LONDON () -- The World Gold Council have just published a report by two Scotland-based academics, Eric Levin and Robert Wright, on the Short-run and Long-run Determinants of the Price of Gold.
The report describes a comprehensive econometric study using cointegration regression techniques to investigate the determinants of the gold price, using data from 1833-2005. Five key findings from the report are listed below and then developed in the following sections. The most actionable finding is number five in the list: buy gold if you believe that the dollar will depreciate in real terms.
Key Findings
- There is a long term relationship between the price of gold and the general price level in the U.S. by which a 1% rise in U.S. inflation leads to a 1% increase in the price of gold.
- However short-term deviations from the long term relationship are caused by factors which affect the supply and demand for gold; particularly U.S. inflation and inflation volatility, credit risk, the dollar trade-weighted exchange rate and the gold lease rate. These short-term deviations can last for years and there is only a slow reversion back to the long term rate.
- There is no significant relationship between changes in the price of gold and world inflation and inflation volatility, world income and gold's 'beta' (specifically gold's correlation with the S&P500 index).
- Gold is profitable as a long run inflation hedge for investors in countries whose currencies depreciate against the dollar by more than the difference between the country's and the U.S.' inflation rate. Historically these countries have included India, China, Turkey, Saudi Arabia and Indonesia. It is no coincidence that these countries are among the top consumers of gold.
- If you believe that real dollar depreciation against other currencies is inevitable (and the authors argue many good reasons why this should be so, the only issue being exactly when and by how much) then U.S. wealth holders should profit from holding gold as:
- Dollar deprecation will lower the price of gold to investors outside the U.S., which will raise their demand for gold and thus raise the dollar price of gold.
- Dollar depreciation will raise U.S. inflation. Gold would then act as an inflation hedge.
Inflation Is the Key Long-Term Determinant of the Gold Price
The authors hypothesise that the long-run price of gold should rise in line with inflation essentially because it is related to the marginal cost of extraction. This in turn is related to inflation. They tested the relationship using both U.S. inflation and world inflation as the explanatory variable and found that:
- The U.S. price level was the best predictor with a 1% increase in the U.S. price level leading to a 1% increase in the gold price (at the 95% confidence level). They conclude that gold is therefore an effective hedge against inflation.
- After a shock it takes a long time to revert to the long term relationship. The 'error correction term' was estimated to be 0.019. This means that each month's error is 2% smaller than the previous month's. So after five years about two thirds of the deviation from the long term relationship would have been eliminated.
The chart below, taken from the report, illustrates the long term relationship. In 1833 the price of gold was $20.65 per ounce equivalent to $415 in 2005 money. In 2005 the actual price of gold was $445 implying a very small change in the real price of gold over the 172 years.
The chart also shows however clear evidence of extended periods when the gold price has moved away from the long run relationship with inflation and when short-term determinants of the gold price have ruled the roost.
Short Term Determinants
Levin and Wright hypothesise that movements in the short-run price are determined by supply and demand where:
- The short-run supply of gold depends on the amount supplied by the gold producers, either through leasing gold or through extraction, and on the willingness of central banks to lease gold. This in turn will depend on several factors such as the gold price, the lease rate, the interest rate, the level of political and financial turmoil and the default risk.
- The demand for gold depends on the 'use' demand (for jewellery, medals, electrical components etc.), which is likely to be negatively related to the gold price, and on the 'asset' demand for gold as an investment. The latter is likely to be affected by expectations about inflation and exchange rates, "fear", the returns on other assets and lack of correlation with other assets, and the real interest rate (as this is the cost of holding gold).
They therefore tested the influence of the following explanatory variables on the short-run price of gold using cointegration techniques on monthly data from 1976-2005.
- U.S. and world inflation and inflation volatility: on the basis that an increase in these variables could increase the need for an inflation hedge and thereby the demand and price for gold.
- World income: as rising income may increase the demand for gold jewellery.
- Dollar exchange rate: a fall in the dollar would make gold cheaper for investors outside the dollar area.
- Gold lease rate: as this is a proxy for the world real rate of interest and thus the opportunity cost of holding gold. It is also a key factor on the supply side.
- Gold's beta (relative to the S&P500 index): as this could determine portfolio demand for gold if it reduces portfolio volatility.
- Credit risk: a rise in financial turmoil/credit risk would increase the demand for gold but reduce the supply as central banks would be less willing to lease gold.
- Risk (as measured by the index of global political risk): greater political risk may drive demand and reduce supply.
The most important determinants were found to be a positive relationship with U.S. inflation, U.S. inflation volatility and credit risk and a negative relationship with the dollar exchange rate and the gold lease rate. World income, world inflation and gold's beta were found to be not statistically significant.
The political risk index was found to need further investigation as there were 10 periods, each of a month's duration, where the explanatory variables above failed to explain all the significant variation in the gold price. For some periods the reason was clear; the October 1999 spike for example was caused by the first Central Bank Gold Agreement. For other periods the reason was less obvious though by Levin and Wright demonstrated the importance of oil risk by using a sub-index of the political risk index based on the political risk of the top ten oil producing countries minus the U.S.
Changes in the U.S. price level were found to be the most important explanatory variable accounting for 36% of the variation, followed by the exchange rate which explained a further 18%%.
Gold as a Hedge for Non-US Investors?
Levin and Wright show charts for each of the main consuming and producing countries of gold which illustrate whether or not gold would have been a good hedge against local inflation. For example the chart below shows that in India gold has been very effective as a hedge over the last 30 years.
The gold price in 1976 was around 1100 rupees. Since then Indian prices have risen about ten times so an inflation hedge would now need to be worth 11,000 rupees. The gold price however rose to over 20,000 rupees by 2005 and has always outperformed the inflation hedge price.
The various charts show that gold would have been effective as an inflation hedge in a number of countries including China and Saudi Arabia as well as in countries that have experienced hyperinflation such as Turkey and Indonesia. However in certain countries such as Australia and Japan it would not have been effective.
Instead of gold the Indian investor could have bought dollar-based assets. These would have enjoyed the same gains as gold but yielded additional interest and eliminated the risks associated with the gold price. However the investor would have lost the utility from wearing jewellery. The authors (they are male!) conclude that it might seem that real interest-bearing dollar-denominated assets are better than gold provided that there is no expectation of real dollar depreciation against the rupee. However for reasons explained in the next section they argue that dollar deprecation is inevitable and that in these circumstances it would be preferable for an Indian investor to hold gold.
Dollar Depreciation
The authors claim that there is growing consensus among economists that deprecation of the dollar is likely. This is partly because of the large increase in China's foreign currency reserves, and partly because the already-huge U.S. current account deficit is likely to get even worse. (They cite an OECD report which gives five reasons for expecting the current account deficit to worsen and fourteen estimates, ranging from 12-90%, of the dollar deprecation necessary to restore balance).
The authors don't discuss in the report whether there will be a soft or hard landing and the extent of the dollar deprecation. They say that it is sufficient to say that there will be a significant degree of depreciation which will make a sizeable impact on wide range of financial assets including gold.
In this instance they claim that U.S. wealth holders would profit from holding gold for the reasons listed at the beginning of this article:
- Dollar deprecation will lower the price of gold to investors outside the U.S. which will raise their demand for gold and thus raise the dollar price of gold
- Dollar depreciation will raise U.S. inflation, and gold would then act as an inflation hedge.
There are many statistical studies into the determinants of the price of gold. Lewin and Wright group these into three types: those which use macroeconomic variables, those which look at speculation and rationality of gold price movements and those which examine gold as a hedge looking at both the short and long term. Their analysis, they claim, extends the third approach by including political and financial risk into the analysis.
