How (Not) to Ask About Undeclared Work: Estimating Prevalence and Patterns Using a Survey Experimental Approach

Reliable estimates of undeclared work in Germany are scarce, yet they matter for tax and social security policy. Using two large-scale surveys of the German general population and recipients of transfer benefits particularly, we elicit the prevalence of the supply and demand of undeclared work via direct questions, a crosswise model and a list experiment. In line with economic theory, opportunity and incentives shape the decision to work undeclared. The self- and marginally employed as well as transfer recipients who face high effective marginal tax rates are particularly likely to work undeclared. The potential, especially for envelope wages, is substantial. Methodologically, the crosswise model consistently yields higher prevalence estimates for both supply and demand than direct questioning. We discuss assumptions and challenges when implementing list experiments.