Recent research: Angel tax credits not showing economic impact

In a new working paper, Sabrina T. Howell of New York University and Filippo Mezzanotti of Northwestern University provide a systematic review of state angel tax credits. One of the most notable aspects of their research is a seemingly-comprehensive index of all of the relevant programs authorized by states over the past 30 years. The results indicate that angel tax credits have some impact on investment activity but not on economic outcomes. The authors provide evidence that the reason for this seeming discrepancy could be due to program design allowing existing activity to benefit from the new credits.

The summary of state angel tax credit programs should be useful for any policymakers still considering these programs. The authors provide two dozen points of comparison, including starting and end years, eligibility for both investors and companies, and the credit’s share of all angel investment in the state (there are difficulties with this data, which is largely based on self-reports from AngelList, but is still interesting as a rough guide). These comparisons are just factual, of course, and do not necessarily indicate which states have designed better incentives.

This dataset is then subjected, along with other data on investment and economic activity, to statistical analysis. The authors report that state angel tax credits produce:

• An increase in number of reported angel deals;
• An increase in number of reported investors in local companies;
• No impact on employment by young firms; and,
• No impact on creation of new firms.

One would expect that if the credits are generating as much as 30 percent more deals, that there would be some measurable impact on startup employment or creation. To test whether the programs are just too small to be noticeable by aggregate analyses, the authors test firm-level employment and do not see significant employment growth relative to other young companies.

Another plausible explanation for the lack of impact is that the investments are helping firms grow through means other than employment. To test this hypothesis, the authors find that firms receiving tax credit-enabled investments are no more likely than other firms to receive venture capital investment within two years.

The authors’ explanation for the finding that angel tax credits generate some activity without achieving economic impact is that the credits are providing tax opportunities for insiders. Evidence for this assertion includes that startups benefitting from the credit are more likely to already be well-funded and, at least in the five states that identify both the companies and investors, 20 percent of the investors were employees of the company at the time they received the tax credit. The authors also point to investments in companies outside of IT as evidence of malfeasance, although this activity outside of the most common high-growth sector could have an economic development rationale.

As with any study of early-stage investment, the authors are hamstrung by data quality. Angel investment data is typically self-reported and always incomplete, and venture capital data, while better, also misses many investments. In recognition of these challenges, the paper does an admirable job of using multiple analytic approaches to test each hypothesis. Further, one might suspect that investors receiving tax credits would be easier to identify than investors not receiving the credits, which would suggest that the authors have presented a best-case scenario for the credits’ value.

Ultimately, this article suggests states using or considering these programs should be particularly cautious and carefully measure the investment and economic activity of both the startups benefitting from the credits and those that do not.