Gold makes new all-time amidst central bank buying - Research

Gold is at an all-time high (see Exhibit 1). This is against all odds given the level of interest rates. We have written extensively in the past how gold prices are driven by three main factors: Central bank policy (Real-interest rate expectations and QE), Longer-dated energy prices and central bank net buying of gold. For readers unfamiliar with our gold price framework, they can catch up here (Gold Price Framework Vol. 2: The energy side of the equation, May 28, 2018, here (Part II, 10 July 2018) and here (Part III, 24 August 2018), as well as some follow up reports that built on the model (Gold Price Framework Update – the New Cycle Accelerates, 28 January 2021), (Gold prices continue to weather the rate storm, 13 April, 2022 and Gold prices reflect a shift in paradigm – Part I and Part II, March 15 & 16, 2023)

Exhibit 1: gold prices are at an all-time high

$/ozt

Source: Goldmoney Research

More recently, our gold price model stopped working. To be precise, gold prices are much higher than what our model predicts (see Exhibit 2). While the rise in real-interest rate expectations would have produced price declines, gold has been on a steep upward trend.

Exhibit 2: Gold prices are much higher than what our model predicts

$/ozt

Source: Goldmoney Research

Does that mean that the drivers we identified are no longer important? The answer is clearly no. When we analyze the month-over-month changes in the gold price vs the changes in the predicted prices[1], we see that both still follow a similar path (see Exhibit 3). However, the predicted changes tend to underperform the changes in the realized gold prices both when they are positive and negative.

Exhibit 3: The month-over-month changes in our predicted prices and realized prices still follow a similar path

% change

Source: Goldmoney Research

So, how can it be that the three drives we identified still follow short-term changes in the price but cannot explain the current levels? We have analyzed this in the past and we remain confident that the explanations we proposed are valid (see “Have central banks lost control over the gold price?”, October 11, 2023). More specifically, we concluded that:

“Despite abandoning the gold standard decades ago, gold prices still largely reflected central bank actions since then. Arguably western central banks so far were not primarily concerned about the price of gold in their respective currency, but they do try to control some of the factors that also drive gold prices. To illustrate this with the three main drivers for gold prices we identified in our model (Central bank net gold purchases/sales, real-interest rate expectations and longer dated energy prices):
Until recently, it was mainly western central banks that bought and sold large quantities of gold (mostly sold).
Central banks set interest rates, which impacted real-interest rates expectations. Central banks also control how much money is in circulation, which impacts inflation and inflation expectations. QE was just another form of manipulating interest rates (and we argue it is behind the surge in inflation).
On a long enough timeline, longer dated energy prices reflect mostly inflation.
In other words, markets understood that central banks and their policies were ultimately behind rising gold prices. But importantly, the market also believed that central banks had the power to reverse the impact of their policies and implicitly had the power to bring gold prices down. Our model shows that over the last 20+ years, sell-offs in gold prices were mainly due to falling longer-dated energy prices and / or rising real-interest rate expectations. The fact that the recent massive rally in real-interest rate expectations – which was entirely driven by central bank rate hikes – was just shrugged off by the gold market, suggests that the power of western central banks, and particularly the Fed, to implicitly control gold prices, has dramatically decreased, if not vanished.
To go one step further, the above interpretation may look at things the wrong way around. One can argue golds reluctance to react to the Fed rate hikes is just a symptom of a much larger issue. Rather than having lost control over gold prices, the Fed may have lost control over the US dollar itself. Gold priced in US dollars simply reflects that. And the Fed is not alone, all central banks are facing the same issue.”

In a nutshell, in our article from October 2023, we argued that gold was rising amidst rising-real interest rates because western central banks had lost control over the price and investors added gold to their portfolios despite rising real-interest rate expectations. However, while we focused mainly on how private investors had shifted their behavior, we think it’s equally, and maybe even more important, to look at sovereign demand for gold in the form of central bank net purchases and sovereign wealth fund purchases.

Unfortunately, non-OECD central banks don’t publish their holding regularly and accurately. In some cases, we believe they are outright hiding their true holdings. This makes it extremely difficult to run statistical models, as we simply don’t have reliable historical data. Our original model uses central bank net purchases as an explanatory variable. However, historically, changes in central bank holdings were primarily driven by Western central banks which report their holdings accurately and timely.

It is undeniable that we have seen a shift in paradigm in central bank demand starting around 2022. Up to that point, non-OECD central banks have been small net buyers of gold while OECD central banks have done largely the opposite. But since 2022, we believe central bank net purchases from non-OECD countries have sharply accelerated.

There are multiple reasons for this. One of the reasons is that many non-OECD countries became concerned about their foreign reserves when Western sanctions targeted the foreign reserves of the Russian central bank in the aftermath of the Ukraine invasion.

The Russian central bank had already declared - prior to the Ukraine invasion and the subsequent financial sanctions imposed by the West - that it no longer held any dollar denominated assets. But the EU sanctions and confiscation of Euro denominated assets meant that any sovereign debt issued by a Western country was no longer an option. While the Russian central bank itself - due to the struggles of the Russian economy - probably wasn’t able to add a lot more gold to their reserves since then, the entire story has rattled leaders and central banks across the globe. All of a sudden, foreign currency reserves were not just at risk of losing value due to inflation (more about that later), they also were at risk of being confiscated. According to a 2023 Financial times article, a survey of 57 central banks and 85 sovereign wealth funds showed that “many sovereign investors were “concerned” by the precedent set by the confiscation of Russian assets, with 96 per cent saying further investment in gold was driven by its status as a safe haven”. We believe this is one of the reasons why non-OECD central banks accelerated their gold purchases since 2022.

Another reason why non-OECD central banks are increasing their gold allocations is the sudden emergency of the inflation spectre. After decades of decline inflation, seemingly at odds with endless money printing and an explosion in sovereign debt, suddenly inflation came back in 2022 in a roaring fashion (see Exhibit 4). The world hasn’t seen this kind of inflation in hard currencies since the early 1980s. For emerging markets, this must have come as a shock. Non-OECD nations are sitting on trillions of dollars in sovereign debt in hitherto seemingly inflation proof currencies.

Exhibit 4: Inflation came back roaring after declining for decades

Source: FRED, Goldmoney Research

Arguably, the hard pivot by the Fed and other central banks has brought down inflation to less scary levels (see Exhibit 5).

Exhibit 5: The Feds hard pivot brought inflation down for now

Fed Funds rate %

Source: FRED, Goldmoney Research

But the cost of that is that OECD budgets face steep increases in debt servicing costs.

In the US, federal debt servicing costs went from USD500bn per annum in 2020 to over USD1000bn per annum by the end of 2023 (see Exhibit 6) and even if rates decline from here, debt servicing costs are set to rise further as more debt matures. At the same time, budget deficits have exploded.

Exhibit 6: US federal debt servicing costs have exploded

$bn per annum

Source: FRED, Goldmoney Research

This has major consequences for national debt levels. The US federal debt is now increasing by $1tn every 90 days (see Exhibit 7). Central banks and sovereign wealth funds must be observing this with growing concerns. And while so far, we haven’t seen a major exodus out of sovereign bonds – with all the earth-shattering consequences this would have – many central banks and probably quite a few sovereign wealth funds feel compelled to at least diversify a bit into gold.

Exhibit 7: US federal debt is on a sharp upward trajectory, adding $1tn every 90 days and accelerating

$bn

Source: FRED, Goldmoney Research

Gold purchases by central banks and sovereign wealth funds are – unlike private investors – much less interest rate sensitive, if at all. Private investors view gold as attractive when real-interest rate expectations are low but feel less drawn to gold when real-interest expectations are high. 10-year TIPS yields are currently at 2%, that’s a sharp increase from -1.5% just two years ago (see Exhibit 8).

Exhibit 8: Real-interest rate expectations rose sharply over the past two years

%, 10-year TIPS yield

Source: Goldmoney Research

This move is the reason why our gold price model predicts a price decline rather than an increase over that timeframe. And in fact, gold holdings in ETFs are down 20% from their peaks (see Exhibit 9), in line with our expectations. Arguably, part of this might be sovereign wealth funds selling their ETF holdings and buying physical gold to be held within the country, but it’s undeniable that the relationship between real-interest rate expectations and investor demand for gold continued to hold at least in ETFs.

Exhibit 9: ETF gold holdings are down 20% from their peaks

Ozt

Source: Goldmoney Research

In stark contrast to private investors, central banks tend not to worry that much about the real return on their assets. They only worry whether the assets are safe from default and devaluation, and more recently also from confiscation. And to some extent that is probably also true for sovereign wealth funds. They are more concerned about preserving wealth for future generations rather than earning a market return.

So, how does this help explain why the monthly changes predicted by our model still move on a similar path to realized changes, but the model fails to explain current levels? We think central bank gold demand is like a level change from 2022 onwards. Underneath that steady demand, the other drivers still play their roles. Unfortunately, because of the unreliable nature of the data for non-OECD central bank gold holdings, this can’t be properly quantified.

More importantly, it’s anybody’s guess how central bank purchases develop going forward. There is the risk of a substantial correction should the trend pause or even reverse, should non-OECD banks need to reduce their asset holdings due to a recession for example. On the other hand, it is hard for us to see how – absent any crisis – this trend should not continue. OECD sovereign debt is unlikely to become more appealing than it is now and there is no other currency that can fill that void as the Chinese Yuan is simply not an alternative at this point. At the moment, long-dated interest rates are on the rise again as the market re-evaluates the likelihood of interest rate cuts by the Fed. But that increases the probability of a hard landing, which then in turn, would inevitably lead to larger rate cuts. Once that happens and real-interest rate expectations decline again, private investor demand will also re-accelerate. This is when gold prices will really start to take off.

[1] The chart shows the month-over-month changes of the original multivariate regression analysis rather than the results of separate multivariate regression analysis of the month-over-over month changes in price.