CICC: Is gold really expensive?

Wallstreetcn
2025.01.15 00:50
portai
I'm PortAI, I can summarize articles.

CICC updates its gold pricing model, analyzing the relationship between gold and U.S. Treasury yields. Despite high U.S. Treasury yields, gold prices have risen against the trend, with returns exceeding 30%. The model indicates that gold is not overvalued and expects a significant increase in gold prices in 2024. The updated model extends from 1971 to 2024, excluding U.S. Treasury yields, and emphasizes the impact of central bank gold purchases and U.S. debt expansion on gold prices

Update on Gold Pricing Model - Focus on Long-Term Real Prices

U.S. Treasury rates have remained high over the past year, recently spiking again, while gold has risen against the trend, with returns exceeding 30% during the same period, further validating the decoupling logic between U.S. Treasury rates and gold.

Chart 1: The decoupling of U.S. Treasury real rates and gold prices has continued since 2022

Source: Wind, CICC Research Department

In the article "Can Gold Still Be Bought?" published in January 2024, we constructed a four-factor cointegration regression model for gold, suggesting that the divergence between gold and U.S. Treasury rates mainly reflects the impact of central bank gold purchases and the expansion of U.S. debt. Gold is not significantly overvalued, and we predict that gold prices may rise significantly in 2024. Entering 2025, we found that last year's increase in gold prices has exceeded the explanatory power of the current model, with model residuals reaching historical highs.

Chart 2: The residuals of the four-factor cointegration regression model (January 2024 version) have exceeded $600, significantly reducing the model's explanatory power for gold prices

Source: Wind, CICC Research Department

Note: The four factors include the scale of U.S. public debt, central bank gold purchases, the real yield on 10-year U.S. Treasuries, and the U.S. dollar. For details, see "Can Gold Still Be Bought?" published in January 2024, where the central bank gold purchase scale for Q4 2024 is predicted based on the average value of the first three quarters, as official data has not yet been released.

To better analyze and predict gold price trends, we are making the following updates to the gold pricing model:

1) Extend the time scale: The previous model was constructed based on data from 2003-2023; however, the macroeconomic environment has changed since the pandemic, and the asset operating patterns during the low inflation period of the past 20 years may not apply to the next 20 years. We have extended the model fitting period to 1971-2024, incorporating the "Great Inflation" period of the 70s and 80s and the "Great Moderation" era after the 90s. Due to the multiple changes in the economic environment during the model fitting period, the regression coefficients are more likely to capture the long-term asset patterns that transcend cycles, excluding some short-term data patterns that may not hold 2) Excluding U.S. Treasury Rates: The negative correlation between the real yield of 10-year U.S. Treasuries and gold has been very strong over the past 20 years, but prior to 2000, the relationship was actually unstable, and the regression coefficient was statistically insignificant. Considering recent market trends, we believe that the negative correlation between gold and U.S. Treasury rates may be a short-term phenomenon. Therefore, in the long-term model, we exclude U.S. Treasury rates from the explanatory variables and only use three factors: central bank gold purchases, the scale of U.S. debt, and the U.S. dollar to explain gold prices.

Chart 3: The price of gold and the real yield of U.S. Treasuries only show a strong negative correlation from 2005 to 2022, but over a longer period, the correlation is not high.

Source: Wind, CICC Research Department

Chart 4: The connection between the real price of gold and the real yield of U.S. Treasuries is also not strong.

Source: Wind, CICC Research Department

3) From Nominal Prices to Real Prices: We found that the long-cycle three-factor gold pricing model struggles to find a stable cointegration relationship, reflected in the fact that the residuals have remained positive over the past 20 years, with the mean not converging to zero. Referring to relevant academic literature, we use a new model framework to decompose nominal gold prices into real gold prices and inflation. Gold has "anti-inflation" properties, and the rise in gold prices is closely related to rising prices, while inflation can be predicted separately using statistical models (as discussed in "New Perspectives on Inflation Variables and Asset Changes"). We found that real gold prices (we adjust historical nominal gold prices to the price level corresponding to 2024) have a long-term stable relationship with some economic variables, allowing us to construct a desirable statistical model, which will be discussed further below. Finally, multiplying the real gold price by inflation yields the equilibrium price of nominal gold and future forecasts.

Chart 5: The failure of the long-cycle nominal gold price model's cointegration regression method is reflected in the fact that the residuals have remained positive and continue to expand over the past 20 years.

![](https://mmbiz-qpic.wscn.net/sz_mmbiz_png/fzHRVN3sYsib1GbmFSdVgqkuHEn61x31QDEAGsL5k6EDdZoVcYxIKGqZ6iboib87KXFt7TDtiaLZoFIotY7fePhnYg/640? wx_fmt=png&from=appmsg)

Data source: Wind, CICC Research Department

Chart 6: Nominal gold prices are significantly positively correlated with inflation levels, reflecting gold's "anti-inflation" value

Data source: Wind, CICC Research Department

In the long term, current gold prices may not be significantly overvalued

The rise in gold value reflects the decline in the credibility of the dollar system. The imbalance of U.S. debt is a key reason for the decline in dollar credibility, and the central bank's increased gold purchases are a result of this decline. Therefore, U.S. debt and central bank gold purchases are key long-term explanatory variables in the gold model. From the data, these explanatory variables have a higher correlation with actual gold prices than with nominal gold prices, which helps improve the model's explanatory power.

Chart 7: The correlation between central bank gold purchases and actual gold prices is higher than that with nominal gold prices

Data source: Wind, World Gold Council, CICC Research Department

Note: Actual gold prices are derived by dividing nominal gold prices by the U.S. CPI index, with the base period set to November 2024.

Chart 8: The correlation between the U.S. deficit rate and actual gold prices is higher than that with nominal gold prices

Data source: Wind, CICC Research Department

Note: Actual gold prices are derived by dividing nominal gold prices by the U.S. CPI index, with the base period set to November 2024.

The rise in gold value also reflects an increase in safe-haven demand, which is driven by the economic cycle. Since nominal gold prices include inflation value, the long-term trend is upward, making it difficult to establish a statistical relationship with cyclical variables. In contrast, the trend of actual gold prices is less pronounced, more cyclical, and easier to relate to other cyclical variables We found that the lower the GDP, the weaker the PMI, and the higher the real gold price. The correlation between the real gold price and consumer expectations as well as economic policy uncertainty indicators is particularly evident.

Chart 9: The real gold price is negatively correlated with the U.S. economic cycle

Source: Wind, CICC Research Department

Note: The real gold price is obtained by dividing the nominal gold price by the U.S. CPI index, with the base period set to November 2024.

Chart 10: The real gold price moves in the opposite direction to consumer expectations

Source: Wind, CICC Research Department

Chart 11: The real gold price is positively correlated with expectations of economic deterioration over the next five years

Source: Wind, CICC Research Department

From the negative relationship between "expectations of deterioration over the next five years" and gold, it seems that the current gold price does not agree with the market's optimistic expectations for a "soft landing." We attempted various combinations of explanatory variables for the real gold price, and the regression results were similar, with residual levels generally between $300 and $500. Among them, the three-factor model of central bank gold purchases/ETF + deficit rate + University of Michigan consumer expectations had the smallest residual this year, indicating that the equilibrium price of gold by the end of 2024 is $2,230 per ounce, corresponding to a residual of $390 per ounce.

Chart 12: Cointegration regression of gold purchases + deficit + consumer expectations three factors

![](https://mmbiz-qpic.wscn.net/sz_mmbiz_png/fzHRVN3sYsib1GbmFSdVgqkuHEn61x31QtvXs8A8fZa0tlX5jtbVPAGVIiccr018m7kSp6LKEo4ExsMJXwh8pNUQ/640? Data Source: Wind, World Gold Council, CICC Research Department

Chart 13: Central Bank Gold Purchases + Deficit + US Dollar Index Three-Factor Cointegration Regression

Data Source: Wind, World Gold Council, CICC Research Department

Chart 14: Gold Purchases/ETF + Deficit + Real Effective Dollar Cointegration Regression

Data Source: Wind, World Gold Council, CICC Research Department

Although the residuals of the above models are all positive, it does not mean that the current gold price is significantly overvalued: due to the long-term nature of the model, which spans a long period, there have been multiple significant changes in the economic structure, resulting in larger residuals compared to short-term models. From the historical comparison of residuals, the current residual level of 300-500 dollars is significantly lower than the residual levels of the 1970s and 2010s, indicating that the deviation is not extreme. Additionally, since the above models use year-end values of annual frequency data, considering the high volatility of gold prices in 2024, the deviation between gold prices and the model from an annual average price perspective will further narrow.

Chart 15: The Volatility of Gold Prices in 2024 Reaches a Historical High

Data Source: Wind, World Gold Council, CICC Research Department

Chart 16: Using Annual Mean Regression Data, Model Residuals are Smaller

![](https://mmbiz-qpic.wscn.net/sz_mmbiz_png/fzHRVN3sYsib1GbmFSdVgqkuHEn61x31Q1JwZZDJAjqBq2LpmvniafMunYT1Qe6xfeFXlQjibI6gNoX5MOLXJXkWw/640? wx_fmt=png&from=appmsg)

Data source: Wind, World Gold Council, CICC Research Department

Considering both the nominal gold price model and the real gold price model, we believe that although the current gold price is above the model equilibrium price, the pricing is not extreme, and the valuation will not exert significant pressure on the gold price. It is still possible to trade gold based on fundamental macro logic.

In the next decade, the gold center may be between $3,000 and $5,000 per ounce

From the key driving factors of the pricing model (U.S. debt, central bank gold purchases, inflation), gold still has ample room for appreciation in the medium to long term: If Trump wins the U.S. election and the Republican Party controls both houses, more aggressive tax reduction policies may be introduced in the future, combined with the interest costs brought by high interest rates, U.S. debt may accelerate expansion in the coming decades, weakening the credibility of the dollar and supporting gold prices. As of the third quarter of 2024, the proportion of gold in foreign exchange reserves for major Asian economies such as China, India, and Singapore is only about 5%-10%. Turkey has a relatively high proportion of gold reserves, but it is still far below the over 70% proportion of gold reserves held by major central banks in Europe and the United States. From 2022 to the present, the gold purchases of the aforementioned four central banks account for over 60% of the total gold purchases by global central banks, indicating their willingness to actively replenish gold reserves.

Chart 17: The proportion of gold reserves held by Asian central banks is significantly lower than that of European and American countries

Data source: Wind, World Gold Council, CICC Research Department

Our calculations show that if Asian central banks can ultimately increase their gold reserves to levels comparable to those of European and American central banks, it would generate a gold purchase demand of over 30,000 tons, which is six times the current annual supply of gold. Therefore, in response to de-globalization and the decline of dollar credibility, gold purchases by Asian central banks will also drive up gold prices. According to our calculations, the inflation center in the U.S. has risen to around 2.5%-3% post-pandemic. However, if Trump's presidency leads to uncontrollable inflation in the U.S., according to PIIE calculations, inflation could surge above 9% in the next 2-3 years. Regardless of whether inflation is out of control, the continuous rise in prices is favorable for gold performance. We use the previously constructed central bank gold purchases + U.S. deficit rate dual-factor real gold price model, combined with reasonable assumptions about the explanatory variables for gold, to estimate the potential appreciation space for gold prices.

1) Neutral scenario: Given the enormous amount of gold reserves that Asian central banks need to replenish, and the more complex global situation after Trump takes office, along with the declining credibility of the dollar system, we expect the pace of central bank gold purchases to at least maintain the current level, remaining at around 1,000 tons per year. For U.S. debt, we adopt the forecast results under the baseline scenario of the U.S. Federal Budget Committee (CRFB), predicting that the U.S. fiscal deficit will increase by about $7.5 trillion over the next 10 years We also assume that Trump will not lead to uncontrollable inflation, and that the CPI will stabilize at 3% in the future, fitting gold prices to rise to $3,300 per ounce over the next ten years.

2) Aggressive Scenario: Assume that Asian central banks wish to replenish their gold reserves more quickly, accelerating gold purchases to 1,500 tons per year (rapid gold purchases by central banks may drive up gold prices and increase purchase costs; we do not make overly high assumptions about the pace of central bank gold purchases). On the fiscal side, we assume that Trump significantly expands fiscal policy, reaching the upper limit predicted by the CRFB, which is an additional increase of about $15 trillion in the U.S. fiscal deficit, while adopting the PIIE's inflation path forecast for Trump's impact, which predicts inflation will first spike to 9% and then drop back to a 3% midpoint, calculating that the value midpoint of gold will rise to $5,000 per ounce over the next ten years.

Chart 18: The future 10-year gold midpoint may range between $3,000 and $5,000 per ounce

Source: Wind, World Gold Council, CRFB, CBO, CICC Research Department

Gold can still be overweighted, but the increase may slow down in 2025, and volatility may increase

According to the results of the above model, gold's short-term valuation is reasonable, and there is still ample room for long-term appreciation, so valuation is not a constraint for further increases. Over the past two months, gold has experienced increased volatility, which may primarily be due to a decrease in "Trump trade" risk aversion sentiment and easing geopolitical risks. Looking ahead to 2025, we expect Trump's policies are likely to lead to "uncontrollable inflation" or "economic stall," both of which are favorable for gold performance. Although the market's expectations for the Russia-Ukraine and Middle East situations have become more optimistic, we caution that the risks of Trump's foreign policy leading to other geopolitical issues should not be overlooked. Under the trend of de-globalization, more funds may flow into safe-haven assets, and gold is expected to perform well. Meanwhile, due to the significant increase in gold prices in 2024, there has been some positive residual accumulation, which presents a certain mean reversion pressure; gold may experience increased volatility in the short term, and the increase in 2025 may be lower than in the past two years.

Therefore, we recommend maintaining an overweight position in gold, moderately downplaying the short-term trading value of gold, and focusing on its long-term allocation value.

December U.S. CPI may be high, pay attention to risk control

The U.S. CPI for December will be announced on January 15 (Wednesday). CICC's macro asset model predicts a month-on-month increase of 0.41% for the U.S. nominal CPI (consensus expectation 0.3%, previous value 0.31%), which is higher than the previous value and market consensus expectation; it predicts a month-on-month increase of 0.25% for the core CPI (consensus expectation 0.2%, previous value 0.31%) Chart 19: U.S. Nominal CPI Month-on-Month Growth Rate Breakdown and Forecast

Source: Haver, CICC Research Department

Chart 20: U.S. Core CPI Month-on-Month Growth Rate Breakdown and Forecast

Source: Haver, CICC Research Department

The nominal CPI month-on-month increase is mainly due to the seasonal rise in energy prices influenced by the cold winter in Europe and the U.S. The core CPI remains high, mainly affected by two factors: (1) Rent inflation may be relatively high month-on-month. The U.S. rent CPI lags behind market rents and market prices by about 1 to 1.5 years. At the end of 2023, the U.S. market rent is experiencing a temporary rebound, which may delay the improvement of rent CPI inflation in the first quarter of this year; (2) The rise in used car prices. Currently, U.S. auto inventory is below pre-pandemic levels, and high-frequency data shows that used car wholesale prices continued to rise month-on-month in December.

Chart 21: December National Gasoline Prices Rise Seasonally

Source: Bloomberg, CICC Research Department

Chart 22: Used Car Wholesale Prices Lead Used Car Inflation

Data Source: Manheim, BlackBook, CICC Research Department

By deriving year-on-year growth rates from month-on-month forecasts, we predict that the nominal CPI in December will rise to 2.9%-3% year-on-year, while the core CPI will slightly decline to 3.3% year-on-year. If December's inflation exceeds market expectations, it may raise tightening expectations, which would be unfavorable for overseas equity and bond performance.

Looking ahead, if we do not consider the impact of Trump's policies, we believe that U.S. inflation is expected to accelerate improvement starting in Q2, ultimately returning to normal, with the CPI falling to around 2.5%. From a component perspective, rent inflation may further decline under the influence of lagging effects, becoming a ballast for inflation improvement; easing supply chain pressures will lower core goods inflation; a cooling labor market will help other core services inflation continue to improve. However, if Trump's tariffs, immigration, and other policies are implemented quickly, they could completely reverse the inflation trajectory: according to PIIE estimates, Trump's policies could raise U.S. inflation by hundreds of basis points, at which point the Federal Reserve may be forced to end the rate-cutting cycle early, leading to stagflation in the market, with both equity and bond assets under pressure. Policy uncertainty leads to inflation uncertainty, and we believe that the uncertainty surrounding Federal Reserve policy and U.S. Treasury yield paths in 2025 is very high and requires further observation.

Chart 23: Model predicts that U.S. rent inflation is still expected to improve

Data Source: Zillow, Apartment List, CICC Research Department

Chart 24: Easing supply chain pressures help maintain low U.S. core goods inflation

Data Source: Wind, Bloomberg, CICC Research Department

Li Zhao Analyst SAC License No.: S0080523050001 SFC CE Ref: BTR923

Qu Botao Contact Person SAC License No.: S0080123080031

Yang Xiaoqing Analyst SAC License No.: S0080523040004 SFC CE Ref: BRY559 Risk Warning and Disclaimer

The market has risks, and investment should be cautious. This article does not constitute personal investment advice and does not take into account the specific investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investing based on this is at one's own risk