The Tax Administration Production Function: Experimental Evidence from Indonesia

Many countries face challenges raising sufficient tax revenue to finance their budgets
independently and provide critical infrastructure and services. While there has been extensive
research on optimal tax rates, there has been far less research on optimal tax administration
and governance structures. An important open question in tax administration and governance
is how to increase revenue most efficiently given the different inputs at the government’s
disposal. What are the returns to inputs like labor and capital in tax administration? Can
adding more people or money to tax offices increase revenue that exceeds their costs?
In collaboration with Indonesia’s Directorate General of Taxation, we are conducting two
large-scale experiments that randomly assigned a national sample of tax offices across
Indonesia to receive an increased budget for enforcement activities like taxpayer letters and
visits and/or more tax officers for two years. Using detailed anonymized administrative tax
microdata for millions of taxpayers and thousands of tax officers, we will estimate the effects
of these capital and labor shocks on tax enforcement activities, revenue, and compliance, as
well as their distributional implications for taxpayers. Together these studies will generate
experimental estimates of the marginal returns to labor and operating capital in taxation in the
same developing country setting, and shed light on the role of these inputs in tax
administration productivity and governance. Given the importance of increased revenue and
state capacity in an era of shrinking aid budgets, the results will provide practical guidance for
both policymakers in Indonesia and other developing countries.