There is a lot of interest in the nature of relationships between commodity prices and financial variables. Indeed, there are many sophisticated techniques available to analyse the myriad factors at work in commodity markets. The following article aims to shed some light on the relationship between gold and a number of key financial and economic variables.

Stylised 'facts' about gold price relationships

Volatility on Wall Street has revived hopes of a recovery in the gold price in some quarters. This reflects a perception that the desire to hold gold increases during periods of heightened uncertainty - that gold is a store of wealth. Intuitively, therefore, the price of gold should respond positively to :

  • Political uncertainty, such as the threat of war, and;
  • Economic uncertainty, such as a stock market sell-off, rising inflation or periods of high volatility in financial markets.

Alternatively, the price of gold may be adversely affected by:

  • Economic optimism, where other assets promise greater risk-adjusted returns, and particularly
  • Conditions favouring US dollar appreciation, as this tends to depress the price of all USD-denominated assets.

While the motivation for holding gold is relatively clear cut, the increasing globalisation, sophistication and deregulation of financial markets have given rise to an array of alternative sale havens. This would tend to weaken the allure of gold over time, depending on the stage of development of individual markets.

Nonetheless, in less -developed countries commodity gold still plays an important role during times of stress. Developing countries account for two-thirds of new gold consumption. Indeed, anecdotal evidence suggests that large volumes of gold flowed back into Korea and Indonesia after the Asian financial crisis, despite the high local currency price of the metal.

The Data Supports the 'Facts'

Correlation coefficients calculated from monthly levels of gold prices and a selection of variables over three decades lend support to these a priori assumptions. The results are presented in chart form below:

  • The estimated coefficients have the right signs (+/-). The gold price is positively correlated with oil prices, inflationary pressures and interest rates (Fed fund s and ten-year Treasuries). It is negatively correlated to episodes of US dollar strength (arising USD effective exchange rate) and equity market optimism (proxied here by the average P/E ratio of S&P 500 stocks, to normalise the exponential rise of stock prices since the mid-1980s).
  • The absolute size of the coefficients has diminished with time for the oil price, inflation and interest rates.
  • The relationship with equity prices became more muted during the 1980s but strengthened again during the 1990s for reasons that remain unclear, given the range of alternative investments available. It is likely that the coincidental impact on the gold price of concerns about central bank sales exaggerated the measured relationship towards the end of the decade.
  • The correlation between the gold price and the exchange rate remained strong. This is because the value of the USD directly affects the calculation of the USD gold price, irrespective of sentiment. While the long-run coefficients highlight gold's theoretical relationships with other variables, the correlations have diminished on average. Of course, the relationship may still become pronounced during shorter periods of extreme volatility, but the evidence surveyed below provides only tenuous evidence of such links.

Gold Prices and Equity Markets

In October 1987, the Dow Jones plummeted from 2508 to 1739 in four trading sessions before bouncing and stabilising above 1900.

  • On six of nine days the gold price and the stock index moved in different directions, with daily price levels registering a correlation coefficient of -0.84.

In April 2000, the NASDAQ also experienced a volatile nine-day period, trading between 4446 and 3321.

  • Gold and equity prices again moved in opposite directions on six occasions (correlation = -0.50).

Six favourable outcomes out or nine represents only modest support for our stylised ' fact', diminishing the power or the strong correlation observed in 1987. Meanwhile, changes in the levels of the prices show that gold generally moves by only a small amount - and sometimes not at all - in response to even quite large changes in stock prices.

Thus, from one day to the next, it appears that gold may provide a modicum of protection against stock market jitters, rather than delivering large returns.

However, what would happen to the investor who was unwilling or unable to trade on a day-to-day basis? After dividing the year 1987 into five periods, co r responding roughly to turning points in the Dow Jones, it can be seen that when optimism was high, gold significantly underperform ed the stock market. But when sentiment turned bearish gold came into its own. A zero correlation for the year as a whole illustrates gold's potential as a hedging instrument.

Period Dow Jones Percent Gold Price Change
I 26.9 7.9
II -7.9 12.6
III 22.8 -3.6
IV -36.1 5.1
V 11.5 1.0

However, similar treatment of year-to-date NASDAQ performance revealed a positive correlation between gold prices and stock prices over four month s. For example, period IV shows the two price series falling together for a time, although gold rallied in late March as the NASDAQ came under increasing pressure.

Period NASDAQ Percent Gold Price Change
I -5.9 -1.7
II 12.7 6.0
III 15.2 -4.3
IV -34.2 -2.5
V 16.3 -2.7

Gold and Oil

The prospect of war in the Arabian Gulf caused a spike in oil prices during 1990. The oil price also rose precipitously when OPEC nations decided to limit supply last year. Oil price spikes in the 1970s caused quite severe and widespread economic problems. However, the impact of oil price spikes has greatly diminished during the 1990s.

In accordance with prior expectations :

  • Gold and oil prices moved together in six of the nine months straddling the peak in oil prices in both 1990 and 1999 -00, and;
  • Gold prices generally moved considerably less than oil prices, the only exception being in October 1999 when the gold price reacted to the European central banks' announcement about their gold holdings.

The medium-term investor who bought gold in mid-July 1990, in response to the rising oil price, would have been about $US40/oz (10 per cent) richer at the end of the year - oil was 50 per cent higher. An early December 1999 gold purchase would have returned 1 per cent return by the time oil peaked in mid-March 2000 - oil rose 30 per cent. Thus, gold provided only limited protection from higher oil prices.

Conclusion

Historically, demand for gold has increased in times of crisis. The gold price has risen with the oil price, inflation and interest rates, on average, and against falling stock prices. However, the relationships have faded with time. Overall, the most enduring link is between the gold price and the USO. This suggests that gold may have useful risk characteristics in portfolios of assets. Gold is a hedge, if not always a strong one.