Bureaucrats, Politicians, and Government Performance: Evidence from City Treasurers∗

This paper investigates whether assigning control of government services provision to a politician or bureaucrat affects the cost of government services. Conceptually, bureaucratic control could reduce costs if (a) costs are hard for voters to observe, weakening the incentives for politicians to exert effort to reduce costs, or (b) bureaucrats’ implicit incentives to reduce costs are strong due to career concerns or intrinsic motivation. I draw on the unique characteristics of California’s city referendum process to obtain clean identification of the causal effect of public official appointive status on costs for the case of city treasurers. I utilize a marginal referendum event study strategy based on the effect of narrowly-passing appointive city treasurer referendums on city borrowing costs. The results indicate that appointive treasurers reduce the interest rate a city pays to borrow by 29 percent. These effects do not appear to be driven by the responses to prior treasurer performance, other governmental reforms, or changes in city economic activity. The results suggest that if all cities in California with elected treasurers appointed them, then total city expenditures on borrowing would be reduced by $50 to $75 million per year. ∗I thank Leah Brooks, Tom Hansford, Shawn Kantor, Josh Kinsler, John Matsusaka, Kevin Milligan, Enrico Moretti, Ronnie Pavan, Todd Sorenson, Guido Tabellini, Jessica Trounstine, Christine Vuletich, Kelley Williams and seminar participants at the University of California Merced, the University of Rochester, and the University of Virginia for helpful comments and discussions. Matt Siordia and Chris Abrescy provided outstanding research assistance. All errors are my own. †Department of Economics; School of Social Sciences, Humanities and Arts; University of California, Merced. Phone: (209) 228-4027, Fax: (209) 228-4007, Email: [email protected]