Rothstein, J., (2005)

‘Does Competition Among Public Schools Benefit Students and Taxpayers? A Comment on Hoxby (2000)’

NBER Working Paper no. 11215

  • Reanalyses the data from Caroline Hoxby’s influential 2000 AER paper ‘Does Competition Among Public Schools Benefit Students and Taxpayers?’, which used instrumental variable analysis to show a significant effect of Tiebout school choice on public school productivity.
  • Hoxby famously used the number of larger and smaller streams in metropolitan areas as instruments for the level of Tiebout choice (hence competition) possible, arguing that as historical barriers to travel, areas with more streams will have more fragmented school districts, increasing choice.
  • The author changes Hoxby’s specifications to test the robustness of her results, and conducts an exhaustive search for coding errors and inadequate instruments.
  • Key results:

  • Many of the author’s alternative specifications give results which differ from Hoxby’s, with the coefficient on competition frequently not statistically different from zero.
  • He points out that Hoxby’s dataset incorrectly assigns some Ohio school districts to an incorrect metropolitan area.
  • He questions Hoxby’s use of larger and smaller streams as instruments for the level of Tiebout choice. He finds that the coefficient on larger streams (which he argues should have been a bigger barrier to travel than smaller streams) often shows a negative impact on choice.
  • He argues that Hoxby uses only streams’ sources in her instruments, not all the areas through which a stream flows.
  • He suggests that since high-scoring students may be more likely to attend private schools in areas with low choice/poor public schools, there may be a self-selection problem. He includes private school students in the sample to correct this, and finds choice effects are no longer significant.
  • The author concludes that it would be premature to conclude that schools respond to Tiebout competition by raising productivity.
  • This comment, and an earlier one by the same author, have drawn a vigorous response from Professor Hoxby.



Back to:

Top of Page

Page updated 13/02/2008 by Alison Taylor