Peak Gold?

The sharp decline in gold prices last week has triggered heated debate among market participants.

The bulls argue that this is just another dip within a secular bull market. They say that poor liquidity and CTA stops in futures markets primarily drove the reversal. Meanwhile, the more consequential demand from ETF investors is resilient and the physicals/ETF premia remains wide.

But others are more cautious. They argue that real rates, which gold prices have tracked, cannot keep falling forever. If a COVID-19 vaccine is found early and inflation continues to recover, DM central banks may become less enthusiastic to add accommodation. 

This debate over tactical/cyclical drivers is missing the big picture. Gold is up 32% YTD. At this point, the debate is really about the long-term drivers, namely valuations vs asset allocation. 

Valuations: Adjusted for inflation, gold prices are very close to previous cycle highs.

• In real terms (i.e. adjusted by US headline CPI), the price of gold is close to 8, similar levels at which it peaked in 1980 and 2011 (Chart 1). The real price high in 1980/2011 was 8.5 and 8.1 respectively, which in today’s nominal terms would translate to $2200/oz and $2100/oz. 

Chart 1

• The highs in 2011 were made almost 10 months before real rates bottomed and ETF holdings peaked. In other words, prices were stymied by valuations well before the cyclical drivers turned.
• In short, gold prices have limited upside, and significant downside from here.

Chart 2

Asset allocation: A permanent increase in portfolio allocation to gold is underway. 

• ETF demand is rising because portfolio investors have a fundamental need for a money-debasement hedge. This demand is indifferent to valuation and is larger in today’s NIRP world than it was in 2011’s ZIRP world.
• Gold ETF ownership volume is a significant driver of gold prices. Every 1mn oz increase in ETF holdings is roughly equal to a $15 increase in the gold price.
• While absolute ownership levels are historically high, ownership relative to stock/bond market cap or cash deposits has scope to increase (Chart 3 shows gold ETF market cap as a percentage of Fed, ECB, BoJ and PBoC total assets).

Chart 3

• As such, ETF investors will continue to drive the value of gold higher. Past cycle highs are an irrelevant benchmark.

Which side is correct? Price is a function of demand/supply. Covid-19 severely disrupted both demand and supply for gold. According to data from the World Gold Council, Jewellery fabrication demand in Q2 fell by half, and physical bar/coins fell by a third relative to pre-Covid-19 levels. The latter has made sensational headlines for its rise in Germany and US, though this was more than offset by falling demand in the Asian giants (China, India and Thailand). Of course, ETF demand made up for all of this. But would Asian consumer demand recover to pre-Covid-19 norms at this price level? Meanwhile, supply variables are just as important as demand. Mine production, down 10% YoY in Q2, will almost certainly recover as the largest gold-producing nations appear to be getting the upper hand on the outbreak. Recycled gold, which is typically responsive to an increase in prices, surprisingly fell 10% in Q2. Once supply is fully operational, producers will have every incentive to go into overdrive to make up for lost revenue.

Table 1

Conclusion

I have sympathy for the valuation (i.e. bearish) camp. However, gold price tops tend to form over several months. And prices could overshoot further until demand/supply variables normalize from Covid-19 disruptions. As such, I do not think the bear view is actionable right now. I would rather study how gold’s beta to real rates and ETF flows evolves. Marginal new highs are possible, but if we see diminishing returns, it would be a sign that supply is ramping up and valuations matter after all.