Novel Evidence of Macroeconomic Regime Change

Bringing new data sources to bear on questions of interest is one of the best ways of leashing our perceptions tighter to reality. As an equities quant, I subscribe to commercial data on firms listed in the US. The earnings data and so on contain macro information that can be recovered if the data is processed and aggregated properly. Here were look at the cost of producing goods and services reported by listed firms in the SF1 filings. Suitably aggregated production costs corresponds closely to the inflation as experienced by the listed firms. Clearly, we ought to expect fluctuations in this variable to covary strongly with general inflation.

More precisely, we expect a stable structural relation between general inflation as experienced by consumers on the one hand, and inflation as experienced by firms on the other. Moreover, divergence between the two contains “cyclical” information on slack in the rest of the system; in particular, the labor market. For when wages rise in a high pressure economy, firms’ cost of production go up. In what follows, we will show that the structural relationship between these two variables changed after Biden took office and consummated Trump’s break with neoliberalism.

As our proxy of general inflation, we use core PCE from FRED. As our proxy of firms’ cost of production, we use the mean cost of revenue (“cor”) of firms listed in the NYSE and the NASDAQ. Cost of revenue is defined as the cost of goods produced and sold and services rendered by the firm during the past year, as of the reporting quarter. Here’s a time-series plot of the two. We’ve standardized both for ease of visualization. We can so that there is strong covariation. But has this relationship changed over time?

From the time-series, the eye is not able to detect the structural change. The next plot shows the scatter of quarterly observations of core PCE and listed firms’ cost of production in 1999-2024. The Biden years are in blue. The pattern is striking and unmistakable: not just the level of inflation but its elasticity with respect to firms’ production costs has increased substantially.

We report the results our baseline regression. In addition to a dummy for the Biden years (that captures the level shift), we also admit a dummy for the interaction to capture the change in the slope coefficient. We use robust standard errors. If we use 4-qtr percentage changes, the Durbin-Watson statistic suggests the presence of auto-correlation due to lack of proper detrending. We detrend by using quarterly percentage change. This gets rid of the autocorrelation (DW=2.00).

We estimate that, Pre-Biden, the underlying rate of general inflation was 1.6% per annum and the elasticity of general inflation to production inflation was 0.027. The new macro regime consolidated under Biden has shifted the former up by 1.1% to 2.7% per annum. The elasticity of general inflation to firms’ cost of production has increased dramatically from 0.027 to 0.1382.

We have only documented the structural change in the relationship between general inflation and firms’ cost of production. We posit that this structural change is due to the new high-pressure economy that I announced two months into the Biden administration. Specifically, we posit that the change is due to tighter labor markets. We leave the task of testing that hypothesis to later work and other researchers.