This paper estimates the intertemporal labor supply (Frisch) elasticity of substitution exploiting an unusual tax policy change in Switzerland. In the late 1990s, Switzerland switched from an income tax system where current taxes were based on the previous two years’ income to a standard annual pay as you earn system. This transition created a two-year long, salient, and well-advertised tax holiday. This change occurred both for the federal and local income taxes. Swiss cantons switched to the new regime at different points in time during the 1997–2003 period. Exploiting this variation in timing and using population-wide administrative social security earnings data matched with census data, we identify the Frisch elasticity. We find significant but quantitatively small responses of earnings with a Frisch elasticity of .05 overall. Some groups, such as high wage income earners and especially the self-employed display larger responses with Frisch elasticities of .1 and .27. We find no effects along the extensive margin at all and almost no effects on hours of work suggesting that responses are driven primarily by tax avoidance rather than real labor supply. Therefore, our estimates constitute upper bounds for the labor supply Frisch elasticity.