In this paper, we present and extend previous theoretical results that show that tax authorities can at the same time reduce tax evasion and boost output with clever audit behavior. We argue that randomized controlled trials or field experiments are impractical for empirical testing in this context. Instead, we turn to laboratory experiments and test if humans follow the theoretical mechanisms underlying the theory. We find that both dividends, less evasion, and higher output, materialize in the laboratory. We show in additional experiments that the behavioral mechanism generating the higher output differs from the theoretical driver.