The effect of reporting institutions on tax evasion:Evidence from the lab

We study the effects of different tax reporting mechanisms in experi-mental double auction markets in the laboratory. The sales tax is paidby the seller, and we compare market outcomes in a no-tax conditionto cases where (i) tax evasion is impossible, (ii) taxes can be evaded butthere is an exogenous (low) audit probability, or (iii) there is double-reporting by both the buyer and the seller, and the seller’s audit prob-ability is endogenously increased if her tax report is inconsistent withthe buyer’s report. The latter case mimics the use of so called third-party reporting in tax enforcement. We find that third-party reportingeffectively deters evasion, and deterrence also has real effects on mar-ket outcomes: market clearing prices, quantities and overall efficiencyreturn to the levels observed when tax evasion was impossible. Whenreporting is costly to buyers, they report significantly less trades. Taxcompliance by sellers however remains at a relatively high level, eventhough payoffs would be maximized for both parties if no trades werereported. This suggests that the mere possibility of the existence ofthird party information may be a fairly effective deterrent on tax eva-sion, and tax administrators might consider making their informationsources more widely publicized.