Essays in Public, Labor, and Financial Economics

Abstract

This dissertation comprises three chapters. The first chapter investigates the real effects of dividend taxation. "C"-corporations and "S"-corporations operate at the same scale and in the same narrow industries across the United States but only C-corps are subject to dividend taxation. Using hundreds of thousands of corporate tax returns and S-corps as a counterfactual, I show that (1) the 2003 dividend tax cut increased total C-corp payouts (dividends plus buybacks) by 45%, implying an elasticity with respect to one minus the tax rate of 1.4; (2) the payout response did not diminish after the 2004 reelection of President Bush and subsequent tax cut extension, undermining intertemporal tax arbitrage as a candidate explanation; and (3) the tax cut caused little or no increase in C-corp gross investment, net investment, total employee compensation, or number of employees. Unlike previous papers, these results reject both the "old" and "new" views of dividend taxation and instead point to some combination of the new view along with agency, tunneling, wealth reallocation, and tax avoidance effects. The second chapter uses the 1996 UC affirmative action ban to study whether and how nondiscrimination laws constrain decisions made behind closed doors. Seventeen years of law school applications reveal (1) pre-ban admissions offices used race, a novel conclusion from purely cross-sectional data; (2) the ban reduced observed black admissions advantages by two-thirds, implying under weak assumptions that the ban substantially reduced the use of race; (3) observed post-ban black advantages were nevertheless large; and (4) post-ban admissions offices used race in at least the first several years after the ban. These facts suggest that nondiscrimination laws can meaningfully constrain private selection decisions but that enforcement frictions may permit modest continued use of race. This paper's methods can improve civil rights litigation. The third chapter investigates why net flows into equity mutual funds are strongly procyclical. Investors' stated beliefs indicate they are trying to time short-run peaks and troughs, but I show they chase returns even with assets that are illiquid over a five-year or longer horizon. The time series of inflows suggests that individuals raise their expectations of the annualized long-run equity return by 10%, 6%, 5%, and 4% of the first-through-fourth annual S&P lags. Such beliefs imply frequent and dramatic revisions of the expected long-run equity return, including an 8-percentage-point reduction from 1999 to 2003.