We estimate Frisch elasticity in a labor market with high job turnover. In a context where only around 18% of the employed labor force has formal and stable jobs, we perform a fixed effects estimation as proposed by MaCurdy (1981) with a Heckman correction for selection into unemployment. We identify the positive slope of the labor supply using firms' size as an instrumental variable for wages. We use Peruvian data from the Permanent Employment Survey of Lima. We find that neglecting wage endogeneity implies a downward sloping labor supply, while the job turnover bias, not accounting for job turnover, overestimates Frisch elasticity. We estimate Frisch elasticity at around 0.38, which indicates fairly adjustable wages and little reaction of hours of work to wage variations. Moreover, we find that the Frisch elasticity is decreasing in income and tended to increase in the last decade.