January 30, 2026

Model Perspectives on Supply and Demand Factors behind a Soft Labor Market

Danilo Cascaldi-Garcia and Camilo Morales-Jiménez¹

U.S. employment growth slowed down notably in the second part of 2025, and a key question is how much of the weakness stems from labor demand and how much from labor supply. In this note, we examine this question from the point of view of two different models: a statistical model that uses an intuitive interpretation of the joint behavior of employment and wage growth to infer the effects of labor supply and demand (VAR), and a structural dynamic stochastic general equilibrium (DSGE) model that uses a much wider array of data. While other models could also be employed, these two are, by design, well suited for this purpose. Our results point to supply factors as the dominant cause of the cooling labor market, although there also are some signs of weakening demand.

Insights from a statistical model

We start with a two-variable VAR model of private employment gains and real wage growth in the United States, using monthly data through August 2025. The model identifies labor supply and labor demand forces based on the premise that lower demand for workers decreases both employment and wage growth, while lower supply of workers decreases employment but increases wage growth, in the same spirit as Baumeister and Hamilton (2015). Having identified these drivers, we use the model to decompose private employment gains into labor-supply and labor-demand factors. As shown in figure 1, according to this model, and after controlling for the lagged effects of the unusual COVID-19 shock in early 2020 (the pink bars), supply forces (the green bars) were the prominent drivers of employment gains from 2022 to 2024, coinciding with the large influx of immigrants into the country during that period. Employment gains have fallen since the start of 2025, and with real wage growth decelerating, the model infers that the contribution of labor demand (the blue bars) decreased and then turned negative in the spring. As the slowing in employment gains since then has not been accompanied by a further slowdown in real wage growth, the model attributes it to lower supply, as indicated by the shrinking size of the green bars, even as labor demand remained a small drag. Although the important role for supply in the model is consistent with the reduction in immigration inflows in 2025, there may be other supply-side factors at work.

Insights from a structural model

One way to see if other supply factors beyond the direct mechanical effects of population growth and immigration might be contributing to the decline in employment growth is to focus on the employment-to-population (EPOP) ratio, which has declined 40 basis points between 2025Q3 and 2024Q4. To conduct this further decomposition, we use the model by Cairó et al (2023), a DSGE model that takes into account a broader set of indicators and allows us to identify the contributions of more specific supply and demand.²

Figure 2 plots the DSGE model's estimates of the driving forces behind quarterly changes in the EPOP ratio since 2022 until the third quarter of 2025. To highlight the contribution of relatively recent shocks, we distinguish between contributions of supply and demand shocks occurring since 2022 (the green and blue bars) and the lagged effects of shocks before 2022 (the pink bars); those lagged effects are small for 2025. Comparing the relative sizes of the green and blue bars, the DSGE model views supply shocks as the main reason for the decline in the EPOP ratio in 2025. The intuition behind this result is as follows: While a much lower-than-expected EPOP ratio could have been driven by negative shocks to aggregate demand, these shocks would have also produced a much larger decline in the workweek, GDP growth, and wage inflation than we have seen thus far.

According to the model, the contribution from aggregate supply shocks mainly comes from a surprisingly weak labor force participation rate and surprisingly strong productivity and workweek hours. In the model, these shocks lead firms to economize on workers, resulting in a lower EPOP ratio. To the extent that demand-side factors have played a role, their negative contributions were mostly driven by adverse direct shocks to demand for investment and consumption of durable goods.

Conclusion

Even though our statistical and structural models react to different data and make different assumptions about how demand and supply shocks are identified, both point to a larger role for supply forces than for demand forces in explaining the softening in the labor market. That said, the models also see some contribution from weak labor demand in 2025.

References

Baumeister, Christiane, and James D. Hamilton (2015), "Sign restrictions, structural vector autoregressions, and useful prior information." Econometrica 83, no. 5 (2015): 1963-1999.

Cairó, Isabel, Hess T. Chung, Francesco Ferrante, Cristina Fuentes-Albero, Camilo Morales-Jiménez, and Damjan Pfajfar (2023), "Endogenous Labor Supply in an Estimated New-Keynesian Model: Nominal versus Real Rigidities," Finance and Economics Discussion Series 2023-069 (Washington: Board of Governors of the Federal Reserve System, November).

Cascaldi-Garcia (2022), "Pandemic Priors," International Finance Discussion Papers 1352 (Washington: Board of Governors of the Federal Reserve System, August). 

Appendix