Evaluating the Effect of Teachers' Performance Incentives
on Pupils' Achievements


Victor Lavy

ABSTRACT

Proposals to use teachers' performance incentives as the basis for
school reforms have recently attracted considerable attention and
support among researchers and policy makers. The main message is that
the most likely way to improve students' achievements is to institute
performance incentives( direct monetary rewards for improvements in
student outcomes. However, there has been very little experience with
applying performance incentives in schools. This paper provides
empirical evidence on the causal effect of a program that offered
monetary incentives to teachers as a function of their students'
achievements. The program offered incentives to schools in the form of
performance awards, part of which were distributed to teachers and
school staff as merit pay and the rest( for the well-being of teachers
in the school. This teachers' incentives program fits closely the
framework of a rank order tournament: the scheme pays prizes to the
winners of a contest or a tournament in which the prizes depends on
the rank order of the winner. In this particular contest the distance
between winners is also taken into account in determining how the fix
prize is distributed among winners. I evaluate the effect of the
program during the first two full academic years of its
implementation, 1996 and 1997. Since participating schools were not
chosen randomly, the issue of identification is central to the
empirical strategy. Indeed, treated schools differed considerably, in
terms of pre-program characteristics and outcomes, from all other
secondary schools in Israel. However, the rules of selection into the
program provide a potential quasi-natural experiment that could form
the basis for credible identification strategy.

The second part of the paper evaluates the effect of a 'twin' program,
implemented simultaneously and with the same objectives as the
incentive program, but in different schools, and based on providing
additional resources to schools. The paper compares the effect and
cost of the two alternative intervention strategies and draws policy
implications. The results suggest that teachers' monetary incentives
had large positive effects, mainly in its second year of
implementation, on many dimensions of students' outcomes. However,
endowing schools with more conventional resources, such as additional
teaching time and on-the-job teacher training, also led to significant
improvement in student performance. The comparison based on
cost-equivalency suggests that the teachers' incentive intervention is
much more cost effective.