- The price of gold is testing new record highs as the yellow metal is expected to cross the $2000/oz mark for the first time in its modern history
- The seemingly unstoppable ascent of gold is due to a combination of negative real market interest rates and sliding dollar in the face of the FED’s continuing massive injections of liquidity into the US and global economy, fiscal stimulus in trillions of dollars (with a new stimulus expected) and the FED’s commitment to maintain ultra-loose monetary policy for the foreseeable future
- The uncertainty related to the expected path of the COVID-19 pandemic is another tailwind for gold, with the recent surge in coronavirus infections and deaths in the US and in other countries
- Going forward, we expect gold to continue its ascending movement favoured by a weakening dollar and persistent macro uncertainty although a correction and consolidation might occur in the short term for technical reasons.
- Investors will also have to consider the increased price volatility associated with an overcrowded market as there have been record inflows into gold-backed funds.
Gold has witnessed an unprecedented rally since the start of the year which is now pushing the price of the yellow metal beyond the $2000/oz mark for the first time in its history.
The performance of gold is impressive both in nominal terms and in real terms as can be seen on the chart below which tracks both prices over the 1968-2020 period. The current rally actually started by year end 2018.
Volatility of gold prices has also been edging up significantly showing the sudden increase in transactions in the futures markets as traders seem to adjust to a New Gold Normal.
The factors behind gold rally
There is a close relationship between the price of gold and the yield-to-maturity on 10-Year Treasury Inflation Indexed Securities (or 10-Year TIPS notes), a straightforward market-based indication of prevailing real interest rates (cf. the chart below). TIPs protect investors against inflation, a traditional quality also attributed to gold although the results in this regard are more mixed than usually thought (cf. our previous article available for PREMIUM subscribers). But this close relationship has been magnified over the last few years by the
Disappointingly enough for high frequency traders (be they humans or most likely robots), the daily correlation between daily gold returns and the variation of real interest rates in percentage points is not as meaningful as the analysis of the overall co-movement between these variables.
Analysing co-movements between real yields, gold prices and the dollar index
In order to assess the later, we performed a VAR (Vector Auto Regressive) analysis over the 2003-2020 period taking just these two variables (gold daily returns, daily variation of TIPS in percentage points).
The results are as expected. TIPS yield do not react to Gold prices whereas the latter react to TIPS, as can be seen from the following estimation results for the regression of gold returns on previous gold returns and previous real interest rates variation. This type asymmetry cannot be captured by a correlation analysis.
Performing the same analysis and including the trade-weighted dollar index as an exogenous variable yields even more interesting results.
Finally we perform the same VAR analysis by including the three variables in the VAR (over the 2006-2020 period). We find that real yields are heavily influenced by a variation in the level of the dollar. The same is true for gold.
Incidentally, the equation for the dollar index shows that it reacts positively to a variation in real yields as expected by the economic theory.
However, including gold in the equation as an endogenous variable might bias the results as it can be seen from the analysis of the correlations between the residuals of the three variables.
One argument that could be put forward to criticise the results of this VAR analysis is that the volatility of the different variables varies considerably over the analysis period. However, econometric theory shows that parameter estimates derived from an estimation by OLS are robust to heteroskedasticity (the academic jargon for non constant volatility), although the later tends to widen the confidence intervals around the estimates.
In addition, we performed two causality tests (Granger, Wald) to determine if the variation of TIPS yields “causes” (statistically speaking) the variation of daily gold returns. Once again the results are quite telling. The null hypotheses that “tips do not Granger-cause gold” and that “there is no instant causality between tips and gold” can be rejected with a very high level of confidence. In plain language, this means that gold prices consistently react to a variation of yields on TIPS. The latter is either caused by a variation of inflation expectations or by a variation of nominal interest rates, which react to inflation but also to a whole set of other factors : growth expectations over the medium to long term – including expected long term potential growth and growth volatility -, international flows of funds – the global demand for safe haven assets and the role of the dollar as the de facto international currency – and last but not least to (un)conventional monetary policies such, primarily government bonds purchases by central banks.
Is this the end of the story? Well it would be all too easy to think so. Actually, what you we would like to understand is what causes the variations in real interest rates. We have seen that real yields – and for that matter nominal yields – react to a variation in the dollar index. As previously mentioned the yield on TIPS is a market-based estimate of real interest rates. These market-implied interest rates can significantly diverge from the equilibrium interest rate“, the so-called “natural rate of interest”. The latter has been defined in 1898 by Swedish Economist Knut Wicksell in 1898 as the interest rate that would prevail in the absence of “monetary manipulations”. Wicksell’s Interests and prices pioneering study has been translated to English only in 1936. In the meantime it heavily influenced the Austrian School of Economics. The natural rate of interest often confused with the “neutral interest rate”, a close but distinct concept that is related to the arbitrage between inflation and unemployment (Philipps curve). The natural rate of interest is an elegant concept. John Williams, the former President of the Federal Reserve of New York has extensively written on what he calls R-Star. But so far attempts to provide a meaningful measure of that concept have proven illusive to say the least.