The Golden Constant
Considering Gold's Role in a Portfolio
By Igor Vasilachi
The famed legions of Rome were commanded by a junior officer corps of centurions, a role roughly equivalent to a captain in today’s US Army. In the age of Augustus, centurions earned a stipend of 3,750 denarii, roughly 38 ounces of gold. Today, in the age of Biden, the average US Army captain earns a salary of about $70K—40 ounces of gold at current prices. Over two millennia, the same weight of gold buys a roughly equivalent amount of labor, preserving real purchasing power despite the rise and fall of empires and wild technological change.
In more recent history, gold has maintained this constant relationship with the price of labor. The chart below shows median personal income in the US priced in gold between 1974 and 2019.
Figure 1: US Median Personal Income in Gold Ounces (1974–2019)
Source: FRED
Inflation may destroy the value of fiat currencies, but gold has remained a constant store of value over long periods of time.
But how should investors think about the role of gold as a financial asset? What role should it play in a portfolio of stocks, bonds, and cash? And how does the price of gold vary with the business cycle, or the macroeconomic environment?
Let’s start by comparing gold to stocks and bonds. Below we show the ratio between gold and 60/40 portfolio, the latter defined as 60% S&P 500, 20% investment-grade bonds, 20% 10-year US Treasurys. We also highlight recessionary periods, as defined by NBER.
Figure 2: Gold Divided by 60/40 Portfolio (1970–2020)
Source: Bloomberg, FRED
Investors see gold as insurance against redenomination risk, so gold’s value is dependent on the level of fear about currency instability. We detailed this relationship with inflation in our recent research article on inflation, where we show that gold is a top performer in inflationary environments, with an inflation beta of 2.4 (i.e., for every 1% rise in inflation, there is a 2.4% increase in gold price). During the period before 1990, when average inflation was 6.3% and even reached a 14% high in 1980, gold was worth 3.2x the 60/40 portfolio on average. In the period since then, when inflation was tamed to an average of 2.4%, gold averaged 0.6x the 60/40 portfolio.
Although inflation was always an important driver of gold returns, this relationship weakened after 1990 as inflation was falling. Below we show gold returns correlation to inflation, real interest rates, and real GDP growth rates.
Figure 3: Gold Returns Correlations (1970–2020)
Source: FRED, Bloomberg
Post-1980, real interest rates and real GDP growth became equally important drivers of gold returns. For example, while gold was 3x more sensitive to changes in inflation than to changes in real GDP growth pre-1980, it became equally sensitive to both since 1980.
In addition, gold has tended to have an inverse relationship with GDP: in recessionary times, when GDP falls and high-yield spreads rise, gold prices soar. In an average recession, gold appreciates 1.3x relative to the 60/40 portfolio.
Gold thus can play an important complementary role in portfolio allocation, providing returns when traditional assets like stocks and bonds are suffering either from high inflation or from recessions.
What Is the Best Way to Invest in Gold?
Investors generally have three options to build exposure to gold: holding the metal physically, which comes with the additional shipping and storage complexities, investing in rolling future contracts, and holding “gold stocks” (i.e., companies mining or trading gold, or otherwise exposed significantly to the metal). It is worth noting that there are managers offering physical gold ownership, such as Sprott, which charges an additional 0.65% fee on top of roughly 1% in annual associated holding fees payable to the US Mint.
Below we show the performance of these three options starting in 1975, the first year gold futures were issued for trading, as well as the 60/40 portfolio.
Figure 4: Performance Indicators for Gold, Gold Futures, and Gold Stocks (1975–2020)
Source: Bloomberg, FRED, Ken French Data Library
In all three cases, a buy-and-hold approach to gold would have delivered mediocre returns coupled with very high volatility and more than double the drawdowns of owning the 60/40 portfolio over the period of this analysis. Investors might therefore consider alternatives to buying and holding gold, such as implementing a “timed” allocation strategy.
It is worth noting that gold futures nearly perfectly track the price of gold after accounting for the fees required to hold it physically. Meanwhile, gold miners have a roughly 70% correlation and a 1.4 beta to gold futures returns, making them a leveraged (and volatile) bet on gold, with a 25% correlation with the S&P 500 (compared to virtually 0% for gold futures). We believe investors who are owning gold primarily to diversify their equity portfolios should prefer their gold exposure through futures.
Timed Allocation Strategy
Our research so far has revealed that, on the one hand, gold seems to have done well when stocks and bonds have performed poorly, especially in inflationary environments. On the other hand, it has had mediocre returns and extreme drawdowns (close to 80%) historically. To improve returns, then, investors seeking exposure to gold could try to (a) estimate the direction of inflation to capture the gold rally while (b) setting up downside protection mechanisms from major drawdowns.
In our previous research, we have shown that the high-yield spread and the slope of the yield curve can help investors predict the broad movements of the business cycle. Below we compare the forward returns of gold futures based on different starting conditions for the yield curve and the high-yield spread. On the left, we show the forward returns of gold based on the slope of the yield curve and, on the right, we show the forward returns of gold based on the trailing three-month change in the high-yield spread.
Figure 5: Annualized Average Gold Futures Returns by Quartiles of Term Slope and Change in High-Yield Spread (1975–2020)
Source: FRED, Bloomberg
Historically, gold has done well in periods when the slope of the yield curve (left table) was at its extremes (quartiles 1 and 4). In quartile 1, gold has been a top performer when the curve is flat or negative (i.e., long-term rates are equal to or lower than short-term ones). This tends to occur at business cycle peaks, when the economy is overheating and on the verge of entering a recession.
In quartile 4, gold tends to do well when the slope of the yield curve is very steep (i.e., long-term yields are significantly higher than short-term yields), which tends to happen deep into recessions when markets panic and the Fed intervenes with a sudden rate drop. These are also the periods when the high-yield spread experiences sudden and extreme increases when markets panic. In congruence with the slope of the yield curve, gold has done best in the fourth quartile of high-yield spread changes.
All of the above suggests that gold has been a top performer when investors needed it most: in late-cycle stages when equities and corporate credit tend to do poorly.
These insights help us understand when and why gold works in a portfolio. But as we showed in Factor Investing in Commodities, gold is a highly trending asset (t-stat of 3.1 and 5.4 for three- and six-month returns).
Below we compare annualized returns of gold and the 60/40 portfolio by gold’s price level relative to its 200-day moving average.
Figure 6: Gold Futures Annualized Returns by Spot Price Relative to 200-Day Moving AVG (1975–2020)
Source: FRED, Bloomberg
These results support a strong case for trend-following in gold. We sought to test a simple trend-following strategy: if the gold spot price is above its 200-day moving average, long the gold futures; otherwise, hold alternatives (e.g., 10-year Treasurys).
Below we show the comparative results by strategy, with 10-year Treasurys as the alternative asset class.
Figure 7: Performance Indicators by Strategy (1975–2020)
Source: FRED, Bloomberg
Applying this rule to gold futures would have improved returns historically, while halving drawdowns. All that, while being invested in futures only 43% of the time. This creates opportunities to further improve returns by allocating to other assets in the remainder of the time or by incorporating a trend-followed gold strategy into a broader portfolio, as shown below.
Figure 8: Performance Indicators by Strategy (1975–2020)
Source: FRED, Bloomberg, Ken French Data Library
Is Now a Good Time to Invest?
Fears of inflation are dominating the market, so investors are asking themselves if gold can help them navigate these dangerous waters. The big-picture question is whether inflation and monetary instability will return to 1970 levels, in which case it would not be unreasonable to think gold could be worth ~5x its current price relative to the 60/40 portfolio.
But over the shorter term, two of the three indicators we looked at are suggesting that now is not the ideal entry point.
The slope of the yield curve, which we compute as the difference between 10-year and 1-year Treasury yields, has reached its 10-year trailing median in Q1 2021, following a steady increase since 2019, suggesting inflation expectations might already be priced in.
Figure 9: Slope of the Yield Curve vs. 10Y Trailing Median (2018–2021)
Source: FRED
The high-yield spread has been narrowing: it is both below its 10-year trailing median and its level three months ago. In addition, real rates have been rising and real GDP is rising in the aftermath of the pandemic. None of these are positive signs for gold.
Figure 10: High-Yield Spread vs. 10-Year Trailing Median (2018–2021)
Source: FRED
On the other hand, the gold spot price has risen above its 200-day moving average at the end of May, even if it was below that level as early as March 2021. This suggests investors could benefit from a price trend in gold.
Figure 11: Gold Price vs. 200-Day Moving Average (2018–2021)
Source: FRED
The macro signals—slope of the yield curve and high-yield spread—point to being out of gold. Trend-following signals, however, suggest investors might do well to ride the wave. These conflicting signals warrant caution from investors. In their recent paper, Erb, Harvey, and Viskanta argue that gold’s real price level might have peaked recently. They show that gold has tended to draw down significantly after real price peaks in 1980 and 2011 (by 55% and 67%, respectively) and suggest now may be the case too.
By March 2021, gold prices have dropped 20% from their August 2020 peak before moving back up to nearly that level. If the thesis of Erb et al. holds true, however, and gold does move towards its long-term ratio to the median average income (i.e., 31 ounces of gold), that might translate to another 80% drawdown from today’s price and back to its Q4 2018 price levels.
Conclusion
We believe gold can be an important tool for investors. It can help them generate returns when they are needed most, in recessions and inflationary periods, and it is uncorrelated with equities. However, gold is also volatile and has experienced massive drawdowns historically. Therefore, investors must time their exposure to it. Today, our gold signals are conflicting: both macro indicators suggest investors should be out of gold, while the trend supports gold allocation.