Ten Years Hence: Business Angel Tax Credits, the Absent Employment Lift, and a Decade of Convergent Evidence
Kevin A. SchulteA 2016 doctoral dissertation employed a Campbell and Stanley dual-pair quasi-experimental design across matched state pairs to examine whether US state business angel tax credits produce the technology employment lift that constitutes their principal legislative intent. Across three independent treatment-control configurations (Kansas-Minnesota, LMM F = .026, p = .794; Wisconsin-Tennessee, F = .023, p = .879; and Wisconsin-Minnesota, p = .797), the null hypothesis of no employment effect was accepted at 95% confidence. Ten years later, the evidence base has grown decisively and convergently. Denes, Howell, Mezzanotti, Wang, and Xu (2023), publishing in the Journal of Finance across 31 US states using difference-in-differences methodology, confirm the null employment finding while identifying the mechanism: angel tax credits increase angel investment by approximately 18% but attract non-professional, inexperienced investors who provide capital without managerial support, producing crowding-out rather than net job creation. International evidence from the United Kingdom, Ireland, Germany, and Sweden sharpens the mechanism further: outcomes track the managerial engagement investors actually provide, and investor subsidies that attract inexperienced, unengaged investors produce the mal-investment signature predicted by Austrian Business Cycle Theory in 2016. Lerche (2025), in the American Economic Review, demonstrates that direct firm subsidies, bypassing the investor intermediary entirely, produce robust employment gains, completing the evidentiary picture. Meanwhile, the federal Angel Tax Credit Act has been introduced in Congress five times since 2014 without enactment, and over 30 states have collectively allocated more than $8.1 billion to these programs since 1989 (Denes et al., 2023), a decade of compounding null evidence notwithstanding. This article argues that the persistence is structural, not accidental.