Gold mining supply varies considerably over time. Gold mining production peaked at around 53 million troy ounces per year between 1963 and 1965, and subsequently shrank to about 34 million troy ounces by the end of the 1970s.  Then, from 1980 to 1999, gold mining supply soared from 34 million to 85 million troy ounces.  Mining output then contracted to 73 million troy ounces by 2008 before expanding again to 102 million by 2016.  Since 2016, mining production of gold has contracted to 94 million troy ounces by 2019. 

The variations of gold supply have been strongly and inversely correlated with gold prices.  The tremendous run up in gold prices during the 1970s, which took gold from $35 to $835 per troy ounce is commonly attributed to macroeconomic factors including the breakdown of the Bretton Woods system of fixed exchange rates, the weak dollar and the roaring inflation of the 1970s.  Undoubtedly, these demand-side factors played a significant role in the gold bull market.  The USD had been pegged to gold at $35 per troy ounce from 1933 to 1971 when considerable inflation had eroded USD’s purchasing power.  Even so, USD’s 32% slide versus foreign currencies between 1971 and 1979 and the cumulative impact of the 543% rise in consumer prices between 1933 and 1980 can’t by themselves explain why gold prices rose 2,285% between 1971 and 1980.  The 36% drop in mining supply, however, contributes to the explanation of gold’s extraordinary bull market.

Likewise, gold’s decline from $835 in 1980 to $280 by the late 1990s is usually attributed to a generally stronger dollar and the Federal Reserve’s (Fed) high real rates of interest, which helped to lower inflation from 13.5% in 1980 to around 2.5% by the late 1990s.  While the changing macroeconomic policies of the 1980s and 1990s almost certainly played a major role in gold’s bear market, the fact that mining supply soared from 34 million to 85 million troy ounces per year between 1980 and 1999 also had a significant role in depressing gold prices. 

The gold bull market from the late 1990s until 2011, which sent the price from $280 to almost $2,000 per troy ounce, is often attributed to a sharp weakening of the USD from 2002 to 2011 as well as Fed funds rate moving from 6.5% in 2000 to 0.125% in the aftermath of the 2008 financial crisis. During the early years of this bull market, gold prices may have been supported by a 15% drop off in annual mining supply. 

Gold prices peaked in 2011, and then began to slide.  This decline is sometimes blamed on the Fed announcing the end of quantitative easing (QE) in 2013 and winding it down by 2014 as well as the Fed eventual lifting rates from 0.125% to 2.375% between late 2015 and late 2018. This period of generally falling gold prices also coincided with a sharp increase in gold mining output from 73 million to 102 million troy ounces.  Mining supply may have played a larger role in depressing gold prices than the relatively minor changes in U.S. monetary policy. 

Finally, gold’s run-up in price from $1,200 to $2,080 per troy ounce from early 2019 to mid-2020 was demand driven.  Mining supply barely changed while the Fed slashed rates to zero and began a $4.9 trillion QE program which included printing $3.0 trillion in March and May 2020.  When viewed over a long period of time, the inverse relationship between gold mining supply and the inflation-adjusted price of gold is strongly apparent (Figure 1).

Unlike mining supply, which appears to be a major driver of prices on a year-to-year basis, secondary supply does not appear to influence prices much at all.  In fact, the opposite appears to be the case: secondary supply is influenced by the level of prices.  Higher prices incentivize more recycling.  Lower prices tend to disincentivize it.  In any case, the market treats secondary recovery as gold that already exists, whereas mining supply is seen as gold that is new to the market.  While gold mining supply has a strongly negative correlation with the year-on-year change in the real price of gold, secondary supply tends to have a positive correlation (Figure 2).   

Central banks hold 35,000 metric tons of gold, which equates to about 20% of all of the gold that has ever been mined in human history.  Their reserves also equate to about 12 years’ worth of the gold mined in 2022.  Over the past several decades, central banks have alternated from being net sellers of gold (1977-78 and 1982-2008) to being net buyers of gold (1979-81, and since 2009).  When central banks are net sellers, they are effectively increasing supply in the market with gold that would not otherwise have been available for private use.  By contrast, when central banks are net buyers of gold, they absorb supply and lock it away in vaults.  Their annual buying and selling of gold can make for noticeable additions and subtractions from primary and secondary gold supplies (Figure 3).  

Central bank buying also appears to be strongly correlated to price (Figure 4).  Net central bank purchases of gold coincided with gold’s 1979-81 price peak.  Net purchases by central banks have also remained robust since the global financial crisis, and is a likely factor supporting gold prices since 2009.  

Gold mining supply appears to be on track to grow by 2% or so in 2023.  With gold trading near $1,950 per troy ounce currently and all-in sustaining costs of around $1,212 for the world’s gold mines, profit margins at gold miners remains robust at over 50%.  Whether the miners’ generally large profit margins translate into increased investment and increased mining output, however, remains to be seen.