The Tax Policy Center has traditionally measured the macroeconomic effects of tax changes based on historical experience and empirical estimates from the economics literature. But that isn’t the only way to estimate how tax changes affect the overall economy. Another one, implemented in the OG-USA model, considers how households respond to expected future changes in policy. Going forward, TPC plans to include estimates using the OG-USA model in its analysis of major tax legislation.
To see how these two methods differ, we looked at Joe Biden’s presidential campaign tax plans using the OG-USA model combined with our microsimulation model. Then we compared the results with our traditional macroeconomic analysis paired with the tax model. The new analysis estimates that Biden’s campaign proposals would have increased economic output modestly in the first year, but reduced output in later years. That top-line result is similar to TPC’s 2020 estimate, but with some differences.
TPC performed the new analysis with the designers of the OG-USA model, Richard Evans of Rice University and Jason DeBacker of the University of South Carolina. In that model, simulated households make choices about how much to work and to save based on future, as well as current, tax policy and macroeconomic conditions.
To put it another way, how will these households work and save in response to their expectations of future changes to policy? For example, marginal tax rates will rise if most individual income tax provisions of Tax Cuts and Jobs Act (TCJA) expire, as scheduled, after 2025. To the extent people expect that will happen, they might work more in the short term to take advantage of today’s lower tax rates.
By contrast, the macroeconomic model TPC used previously looked to how taxpayers responded to tax changes in the past and empirical estimates of behavioral responses to current conditions. For example, how will labor supply respond to current changes in marginal tax rates on labor income?
Building TPC modeling capacity to use the OG-USA model was a major undertaking. First, TPC had to use its microsimulation model to create a set of tax rate and income measures for each of the roughly 280,000 records in its microsimulation model. Those measures were then used to estimate the tax schedules that applied to the simulated households in the OG-USA model.
The nature of Biden’s campaign plan created special challenges and opportunities. Biden would raise taxes primarily on very high-income households. But, while the simulated households in the OG-USA model spanned a range of incomes, it excluded the highest. As a result, Evans and DeBacker added categories of very high-income households. The model now has a unique capacity to model the economic effects of taxes on high-income taxpayers within an overlapping generations framework.
The new estimates are largely consistent with TPC’s previous analysis. They project the plan would boost economic output by a few tenths of a percent in 2021, when most Biden tax cuts would benefit low-income workers. But, it would reduce output by a few tenths of a percentage point on average over the 2022-2030 period, when it primarily raises tax rates on high-income households and corporations.
It may seem surprising that the estimates of both models are so similar, since their analytical methods are quite different. However, in some years the similar estimates stem from different effects.
For example, both approaches estimate a positive effect on output in 2021. In TPC’s traditional model, that growth results from an increase in aggregate demand due to Biden’s tax cuts. In the OG-USA model, the boost in short-term output stems primarily from high-income people working more in 2021 in anticipation of increased taxes in 2022.
The macroeconomic effects of the tax plan in turn influence the amount of revenue it would raise. When output rises, so do taxable incomes and revenues. The OG-USA model predicts lower revenues than TPC’s traditional model, mostly because those high-income simulated households will work less. Because those households would face relatively high marginal tax rates, the change in output has a relatively higher effect on revenues.
TPC is releasing two publications along with the new results. One paper explains the details behind TPC’s new analysis. The other paper performs a sensitivity analysis of the results. Looking ahead, we anticipate continuing to refine our methods to produce the best available macroeconomic estimates of proposed tax changes along with results from our microsimulation model.
