One of the best-known fallacies in economics is the notion that there is a fixed amount of work to be done – a lump of labor – which can be shared out in different ways to create fewer or more jobs. For instance, suppose that everybody worked 10% fewer hours. Firms would need to hire more workers. Hey presto, unemployment would shrink. In 1891, an economist, D. F. Schloss, described such thinking as the lump of labor fallacy because, in reality, the amount of work to be done is not fixed. Government-imposed restrictions on the amount of work people may do can actually reduce the efficiency of the labor market, thereby increasing unemployment. Shorter hours will create more jobs only if weekly pay is also cut (which workers are likely to resist) otherwise costs per unit of output will rise. Not all labor costs vary with the number of hours worked. Fixed costs, such as recruitment and training, can be substantial, so it will cost a firm more to hire two part-time workers than one full-timer. Thus a cut in the working week may raise average costs per unit of output and cause firms to buy fewer total hours of labor. A better way to reduce unemployment may be to stimulate demand and so increase output; another is to make the labor market more flexible, not less.