the government paper that worries so many gold bugs.

Every commodity has its own unique quirks. Gold is distinctive in being non-destructible as well as being homogenous and fungible, unlike oil, which comes in different forms. Non-destructibility, as well as gold’s former role as money or backing for money, has important implications for the nature of the market for gold. Above-ground stocks are large (153,000 tonnes as of 2004), relative to the annual newly mined supply (2,464 tonnes in 2004).  The gold price is thus determined largely on the basis of existing stocks rather than supply that is newly mined. This is in sharp contrast to oil, where new supply is mostly burned rather than put into storage. This means the holders of gold have market power, not the miners. And the biggest holders of gold are the world’s central banks, a legacy from when gold provided backing to either the currency or the exchange rate of many countries.

There is a certain irony in some anti-government types and investors seeking protection from inflation by central banks in a commodity that is largely owned by central banks, who these days would probably rather not hold it! They would rather hold income-producing assets rather than an asset that has a negative yield after storage and transaction costs.

With central banks and other institutions accounting for about 20 per cent of above-ground stocks, it should not be surprising that official sector transactions influence the gold price. Between 2002 and 2006, central bank sales accounted for an average 14% of global supply.

The sensitivity of the gold price to changes in central bank gold holdings should make gold bugs nervous. The most recent upward trend in the US dollar gold price began in September 1999 only when the world's major central banks entered into a five-year agreement that capped their gold sales at 400 tonnes annually, subsequently revised to 450 tonnes in 2004. The third Central Bank Gold Agreement signed in August 2009 again limited sales to 400 tonnes annually, including the IMF's planned sales. Before the Central Bank Gold Agreement in 1999, central banks found themselves selling into a declining market.

Australia's Reserve Bank provides a case in point. In July 1997, the RBA announced it had sold 167 of its 247 tonnes of gold, reducing gold's share of the RBA’s international reserves from 20 per cent to about 7 per cent. The announcement after the fact still had a depressing effect on the world gold price.

As a major producer and exporter, Australia is naturally long gold. From a portfolio management perspective, there is little diversification value to be obtained from hoarding in the RBA's vaults what can already be found beneath the Australian outback.

Moreover, the gold reserves had been purchased at the old Bretton Woods US dollar-gold parity price of $US35 an ounce. Even at the (now seemingly modest) July 1997 price of around $US 330 an ounce, the RBA realised a tidy capital gain, the proceeds of which were used to purchase other higher yielding assets. One Australian journalist mistakenly accused the Reserve Bank of ‘cost[ing] the nation $5 billion’ by selling into at a major low in the gold price.  In fact, the RBA realised a profit on the sale while re-investing the proceeds in assets that were almost certainly higher yielding in the long-run.

Selling the publicly-owned gold stock is frequently mooted as an (admittedly poor) solution to a country’s fiscal problems. Both social democratic and conservative think-tanks in the United States have advocated either selling or monetising the gold stock to reduce the US government’s debt.  The International Monetary Fund has been selling its Bretton Woods-era gold holdings and converting them to income-producing assets in an effort to stave-off reform by making the Fund more financial independent and thus less accountable to its member countries. With COVID-19 smashing budgets around the world, offloading official sector gold stocks and replacing them with income-producing assets has some appeal.

A favourite gold bug indicator is the Dow/gold ratio, which is said to represent the Dow denominated in ‘real money.’  This is true only if you think the gold price benchmarks real purchasing power.  We could equally measure the Dow in terms of another commodity (as is sometimes done with oil).  Since commodity prices are typically much more volatile than the prices of consumer goods and services, they are a poor benchmark for consumer purchasing power, which is best measured with respect to the Consumer Price Index. 

Rather than a real price, the Dow-gold ratio is better viewed as a relative price or a measure of the relative performance of stocks and gold. This is only useful as a valuation tool if you believe there is a long-run relationship between gold and stock prices.   The long-run increase in the Dow/gold ratio highlights the long-term underperformance of the gold price relative to income producing financial assets and not the over-valuation of stocks, as the gold bugs would have it. Time the cycle right and you might make money going long gold and short stocks, but the secular trend in the Dow/gold ratio is not your friend.

Here is the real gold price plotted against US consumer price inflation:

Note the considerable volatility in the gold price relative to inflation. Only at very high rates of inflation does the real gold price offer some protection based on a few data points that have somehow become firmly lodged in the collective consciousness.

For some gold bugs, the high inflation scenario is the whole point. Whether that scenario is likely to come to pass anytime soon is left as an exercise for the more fevered moments of the reader’s imagination.