To estimate demand for labor, we use a combination of detailed employment data and the outcomes of procurement auctions, and compare the employment of the winner of an auction with the employment of the second ranked firm (i.e. the runner-up firm). Assuming similar ex-ante winning probabilities for both firms, we may view winning an auction as an exogenous shock to a firm’s production and its demand for labor. We utilize daily data from almost 900 construction firms and about 3,000 auctions in Austria in the time period 2006 until 2009. Our main results show that the winning firm significantly increases labor demand in the weeks following an auction but only in the years before the recent economic crisis. It employs about 80 workers more after the auction than the runner-up firm. Most of the adjustment takes place within one month after the demand shock. Winners predominantly fire fewer workers after winning than runner-up firms. In the crisis, however, firms do not employ more workers than their competitors after winning an auction. We discuss explanations like labor hoarding and productivity improvements induced by the crisis as well discuss implications for fiscal and stimulus policy in the crisis.