Insight

Uncommon truths: Could gold reach US$7,000?

Uncommon truths: Could gold reach US$7,000?

Is gold cheap or expensive?  Its own history suggests the latter but a return to some form of gold standard could imply a price above $7000 (though could come with holding restrictions).  We have recalibrated our own gold model (using a presidential dummy) and it suggests a current fair value of $1613.  The premium to our model seems reasonable given current uncertainties.  

We have considered gold to be expensive for some time (though we did go Overweight earlier in the crisis). To see why, look at Figure 1 which shows how many barrels of WTI can be bought with an ounce of gold.  The data is annual, so we lose the inter-year detail, but the chart shows that at 51 barrels of WTI, gold has never been so expensive (not in the last 150 years, at any rate).  That ratio has been boosted by the weakness of oil and was even more extreme when WTI temporarily went negative, at which point the calculation ceased to make sense.   

Looking at the price of gold in the traditional way reveals that it has been higher, with a peak of $1898 in September 2011 (based on daily data), versus the current price of $1732.  However, when putting the price in real terms (using US CPI), there were only two periods when gold has been higher than today (based on annual data using year-end levels).  In 1979 and 1980 the level was $1856 and $1831, respectively, while in 2011 and 2012 it was $1791 and $1853.  The annual average since 1833 is $605 (including 2020) but this may be overly harsh as the price was effectively controlled until 1971, since when the average has been $939. 

Of course, we live in extraordinary times, with government deficits once again exploding.  Rising debt is one reason to fear for the stability of the financial system, with the risk that government actions could eventually debase the value of their currencies.  This is even more so if we consider that the aggregate balance sheet of the Fed, ECB, BOE, SNB and BOJ will have increased seven-fold in the 15 years to the end of 2021 (according to our projections).  Did anybody foresee this at the turn of the century? 

If these are the sorts of concerns that encourage investors to buy gold, what is the end point?  How high could gold go in a catastrophic scenario?  One way to think about this is to imagine an outcome so bad that policy makers opt for a return to a form of gold standard.  What would this mean? 

According to the US Department of the Treasury, on 30 April 2020 the US government held 261.5 million ounces of gold.  At a price of $1732 per ounce, that gives a market value of $453 billion.  What if that stock of gold had to back the US currency?  If it were to replace the monetary base ($4.85 trillion in April 2020), that would imply a gold price of around $18,500.   

That is dramatic but is perhaps an overstatement because a large part of the monetary base is the reserves held by banks at the Federal Reserve system (which has increased rapidly since the launch of QE).  If gold is only required to back notes and coins in circulation ($1.89 trillion in April 2020), then the implicit price of gold becomes around $7225. 

Taking a more global view, the World Gold Council reckons 33,919 tonnes of gold were held by official entities at the end of 2019, with a current market value of $1.9 trillion (there are 32,150.75 ounces of gold in a tonne).  If those official holdings were used to back notes and coins in circulation (around $8 trillion), the value of an ounce of gold would be approximately $7336 (based on Bank for International Settlements data showing there was $6.7 trillion of cash in circulation at end-2018 in economies accounting for 85% of world GDP, using World Bank GDP data). 

Before getting too excited, remember that those official holdings represented only 17% of total “above ground” gold holdings (based on World Gold Council data).  We doubt that if a gold standard were reintroduced private citizens and entities would be allowed to freely hold gold.  For example, there were restrictions on the holdings of gold in the US between 1933 and 1974 and Federal law currently allows for the confiscation of gold bullion.  If we include all “above ground” gold in the calculation, the value of gold falls to $1259 or $2378 if we exempt jewellery from the confiscation (jewellery accounts for 92,947 of the 197,576 tonnes of the “above ground” total). 

As the answer appears to be whatever we want it to be, what about an analysis more grounded in market behaviour?  We first introduced our gold model in 2016 (in a document called What price gold?).  This was a simple regression model that explained the movement of gold by changes in three variables: the real 10-year US treasury yield, the 10-year US inflation breakeven and a trade-weighted US dollar index.  We noted an apparent change in behaviour around 2007: before that date, gold increased as inflation expectations rose (see the pre-2007 model in Figure 2), whereas afterwards it declined (the coefficients on the other two explanatory variables were always negative).  Our interpretation was that buyers of gold were more fearful of deflation than inflation after 2007. 

However, the model ran into problems at around the time that President Trump was elected (since November 2016, the actual price of gold was well above the model predicted value).  We have now re-run the model and find that the best statistical fit (R-squared of 0.89) is given by a version that includes a dummy variable that was switched on in November 2016 (the “president dummy” shown in Figure 2).  Importantly, all the coefficients are still negative, so that gold goes up when real yields, inflation expectations or the dollar decline (the coefficient on the dummy variable is positive and suggests gold has been boosted by around $230 during this presidency). Buyers of gold appear to remain more concerned about deflation than about inflation. 

As seen from Figure 2, gold is above the current model predicted value of $1640.  Hence, gold would appear to be slightly expensive based on this model but not inordinately so, considering the uncertainties that exist.  Threats to gold would come from a rise in real yields, inflation expectations and the dollar (and, maybe, a change of US president in November).  

Unless stated otherwise, all data as of 29 May 2020. Paul Jackson is Global Head of Asset Allocation Research with the Global Market Strategy Office. András Vig is Multi-Asset Strategist with the Global Market Strategy Office.