Welfare Added? Optimal Teacher Allocations with Value-Added Scores

• In Event: School Finance and Teacher Effectiveness

Abstract

How might value-added measures of teacher effectiveness improve teacher-assignment
policies? Intuitively, measuring how teachers’ impacts vary across student subgroups
and different outcomes could enable reallocations based on comparative advantage. At
the same time, estimating more effects could increase the risk of making allocations
based on estimation error. Using two decades of data from a large urban school district,
we compare allocations using estimates of teacher impacts across multiple outcomes
and student groups. While many reallocations produce gains over the status quo, sparse
models with only two outcomes that separately estimate teacher effects for above- and
below-median achieving students optimally trade of matching gains and misallocation
losses. The improved allocation would increase present-value earnings by $2000 per
student—39% more than an allocation ignoring subgroup effects and 34% more than
an allocation using only math scores as an outcome.

Author

• Michael Ricks, mricks4@unl.edu