Does Gold Go Up With Inflation? Testing U.S. CPI Against Gold’s 5-Day Returns
Published on September 11, 2026
Does Gold Go Up With Inflation?
Gold is often described as an asset that performs well during inflation. When prices rise and the purchasing power of money falls, gold tends to receive more attention as an inflation hedge. This has helped create a familiar view: higher inflation should mean higher gold prices.
But what does the historical data actually show? When U.S. inflation is high, does gold tend to rise over the days that follow?
In this analysis, we compare U.S. CPI year-over-year with XAU/USD’s return over the following five trading days. The question is whether “buy gold because inflation is high” has historically worked as a simple short-term idea.
The result is not that straightforward. Higher CPI did not consistently correspond to higher XAU/USD returns over the next five trading days. That does not mean gold is not an inflation hedge. Long-term inflation protection and using CPI as a short-term trading input are different questions.
Why Is Gold Considered an Inflation Hedge?
Gold has long been discussed as a way to preserve value when inflation reduces the purchasing power of money. This is one reason investors often associate rising inflation with stronger gold prices.
But there is an important distinction: gold potentially serving as a long-term inflation hedge is not the same as high CPI being followed by higher gold prices a few days later.
Testing whether gold preserves purchasing power over several years and testing whether today’s inflation rate is associated with XAU/USD returns over the next five trading days involve different time horizons and different questions.
This analysis focuses on the second question: when CPI year-over-year was high, what relationship did we observe with XAU/USD returns over the following five trading days?
How We Tested CPI Against Gold’s 5-Day Forward Return
The asset analyzed here is XAU/USD. For each point in time, we calculate its return over the following five trading days and compare that return with the U.S. CPI year-over-year information available at the time.
Using U.S. CPI Year-over-Year
The main Factor in this analysis is CPIAUCNS lag pct change 12.
CPIAUCNS is the U.S. CPI-U All Items, not seasonally adjusted series. Applying lag pct change 12 measures the percentage change from 12 months earlier—in other words, the year-over-year inflation rate.
So when this article refers to “high CPI,” it generally means a high year-over-year inflation rate, rather than a high level of the CPI index itself.
CPI is also a monthly series. FactDecode aligns it with the daily XAU/USD series while accounting for when the CPI information becomes available, and the same CPI value remains in use until the next update. This matters later when interpreting the SHAP dependence chart: many plotted daily observations do not represent an equal number of independent CPI releases.
Why Five Trading Days?
This analysis is not designed to test whether gold preserves purchasing power over long periods. It focuses on the relationship between the inflation environment and relatively short-term gold performance, using a five-trading-day horizon.
It is also not a study of gold’s immediate reaction to a CPI release. We are not testing whether CPI came in above or below market expectations or how XAU/USD moved immediately after the announcement.
Instead, we examine the historical relationship between the CPI year-over-year information available at each point in time and XAU/USD’s return over the following five trading days.
Higher CPI Did Not Consistently Mean Higher 5-Day Gold Returns
We first divide CPIAUCNS lag pct change 12 into ten quantiles, from the lowest to the highest CPI year-over-year observations, and compare the subsequent five-trading-day XAU/USD returns.
If the simple idea that “higher inflation means stronger gold” also held over this short horizon, we would expect average forward returns to generally rise as we move from lower-CPI to higher-CPI quantiles.
That is not what the Quantile Profile shows.
The differences are clearer in the underlying numbers. In Q5, where CPI year-over-year ranged from 1.91% to 2.20%, the average XAU/USD return over the next five trading days was +0.41%, with positive returns in 61.9% of observations.
By comparison, in Q9, where CPI ranged from 4.98% to 7.04%, the average forward return was +0.11%, with a positive-return rate of 54.7%. In Q10, the highest-inflation bucket at 7.11% to 9.06%, the average return was only +0.08%, with positive returns in 50.8% of observations.
The lower-CPI buckets were not consistent either. Q1 produced an average return of −0.06%, Q2 +0.31%, and Q3 −0.24%. Q4 through Q8 were positive on average, but the returns did not improve steadily as CPI increased.
The important point is therefore not that one particular inflation level was “best” for gold. It is that higher CPI did not produce a simple upward progression in subsequent five-day gold returns.
Likewise, the relatively strong Q5 result does not establish that inflation around 2% is an optimal level for buying gold. Each quantile contains observations from different market environments, and this comparison alone cannot validate a specific CPI level as a trading threshold.
Extreme Inflation Looked Different in the Model
The Quantile Profile does not show a simple relationship in which gold returns increase with inflation. The SHAP dependence view, however, reveals a more complex pattern.
Across much of the normal CPI range, the model contribution associated with CPI is relatively small or changes gradually. But at extremely high inflation rates—around 8% and above—the shape of the SHAP contribution changes noticeably.
This result needs careful interpretation. SHAP dependence is not a simple correlation between CPI and gold. It shows how the CPI Factor contributed to the output of a model that also includes other Factors.
Why 8% Is Not a Trading Threshold
The roughly 8% level is where the shape appears to change in this particular chart. It has not been statistically validated as an optimal buy or sell threshold.
Periods with CPI this high are also relatively unusual. Observations in the extreme-inflation region may be concentrated in a limited number of historical high-inflation episodes, which makes it risky to generalize the same pattern to ordinary inflation environments.
CPI is also monthly, while the target series is daily. Multiple daily observations on the SHAP dependence chart can therefore correspond to the same underlying monthly CPI information. The number of plotted points should not be interpreted as the number of independent high-inflation episodes.
The appropriate conclusion is narrower: during extreme inflation periods, CPI’s contribution within the model appeared different from its behavior across the more typical inflation range.
That does not justify a rule such as “sell gold when CPI exceeds 8%.”
Was CPI a Major Factor in Explaining Gold’s 5-Day Returns?
Next, we look at how much information CPI contributed within the model as a whole.
A lack of a simple monotonic Quantile Profile does not mean CPI is completely unrelated to XAU/USD. Conversely, an interesting shape in SHAP dependence does not mean CPI is the dominant Factor behind short-term gold movements.
This is where Factor Contribution provides additional context.
CPIAUCNS lag pct change 12 remained in the final model, but it was not among the dominant contributors. The appropriate interpretation is therefore that CPI was not completely irrelevant, but it also did not explain XAU/USD’s five-day forward returns on its own.
That does not automatically mean another specific variable—such as real yields, nominal yields, or inflation expectations—is “what really drives gold.” Those relationships require separate analysis.
This article remains focused on one question: is CPI year-over-year by itself sufficient to support a simple short-term view on gold?
The results suggest that it is not.
Does This Mean Gold Is Not an Inflation Hedge?
No. That is not what this analysis shows.
The question tested here is whether high CPI year-over-year was associated with higher XAU/USD returns over the following five trading days. That is different from asking whether gold preserves purchasing power against inflation over the long term.
Long-Term Inflation Hedge vs. Short-Term CPI-Based Decisions
Testing gold as an inflation hedge over several years would require a different time horizon—months, years, or even longer—not five trading days.
The result established here is much narrower: high CPI alone did not consistently translate into stronger XAU/USD returns over the next five trading days.
It would therefore be incorrect to turn this finding into the broader claim that “gold is not an inflation hedge.” Changing the time horizon changes the question itself.
This Is Not a CPI Release-Day Study
There is another important distinction.
This analysis does not test what happens when CPI is released above or below market expectations. It does not examine whether an upside CPI surprise causes gold to rise or fall immediately after the announcement, and it does not isolate CPI release days.
So the results should not be used to answer a question such as:
“If CPI comes in above expectations today, will gold go up or down?”
Answering that would require a separate event study using CPI actual versus consensus, the precise release timestamp, and XAU/USD prices before and after the announcement.
Conclusion: High Inflation Alone Is Not Enough for a Short-Term Gold Call
The idea that “gold goes up with inflation” is intuitive and widely repeated. But when we compared U.S. CPI year-over-year with XAU/USD’s subsequent five-trading-day returns, higher inflation did not consistently translate into stronger short-term gold performance.
The Quantile Profile did not show a steady improvement in forward returns as CPI increased. Q5, with CPI between 1.91% and 2.20%, produced an average five-day return of +0.41%, while the highest-inflation Q10 bucket, between 7.11% and 9.06%, produced only +0.08%.
SHAP dependence showed a different model shape in extreme high-inflation conditions, but that does not validate an 8% CPI trading threshold. Factor Contribution also showed that CPI retained some information while remaining far from a dominant standalone Factor for short-term XAU/USD performance.
The conclusion is therefore straightforward: “buy gold because inflation is high” is too simple a rule for a five-trading-day decision based on this historical analysis.
That is not the same as saying gold cannot serve as a long-term inflation hedge. Whether gold protects purchasing power over years and whether high CPI predicts higher gold prices over the next few days are separate questions.
This analysis answers only the latter.