Zhongjin: Is gold actually expensive?

Updated Gold Pricing Model - Focus on Long-Term Actual Prices Despite the high levels of US Treasury rates maintained over the past year, with recent rapid increases, gold prices have risen against the trend, yielding returns of over 30% during the same period, further validating the decoupling logic between US Treasury rates and gold. Chart 1: The real interest rate of US Treasuries and gold prices have continued to "decouple" after 2022 Data Source: Wind, CICC Research Department In the article "Can Gold Still Be Bought?" released in January 2024, we constructed a gold four-factor cointegration regression model, suggesting that the deviation between gold and US Treasury rates mainly reflects the impact of central bank gold purchases and the expansion of US debt. Gold is not significantly overvalued, and it is predicted that gold prices may rise significantly in 2024. As we enter 2025, we find that the increase in gold prices last year has exceeded the explanatory power of the current model, with model residuals reaching historically high levels. Chart 2: Residuals from the four-factor cointegration regression model (January 2024 version) have exceeded $600, indicating a significant decrease in explanatory power for gold prices Data Source: Wind, CICC Research Department Note: The four factors include US public debt size, central bank gold purchases, real interest rates of 10-year US Treasuries, and the US dollar. For more details, see "Can Gold Still Be Bought?" published in January 2024, where the central bank's gold purchase size for Q4 2024 is predicted based on the average values of the previous three quarters, as official data have not yet been released. To better analyze and forecast the trend of gold prices, we have made the following updates to the gold pricing model: 1) Extend the time scale: The previous model was constructed based on data from 2003-2023, but the macroeconomic environment has changed since the pandemic, making the asset operation rules of the low inflation period of the past 20 years not necessarily applicable for the next 20 years. We extended the fitting period of the model to 1971-2024, encompassing the "Great Inflation" period of the 70s and 80s and the "Great Softening" era after the 90s. Due to the economic environment having undergone several changes during the fitting period, the regression coefficients of the model are better able to capture long-term asset rules across cycles, excluding certain short-term established data rules. 2) Eliminate US Treasury rates: The negative correlation between real interest rates of 10-year US Treasuries and gold has been very strong over the past 20 years, but before 2000, the relationship between the two was unstable and the regression coefficients were not statistically significant. Considering recent market trends, we believe that the negative correlation between gold and US Treasury rates may be a short-term rule. Therefore, we have removed US Treasury rates from the explanatory variables in the long-term model, only using central bank gold purchases, US debt size, and the US dollar as the three factors explaining gold prices. Chart 3: The relationship between gold prices and real interest rates of US Treasuries has been significantly negative only from 2005-2022, but the correlation is not high over a longer period Data Source: Wind, CICC Research Department Chart 4: The linkage between actual gold prices and real interest rates of US Treasuries is also not strong Data Source: Wind, CICC Research Department 3) From nominal prices to real prices: We found it difficult to find a stable cointegration relationship in the long-term three-factor gold pricing model, as reflected by residuals that have remained positive over the past 20 years, with the mean unable to converge to zero. Referring to relevant academic literature, we have used a new model framework that decomposes nominal gold prices into real gold prices and inflation. Gold has an "anti-inflation" property, with a close correlation between rising gold prices and rising prices, while inflation can be predicted separately using statistical models ("New Perspectives on Inflation Variables and Asset Transformations"). We found that real gold prices (where historical nominal gold prices are adjusted to correspond to prices in 2024) have a long-term stable relationship with some economic variables, allowing for the construction of appropriate statistical models, further discussed below. Multiplying real gold prices by inflation yields the equilibrium price of nominal gold prices and future predictions. Chart 5: The long-term nominal gold price model's cointegration regression method has failed, with residuals continuously positive and expanding over the past 20 years Data Source: Wind, CICC Research Department Chart 6: Nominal gold prices are significantly positively correlated with inflation levels, reflecting the "anti-inflation" value of gold Data Source: Wind, CICC Research Department In the long term, current gold prices may not be significantly overvalued The increase in gold value reflects a decline in the credibility of the US dollar system. The imbalance in US debt is a key reason for the decline in US dollar credibility, with central bank gold purchases being a result of the decline in the US dollar system's credibility. Therefore, the situation of US debt and central bank gold purchases are key long-term explanatory variables in the gold model. Data show that the correlations between these explanatory variables and actual gold prices are higher than those with nominal gold prices, aiding in increasing the model's explanatory power. Chart 7: The correlation between central bank gold purchases and actual gold prices is higher than with nominal gold prices Data Source: Wind, World Gold Council, CICC Research Department Note: Actual gold prices are derived from nominal gold prices divided by the US CPI index, with the base period set in November 2024 Chart 8: The correlation between the US deficit rate and actual gold prices is higher than with nominal gold prices Data Source: Wind, CICC Research Department Note: Actual gold prices are derived from nominal gold prices divided by the US CPI index, with the base period set in November 2024 The rise in gold value also reflects an increase in risk aversion, driven by economic cycle operations. Since nominal gold prices include inflation value, its long-term trend is upward, making it difficult to establish statistical connections with cyclical variables. The degree of increase in real gold prices is lower in the long term, with stronger cyclical properties, making it easier to establish connections with other cyclical variables. We found that the lower the GDP and weaker the PMI, the higher the real gold price. The correlation between real gold prices and consumer expectations and economic policy uncertainty indicators is particularly evident. Chart 9: Real gold prices are negatively correlated with the US economic cycle Data Source: Wind, CICC Research Department Note: Actual gold prices are derived from nominal gold prices divided by the US CPI index, with the base period set in November 2024 Chart 10: Real gold prices move opposite to consumer expectations Data Source: Wind, CICC Research Department Chart 11: Real gold prices are positively correlated with expectations of a deteriorating economy in the next 5 years Data Source: Wind, CICC Research Department Looking ahead, the expectation of a worsening economic situation in the next 5 years...From the perspective of the negative relationship with gold, the current gold price does not seem to agree with the market's optimistic expectations of a "soft landing". We have tried various combinations of explanatory variables for the actual gold price, and the regression results were generally similar, with residual levels ranging from 300 to 500 dollars. Among them, the 3-factor model of central bank purchases of gold/ETFs + deficit rate + University of Michigan consumer expectations had the smallest residuals this year. The model shows that the equilibrium price of gold by the end of 2024 is $2230 per ounce, corresponding to a residual of $390 per ounce.Chart 12: Co-integration Regression of Purchase of Gold + Deficit + Consumer Expectations Data Source: Wind, World Gold Council, CICC Research Department Chart 13: Co-integration Regression of Purchase of Gold by Central Bank + Deficit + US Dollar Index Data Source: Wind, World Gold Council, CICC Research Department Chart 14: Co-integration Regression of Purchase of Gold/ETF + Deficit + Real Effective US Dollar Data Source: Wind, World Gold Council, CICC Research Department Although the residuals of the above models are positive, it does not mean that the current gold price is significantly overvalued. Due to the long-term nature of the models spanning a long period, there have been significant economic structural changes during this time, resulting in larger residuals compared to short-term models. Looking at historical comparisons of residuals, the current residual level of 300-500 dollars is significantly lower than the residual levels in the 1970s and 2010s, indicating that the deviation is not extreme. Furthermore, as the models use end-of-year final values of annual frequency data, considering the high volatility of gold prices in 2024, the deviation between gold prices and the models in terms of annual average prices is expected to narrow further. Chart 15: Highest Historical Volatility of Gold Prices in 2024 Data Source: Wind, World Gold Council, CICC Research Department Chart 16: Using Data Annual Average Regression, Smaller Residuals in Model Data Source: Wind, World Gold Council, CICC Research Department Taking into account both nominal gold price models and real gold price models, we believe that although the current gold price is above the model's equilibrium price, the pricing is not extreme. Valuation will not exert significant pressure on gold prices, and trading gold based on fundamental macro logic is still feasible. In the next ten years, the central price of gold may range from 3000 to 5000 USD per ounce. From the key driving factors of the pricing model (US debt, central bank purchase of gold, inflation), gold still has sufficient upward potential in the medium to long term: if Trump wins the US election and the Republicans control both houses, more aggressive tax cut policies may be implemented in the future, coupled with the interest costs brought by high interest rates, the US debt may accelerate expansion in the coming decades, weakening the credibility of the US dollar and supporting gold prices. As of the third quarter of 2024, Asian major economies such as China, India, and Singapore hold gold as a proportion of their foreign exchange reserves of only about 5%-10%, while Turkey's gold reserves are relatively higher but still far lower than Western central banks where gold reserves exceed 70%. From 2022 to present, the gold purchases of the four central banks mentioned above accounted for over 60% of the global central bank's total gold purchases, indicating their active desire to replenish gold reserves. Chart 17: Significantly Lower Gold Reserves to Foreign Exchange Reserves Ratios in Asian Central Banks Compared to European and American Countries Data Source: Wind, World Gold Council, CICC Research Department Our calculations show that if Asian central banks can eventually increase their gold reserves to levels comparable to Western central banks, it will bring a demand for gold purchases exceeding 30,000 tons, which is six times the current annual supply of gold. Therefore, in response to deglobalization and declining US dollar credit, Asian central bank purchases of gold will also push up gold prices. Based on our calculations, the US inflation center may rise to around 2.5%-3% after the pandemic ("Is America Entering a High-Interest Rate Era?"). However, if Trump's presidency leads to uncontrollable inflation in the US, according to PIIE estimates, inflation could rise to over 9% in the next 2-3 years. Regardless of whether inflation is out of control, rising prices are positive for gold performance. Using a model of central bank gold purchases + US deficit rate dual-factor real gold price, along with reasonable assumptions about gold explanatory variables, we calculate the potential upside of gold prices. 1) Neutral Scenario: Because Asian central banks need to replenish a significant amount of gold reserves, and the global situation may become more complicated after Trump's presidency, with the credibility of the dollar system deteriorating, we expect central bank gold purchase rates to remain at least at current levels, around 1000 tons per year. For US debt, we use the forecast results under the baseline scenario of the US Congressional Budget Office (CRFB), where the US fiscal deficit is expected to increase by around $7.5 trillion over the next 10 years. We also assume that Trump will not cause uncontrollable inflation, with the CPI stabilizing at 3%, estimating the gold price to rise to $3,300 per ounce over the next decade. 2) Aggressive Scenario: Assuming Asian central banks wish to replenish their gold reserves more quickly, accelerating gold purchases to 1,500 tons per year (too rapid central bank gold purchases may push up gold prices, raising the cost of purchases, and we do not make overly high assumptions about the pace of central bank gold purchases). On the fiscal front, we assume that Trump significantly expands the budget, reaching the upper limit forecast by the CRFB, with an additional increase in the US fiscal deficit of around $15 trillion, while also adopting the PIIE's forecast for the inflation path of the Trump impact, with inflation initially rising to 9%, then falling back to the 3% center. This calculation estimates the central value of gold over the next decade to rise to $5,000 per ounce. Chart 18: Potential Range of Central Prices for Gold in the Next 10 Years Between $3,000 - $5,000 per Ounce Data Source: Wind, World Gold Council, CRFB, CBO, CICC Research Department Gold can still be over-allocated, but the rate of increase by 2025 may slow down, and volatility may increase Based on the results of the above models, gold's short-term valuation is reasonable, and there is still ample room for upward movement in the long term, therefore valuation is not a constraint on further increases. Gold prices have fluctuated greatly in the past two months, possibly due to a decrease in 'Trump trade' risk aversion sentiment and a reduction in geopolitical risks. Looking ahead to 2025, we expect that Trump's policies are likely to lead to "uncontrolled inflation" or "economic slowdown", both of which are favorable for gold performance ("Asset Class Outlook 2025: Adapting to Change"). Although the market's expectations for the Russia-Ukraine and Middle East situations are becoming more optimistic, we caution against overlooking the risks of other geopolitical issues caused by Trump's foreign policies. Under the trend of deglobalization, more funds may flow into safe-haven assets, with gold likely to outperform. Meanwhile, as gold prices have risen significantly in 2024, accumulating a certain positive residual, there is some mean reversion pressure, and gold may experience increased volatility in the short term, with growth in 2025 potentially lower than in the past two years. Therefore, we recommend maintaining an over-allocation to gold, moderating the short-term trading value of gold, and focusing on the long-term allocation value of gold.Bonjour! Comment a va?December US CPI may be biased towards the high side, pay attention to risk control The December US CPI will be released on January 15 (Wednesday). CICC's large asset class model predicts a month-on-month nominal CPI increase of 0.41% (consensus expectation of 0.3%, previous value of 0.31%), higher than the previous value and market consensus expectation; predicts a month-on-month core CPI increase of 0.25% (consensus expectation of 0.2%, previous value of 0.31%). Chart 19: Split and forecast of month-on-month growth rate of nominal CPI in the US Data source: Haver, CICC Research Department Chart 20: Split and forecast of core CPI month-on-month growth rate in the US Data source: Haver, CICC Research Department The increase in nominal CPI is mainly due to the super-seasonal rise in energy prices under the influence of the cold winter in Europe and America. The high level of core CPI is mainly influenced by two factors: (1) Rent inflation may be biased towards the high side. The lag between the US rent CPI and market rent and market house prices is around 1 year to 1.5 years. By the end of 2023, there may be a phased rebound in the US market rent, which may slow down the progress of rent CPI inflation improvement in the first quarter of this year; (2) Used car prices are rising. Currently, US car inventories are lower than pre-pandemic levels, and high-frequency data shows that wholesale prices of used cars in December continued to rise. Chart 21: Seasonally high gasoline prices in the US in December Data source: Bloomberg, CICC Research Department Chart 22: Wholesale prices of used cars leading used car inflation Data source: Manheim, BlackBook, CICC Research Department By predicting the year-on-year growth rate from the month-on-month forecast, we predict that the year-on-year nominal CPI will rise to 2.9%-3% in December, while the year-on-year core CPI will decrease slightly to 3.3%. If inflation in December is higher than market expectations, it may raise tightening expectations, which is not conducive to the performance of overseas stocks and bonds. Looking ahead, if we do not consider the impact of Trump's policies, we believe that US inflation is expected to accelerate in Q2 and eventually return to normal, with CPI falling to around 2.5%: In terms of components, rent inflation may further decrease under the influence of lagging effects, serving as a ballast for inflation improvement; alleviation of supply chain pressures may keep core goods inflation low; cooling in the labor market may help to continue improving inflation in other core services. However, if Trump's tariffs, immigration, and other policies are implemented quickly, they may completely reverse the path of inflation: According to calculations by PIIE, Trump's policies could raise US inflation by hundreds of basis points, in which case the Fed may be forced to end its rate-cutting cycle early, leading to stagnation in the market, with pressure on stocks, bonds, and assets. Policy uncertainty leads to inflation uncertainty - we believe that the uncertainty in Fed policy and US bond rate paths in 2025 is very high and requires further observation. Chart 23: Model predicts that US rent inflation is still expected to improve Data source: Zillow, Apartment List, CICC Research Department Chart 24: Alleviation of supply chain pressure helps to keep US core goods inflation low Data source: Wind, Bloomberg, CICC Research Department Asset allocation suggestions Chinese stocks: Standard allocation, recommend buying on dips and maintaining tactical flexibility In December 2024, the A-share market showed a volatile trend, with a significant adjustment on the first day of 2025, mainly due to internal and external factors: the internal economic fundamentals are still to be improved, with the domestic manufacturing PMI in December at 50.5, lower than market expectations, and the issue of low inflation has not been resolved; with Trump's inauguration nearing, external uncertainties have increased, affecting A-share risk appetite. Looking ahead, the strategic value of Chinese stocks in 2025 is expected to be further enhanced compared to last year. Firstly, this year's countercyclical macroeconomic policies are expected to continue to strengthen. Last December's Central Economic Work Conference set an overall positive tone, with increased focus on stabilizing the real estate and stock markets, a shift in monetary policy towards "moderate easing," fiscal policy emphasizing "more proactive" measures and "increasing the issuance of ultra-long special government bonds," among other measures. Secondly, from a valuation perspective, A-shares still have significant investment attractiveness both vertically and horizontally. After a quarter of volatile adjustments, the forward price-earnings ratio of the non-financial sector of the CSI 300 is only 14.5 times, lower than the historical average of 15.8 times, and also lower than the valuations of major global markets. However, during the game between policy and economic dynamics, the upward trend in the stock market may not be smooth, so tactical flexibility is still needed. Currently, the domestic financial cycle is still declining, there are bottlenecks in the transmission of monetary policy to the economy, and there is still room for improvement in the coordination between fiscal and monetary policy. The new year's fiscal budget needs to be reviewed by the National People's Congress, and it will take some time for policies to be implemented and transmitted to the economic fundamentals. After Trump takes office, increased external uncertainties may also temporarily affect the risk sentiment of the domestic stock market. Structurally, during the short-term adjustment period, the dividend style may have a competitive advantage in stages but attention should be paid to the structure. Consumer-oriented and natural monopoly dividends with attractive dividend yields may be better choices, with a focus on the media, food and beverage, transportation, and telecommunications sectors; during the period of increased external uncertainties, attention may be given to sectors related to industrial autonomy and domestic substitution, such as semiconductors and high-end manufacturing; additionally, we believe that companies with annual forecasts exceeding expectations and good progress in supply clearance will have relative performance, such as industrial metals, home appliances, lithium batteries, etc. In terms of policy support, mergers, reorganization, book-to-market restoration, and debt-for-equity swaps are more certain directions. (See "A-share strategy: Beginning with volatility, how will the future evolve?") Chinese bonds: Maintain standard allocation Currently, real interest rates are still relatively high, and monetary policy is expected to remain accommodative. Although the ten-year government bond yield has fallen to historical lows, the real interest rate calculated by subtracting the GDP deflator from the seven-day reverse repurchase rate is 2.84%, still higher than the natural rate of interest. In addition, although liquidity is relatively abundant, it is mainly concentrated in the interbank market, and financing liquidity has not yet transformed into market liquidity, as evidenced by the continuous increases in the amount of collateral repo transactions in the interbank market over the past 12 months, averaging nearly 14 trillion yuan per month, while new social financing is only 3The total amount is trillion yuan. Looking at the risk premium and credit premium, the financing cost of the real economy has not decreased significantly. With high real interest rates and inefficient liquidity transmission, the domestic economy's fundamentals are moderately recovering. Therefore, monetary policy will still need to be loosened in the near future, and policy rates and money market rates will ultimately need to be lowered. Combined with short-term exchange rate constraints, we believe that the bond market in the first quarter may still be relatively positive.However, after the second quarter of 2025, with the weakening of the exchange rate constraints, the domestic monetary policy relaxation space is expected to open up, and the long-term interest rates may decrease or slow down amid the pre-emptive situation in the market, at which time the yield curve may steepen again. Overseas Bonds: The probability of an overseas "soft landing" has increased, short-term attention should be focused on controlling risks. The unexpectedly strong US non-farm payroll data released on January 10th has increased the probability of the US economy avoiding a "soft landing", with the ten-year US Treasury bond yield surging to 4.8% at one point. Looking ahead, considering that US inflation may remain strong for the next 1-2 months, and with the possibility of Trump advancing tariffs and deportation policies quickly after taking office on January 20th, we believe it may lead to renewed market tightening expectations, triggering a "double kill" in stocks and bonds. Therefore, the short-term trend of US Treasury bonds remains uncertain. In the medium term, if Trump's policies lead to uncontrollable inflation, the Federal Reserve may pause rate cuts, causing US bond yields to return to an upward trend, creating negative pressure on overseas bonds. However, if Trump's policies not only raise inflation but also suppress growth, the Federal Reserve may resume rate cuts, and US bonds may rebound from the bottom. Referring to the market situation in the US bond market after Trump's election, US bond yields also fell in early 2017 after oscillating upwards. It is recommended to remain relatively neutral on US bonds in the short term (1-3 months), moderate risk control, and decide whether to increase holdings of US bonds based on the economic path in the medium term. Overseas Stocks: Suggest reducing allocations gradually. Currently, the US economy has strong resilience, with the risk of the US economy avoiding a "soft landing" increasing, causing concerns in the market about the possibility of the Federal Reserve ending rate cuts early in 2025. Looking ahead, after Trump officially takes office, his policies are likely to raise inflation or depress growth, both of which are unfavorable for the performance of overseas stock markets. If Trump's tariffs and immigration restrictions policies take effect first, they will significantly increase US inflation, leading to an increase in market tightening expectations, increasing the risk of a correction in overseas stock markets. If Trump's government policies fail to effectively support the economy and US growth declines, it may be negative for US stocks. From the perspective of the relative valuation of stocks and bonds, both the 3-month US Treasury bill yield and the 10-year Treasury bond yield are higher than the S&P 500 forward price-earnings ratio (expected earnings over the next 12 months/stock index price), implying that the market's expected return on safe assets (US bonds) is higher than that of risk assets (US stocks), with the market pricing risk almost at 0. This extreme pricing reflects the possibility that investor sentiment may be overly optimistic. Once facing negative shocks, high valuations may amplify the vulnerability of stock assets. Commodities: Maintain underweight. In December, weather-related trading activities picked up, and global commodity prices showed a slight uptrend. We expect the risk of a subsequent adjustment in commodity prices to increase: first, due to the recent strong economic data released overseas, the main theme in overseas markets is gradually shifting from a "soft landing" to a tightening trade environment. The increase in market tightening expectations may increase downward risks for commodity prices. Secondly, the supply and demand balance of commodities is also weakening marginally: current oil production is gradually recovering, overseas copper ore continues to accumulate inventory, and from a supply and demand perspective, it does not support the short-term performance of commodities. Therefore, we believe that commodity prices may face adjustment risks in the short term overall. This article is reprinted from the WeChat public account "Zhongjin Point Insight". Edited by GMTEight: Chen Xiaoyi.