Tariffs, Inflation, and Returns: How Investments Respond to Supply Shocks

Tariffs have regained the economic spotlight. However, the uncertainty of timing and scale gives investors the advantage. The attractive history of tariffs and impact on investment returns are provided by Baltussen et al in recent years Entertainment investor Blog. This blog uses a complementary approach to investigating the possible impact on returns.

Customs duties will change Relative price. Just as a massive change in crude oil prices drives energy costs compared to other commodities, tariffs make imports relatively expensive. In economics terminology, tariffs are “supply shock.” Also, price adjustments are expensive for businesses in the short term, so import prices will rise in response to large amounts of tariffs, but other prices will not change anytime soon, despite the possibility of easing demand (see Romer 2019 (for the explanation of modern macros of “nominal rigidity.”) This causes average The rising price level. In other words, tariffs will increase the inflation rate of headlines (all items).

This post provides a framework for thinking about the impact of tariffs on the revenues of key asset classes by estimating the response to the supply shock of asset classes. By separating the inflation “signal” or trending components (determined by the basic force) from the shock-driven “noise” components, the latter can be estimated to the latter. This may suggest lessons about the possibility of asset class responses to one-off tariffs.

Quantifying inflation shock using CORE and median CPI

Economic theory and a little analysis allow us to infer how asset classes respond to the inflation shock effects of tariffs.

In terms of theory, modern macroeconomics uses the “Philips Curve” framework to explain inflation. This was named after an economist who first pointed out that inflation was negatively associated with economic lag (Philips used unemployment and wages). Philips curves can be specified in a variety of ways. In general, inflation is explained by three variables: inflation expectations (consumer, business, or professional forecaster), output gaps (e.g., unemployment rates and vacancy and unemployment rates), and shock terminology.

In this blog, we use the Philips Curve approach to isolate the inflation expectations and output gaps, as well as inflation signals or trends caused by noise and fleeting factors of inflation.

This avoids two issues. That tariff is shocking I’ll go through To drive inflation by raising expectations and production costs and other channels. In fact, there is already evidence Consumer inflation expectations are rising. However, incorporating these effects makes this analysis quite complicated and is ignored for now.

Phillips’ curves show that inflation can be broken down into trends and shock components. This is usually done by subtracting the trend inflation from heading (all items) inflation. Instead, this blog uses the median consumer price index (CPI) inflation rate, as calculated by the Federal Reserve Bank of Cleveland as a proxy for trend inflation, due to the attractive characteristics of CPI.[1]

Additionally, instead of using headline CPI inflation as a starting point, we use core CPI inflation, which excludes food and energy (XFE CPI). XFE CPI is preferred as the difference in median XFE and CPI results in a measure of impulse purified from large changes in relative food and energy prices. This measure is called “non-XFE shocks.”

The chart in the panel in Figure 1 shows the sense of frequency and size of non-XFE shocks. Scatter plots show monthly XFE and median inflation. If they are equal, the points are on the 45 degree line. The pair above the 45 degree line is a positive non-XFE shock and vice versa. (The R code used to create the charts presented in this blog and perform the analysis is r-pubs page). The histogram shows the distribution of these shocks. Major disturbances are rare.

Figure 1. The top panel shows the median and XFE CPI from 1983 to 2025. The lower panel shows the distribution of shocks (distance from the 45 degree line on the top panel). The frequencies for each of the 11 “bins” will be displayed in the bar.

Source: Fred

Asset class sensitivity to inflation surprises

With the non-XFE shock defined, we can estimate how the major asset classes responded. This may provide a preview of how these asset classes respond to inflation shocks caused by tariffs.

Relationships are estimated in a conventional way. By regressing asset class returns for non-XFE shocks. The estimated coefficients in the result are the non-XFE shock “beta” for the variable on the left. This approach is traditional, and mirrors photographed in me Entertainment investor Did the actual assets offer inflation hedge when blog investors needed them most?

Regression uses the monthly change in non-XFE shocks as the right-hand variable. Monthly returns for the S&P 500 Total Return (S&P 500) index, Northern Trustal Asset Allocation Total Revenue Index, Bloomberg Total Return (BCI) Index, Bloomberg Tip Index, and 1-3 Month Bill Return (T-Bills) Index. The inflation data is from Fred and from index returns from YCHARTS. Sample sizes vary depending on asset class regression, so they run over the longest sample period available for each asset class. This will end in March 2025 in each case.

One caveat before discussing the results. Non-XFE shocks may be the cause Any Of course, a significant relative price change except for food and energy changes. That is, the supply shock includes all that supply chain shock.

Unfortunately, there is no obvious way to isolate the obstacles that are most interested in using public inflation data. However, because we cannot know exactly what form such tariff-induced interference will take, a survey of asset class responses to non-XFE shocks is a reasonable place to begin. However, the results are shown in Appendix 2.

Figure 2. Regression results.

A positive and significant estimate of the “non_xfe_shock” coefficient suggests that the asset class is hedging against non-XFE shocks. Estimating non-positive but insignificant coefficients suggests that they may hedge non-XFE shocks, but the sample size does not allow us to reject the claims of inconfidence. Confidence intervals provide a sense of the magnitude of the impact of inflation on returns, and of course, the reliability of estimates.

These findings suggest that the product (BCI) reacts positively to shock and that T-bill has a negative response, but it is estimated that the former relationship is less accurate than the latter (i.e., the T-bill confidence interval is strict). Of the remaining asset classes, hints, stocks and actual assets, it comes with the correct indications of impact hedges (positives), but is estimated too inaccurately to weakly support the claim. These conclusions are robust to estimates of general sample periods (2015:12–2025:3).

Customs Price Shock Brace

This short exercise suggests, on average, “hedge” the “hedge” shock to inflation caused by large relative price changes (other foods and energy). T-Bill wasn’t like that. (The relationship of Shock T-Bill can be explained by the fear that price-level jumps can trigger a tighter response in monetary policy.

If the empirical relationship estimated here is stable and tariffs affect inflation, like non-XFE shocks, the approach that follows may help inform directional estimates of how tariffs affect investment returns.

[1] Outlier emission measures like median are more efficient measures of more efficient measures of population mean (in our case trends), such as those shown by the distribution of monthly price changes over the sample mean in the presence of a “fat tail.” Furthermore, median and other trimmed average inflation measures are better predictors of future inflation and are less correlated with future increases in money supply (suggesting that central banks generally respond to “supply shocks” rather than traditional “core” (e.g. food and energy) inflation).