Business dynamism has been linked to innovation and employment creation, yet there is little empirical research on the relationship between business taxation and firms births and deaths. Using county-level panel data and a state border identification strategy, I document a negative effect of state corporate tax rates on establishment birth rate: an increase in the top state corporate tax rate of 1 percentage point leads to a decline in the birth rate of about 1.5-2%. The effect on exits is also negative but smaller in magnitude with a net-of-tax rate elasticity around 1. Using data at the census tract level, I show that spillovers are unlikely to be driving the main results, but can be large in areas close to the border – within 3 to 5 miles. These findings are robust to changes in other state level policies and sample restrictions.
"Structural Shocks, Robotization, and Local Public Finance"
I document the effect of increased exposure to robotization on local public finance outcomes - revenues, taxes, transfers, and expenditures - as well as the role of local fiscal autonomy and local tax limits. We find that an additional robot per thousand worker in the United States leads to a decline of roughly 1.5-2% in total local revenue, and around 2.5% in tax collections, while the overall effect on expenditure is around 1-1.5%. We show that these results are driven by states with high fiscal autonomy - defined as states that grant cities functional home rule - where the decline in revenues and expenditures is larger in magnitude. Using a newly collected data-set on property tax rates and assessment values in the United States, we show that in low autonomy areas, increased exposure to robotization leads to a relative decline in property tax rates, while rates are largely unaffected in high autonomy jurisdictions. We highlight the importance of accounting for policy changes when investigating outcomes that are highly policy dependent. Including property tax changes in our estimation shows that the mechanical effect of the shocks - through lower relative economic activity and property values - is similar across jurisdictions. The relative decline in low autonomy area creates a negative bias and the small increase in taxes in high autonomy area leads to a small positive bias on the effect of robotization on property tax collections.
Work in progress
"The Effect of the Great Recession on Local Property Taxes" (with Chiara Ferrero)
Using a newly collected dataset of property tax rates and assessment ratio across the United States between 2000 and 2017, we document how rates and home assessed values have evolved during and after the Great Recession. We look at policy responses as a size of the shock, measured by the decline in 1) local employment 2) local home values. We document the relationship between local public good provisions and local expenditures, and local property tax collections, rates and assessment at the local level. We look at the role of local fiscal autonomy and property tax limits, as well as the role of intergovernmental transfers (mainly from the ARRA) on overall property tax collections and local rules.
"The Effectiveness of Job Creation Tax Credits on State Employment"
Job creation tax credits (JCTCs) - subsidies given to firms who create new jobs - have become prevalent policies at the state level, with more than 120 JCTCs in place in 2015. We find some evidence that firms anticipate credits and shift hiring a few months or more to take advantage of the policy. However, JCTCs are not associated with a higher growth rate of employment during and after the Great Recession. The most effective credits are ones without additional investment in capital required and those that do not specify a minimum wage for eligibility. We estimate an average value of benefits per job created for the set of JCTCs introduced after 2000, and find that the average credit per job created is about $4,600 in 2002 dollars. The median is lower, around $3,800 in 2002 dollars, as a large number of JCTCs provide relatively low benefits. Credits of higher value do not seem to be more effective. Employment in firms of medium size - between 50 and 250 employees - shows the strongest reaction to the introduction of the policy.