There is considerable anecdotal evidence of US companies moving from high-tax states to low-tax states, but what do the data reveal about the impact of state taxation on economic activity? Analyzing establishment-level data from the US Census Bureau for the period 1977-2011, this research finds that firms subject to state-level corporate taxation respond to higher corporate tax rates by closing establishments and reducing employment; those subject only to state-level personal income taxation respond similarly to individual income tax rates, though to a lesser extent. Since half of these responses are due to reallocation of business activity to lower-tax states, tax competition across states clearly plays a first-order role in corporate decision-making.
What is the impact of state taxation on economic activity? Anecdotal evidence abounds with examples of companies moving from high-tax states to low-tax states. A recent example is General Electric’s move from Connecticut to Massachusetts. Many commentators attribute this move to the increase in Connecticut’s corporate tax from 7.5% to 9% (for example, Kudlow, 2016).
Similarly, in the context of a proposal to increase corporate tax in Illinois, the head of the Illinois Manufacturers’ Association was quoted in 2011 in the Financial Times saying that ‘The only businesses that will benefit are the moving companies that will be helping many of my members move out of this particular state’.
Ultimately, whether and how state taxation affects companies is an empirical questionThese anecdotes are consistent with the view that state taxation may depress economic activity – or move it elsewhere – by putting a wedge between pre-tax and after-tax profits. Yet in principle, it could also be that state taxation does not matter.
That could be the case, for example, if the costs of moving to a different state outweigh the tax savings, or if firms have ways to manoeuvre around the taxes – for example, by negotiating tax subsidies under threat of moving out. Ultimately, whether and how state taxation affects companies is an empirical question.
Measuring the impact of state policy on firms’ location decisions
There are two main challenges to measuring the impact of state policy on firms’ location decisions. First, addressing this question requires detailed micro data on firms’ operations. Indeed, examining whether a given company relocates operations across states requires a dataset that specifies the location and size of each of its facilities.
Second, state taxation is likely to be correlated with changes in local economic conditions, making it difficult to attribute firms’ relocation decisions to changes in taxation. For example, it could be that a state is hit by a local economic shock (say, a local recession) and, as a result, firms leave the state. At the same time, the poor performance of the state may hurt the state budget, so that ultimately the state has no choice but to increase taxes to raise revenues.
In this scenario, we would observe an increase in state taxation and a decrease in economic activity. Yet this correlation would not imply that higher state taxes cause a decrease in economic activity. Instead, this correlation would simply reflect the fact that local economic shocks drive both economic activity and state taxation.
Our research addresses these two obstacles and ultimately aims to measure the extent to which state taxation affects economic activity. To overcome the first obstacle, we use establishment-level data from the US Census Bureau for the period 1977-2011. An establishment is a physical location with at least one paid employee. By considering all establishments of a given firm, we therefore know all their facilities and how many employees work at each facility.
To overcome the second obstacle, we use information on the tax filing status of companies to distinguish between two types of firms: C-corporations; and ‘pass-through entities’.
C-corporation profits are subject to the corporate income tax, and then their shareholders are subject to dividend or capital gains taxation when those individuals receive dividends or sell shares. The profits of pass-through entities – which include S-corporations, LLCs (limited liability companies) choosing to file taxes as S-corporations, and partnerships – are subject to the personal income tax: their profits ‘pass through’ to the owners who are then taxed at the personal rate.
This distinction allows us to account for the confounding effect of local economic shocks. Imagine the ideal set-up in which we have two companies that are very similar (and operate in the same states) but one is a C-corporation, and the other a pass-through entity. By construction, a change in the corporate tax in a given state should only affect the C-corporation.
Accordingly, by comparing the response of the C-corporation with that of the pass-through entity, we can disentangle the effect of the corporate tax (which only affects the C-corporation) and the effect of local economic conditions (which affect both the C-corporation and the pass-through entity).
In other words, the pass-through entity provides a counterfactual of what would have happened at the C-corporation in the absence of the change in the corporate tax. We further exploit this set-up to study the reverse – that is, how pass-through entities respond to changes in the personal tax, using C-corporations as a counterfactual.
Our analysis approximates this ideal set-up in which we examine first, the differential response of C-corporations versus pass-through entities following changes in the state corporate tax; and second, the differential response of pass-through entities versus C-corporations following changes in the state personal tax.
Effects of changes in state-level corporate and personal taxes
In our baseline analysis, we find that a one percentage point increase in the state corporate income tax leads to the closing of 0.4-0.5% of a C-corporation’s establishments in the state. In other words, if a state increases the corporate income tax by one percentage point (for example, if the rate increases from 5% to 6%), C-corporations would close down one establishment out of 200-250 in the state.
Remarkably, we find that half of these closures are offset by reallocation across states – that is, for each two establishments that are closed in the state, companies open one establishment in other states.
We obtain similar results – albeit smaller in magnitude – for pass-through entities with respect to the personal income tax. Specifically, we find that a one percentage point increase in the state personal income tax rate leads to the closing of 0.2-0.4% of a pass-through entity’s establishments in the state. Again, we find that about half of these closings are offset by the opening of new establishments in other states.
These results refer to the ‘extensive margin’ – that is, the closing and opening of establishments. We also study the ‘intensive margin’ – that is, changes in the number of employees within existing establishments.
An increase in a state’s corporate income tax leads C-corporations to close some establishments and reduce staff in others
Our findings are similar. For C-corporations, we find that a one percentage point increase in the state corporate tax leads to a 0.4% reduction in the number of employees within establishments of C-corporations; for pass-through entities (with respect to the state personal tax), we observe a 0.2% reduction. We find again that roughly half of these effects are offset by increases in the number of employees in the firms’ establishments in other states.
For the manufacturing sector, the establishment-level data of the US Census Bureau contains information about physical capital (plant and equipment). We use these data to examine changes in physical capital. We find that capital shows similar patterns to labor in its response to taxation, although not of a larger magnitude. This implies that firms respond to state taxes as much through changing the location of their labor inputs as through the reallocation of capital.
We conduct a series of extensions of our baseline analysis. Here, we describe three of them.
First, we focus on ‘large’ changes in state taxes, meaning changes of at least 100 basis points. The general patterns are consistent with our baseline results. A useful feature of this analysis is that we can separately examine tax increases and tax decreases. We find that the effects are approximately symmetric: while tax increases hurt business activity, the opposite is true of tax cuts.
Second, focusing on this subset of large tax changes, we apply the narrative approach of Romer and Romer (2010). Specifically, for each (large) tax change, we read newspaper articles around the time of the tax change to determine the ‘narrative’ and hence the rationale underlying the tax change. We then focus on the subset of tax changes for which the rationale was unlikely to be related to changes in economic conditions.
While tax increases hurt business activity, the opposite is true of tax cuts
Using this subset, we find that our results continue to hold. In further analysis, we consider a subset of these tax changes that were triggered by federal tax reforms – specifically, the Economic Recovery Tax Act of 1981 and the Tax Reform Act of 1986 – and hence were less likely to be related to the economic situation of individual states. We again find that our results are robust.
Third, we examine heterogeneity across firms. We find that our results are stronger for firms operating in ‘footloose industries’ (industries whose activities are less concentrated in specific states), for larger firms, and – among the larger firms – for multinational corporations. These results are intuitive since these firms are likely to be relatively more mobile. As such, their location decisions are likely to be more responsive to changes in state taxation.
In sum, our study indicates that firms respond substantially to changes in state taxation. C-corporations reduce the number of establishments per state as well as the number of employees and the amount of physical capital per plant when state-level corporate tax rates increase. Pass-through entities respond similarly to changes in state-level personal tax rates, although by a somewhat smaller magnitude.
Moreover, our results suggest that around half of these responses are due to reallocation of business activity to lower-tax states. This indicates that the notion of tax competition across states is of economic importance, and plays a first-order role in companies’ decision-making.
This article summarizes ‘State Taxation and the Reallocation of Business Activity: Evidence from Establishment-Level Data’ by Xavier Giroud and Joshua Rauh, which is forthcoming in the Journal of Political Economy.
Xavier Giroud is at Columbia Business School. Joshua Rauh is at Stanford University.