Scholz and Wood: Controlling the IRS

Scholz and Wood. 1988. Controlling the IRS: Principals, principles, and public administration. AJPS.

In Brief

Most studies of bureaucratic responsiveness assume that democracy requires political control of bureaucrats. But think again: Given that we question the extent of the public's control of politicians, why would we want perfect bureaucratic resopnsiveness? After all, we have a Constitutional system of checks and balances because we don't fully trust our politicians. Democracy, then, requires more than responsiveness; it also requires equity and efficiency.

Concerning the IRS, pure responsiveness might be a dangerous path to selective enforcement, politically-motivated audits (think of Nixon), and so on. Ideally, then, the IRS's decision of how to allocate auditing resources should correlate not only with political variables, but also with indicators of equity and efficiency.

In principle, Democrats might want the IRS to audit corporate returns more heavily, while Republicans might back off auditing the corporate returns. Thus, by examining the ratio of corporate to individual audits from 1974-92, the authors show that this ratio responds not only to national political control, but also to principles of equity and efficiency.

Operationalization and Findings

Y: IRS Audits

• The number of corporate audits in each state divided by the number of individual audits. A state-level measure.

X1: Responsiveness

National Level

• President: Dummies for Carter, Reagan, and Bush. Carter increased corporate audits slightly, Reagan decreased them, and Bush (strangely) increased them quite a bit.
• Congress: Control for the Chamber of Commerce scores for members of Senate Finance and House Ways and Means committees. More conservative committees push for fewer corporate audits.

State Level

• State presidential vote, governor's party, state legislative balance, partisan mix of state's Congressional delegation, etc. None of these variables matter, nor should they: They IRS was designed not to be influenced by state governments.

X2: Equity

• Look at the number of corporate returns filed, divided by number of individual returns. Principles of equity suggest that an increase in this ratio should lead to an increase in Y. The results are statistically significant, especially with a time lag: the IRS responds to a change after about 2 years.

X3: Efficiency

• Changes in expected revenue from corporate vs individual returns (the IRS posts these expectations). None of the first three lags is individually significant, though the three lags are collectively significant.

X2 and X3: Equity and Efficiency: Additional effects

• State-level controls: Income per capita, unemployment, manufacturing jobs per capita, service jobs per capita, farm jobs per capita, population density, poverty, etc. Though many of these variables are statistically significant, their collective effects contribute little to the model.