Arbetslöshetens bestämningsfaktorer i ekonomisk-historisk belysning - En analys av lönebildning, totalfaktorproduktivitet och löneutrymme under perioden 1911-1960.
This paper analyzes the Swedish labor market during the interwar and early postwar period within the framework of modern labor market theory. The development of unemployment during this period - according to the commonly cited source of labor union reports - represents a conundrum for research. The unemployment rate rose after the initial diverse shock of 1921 and stayed at a permanently higher level for the rest of the interwar period. This development was reversed after World War Two when the unemployment rate decreased and stayed permanently low for the rest of the postwar period until the oil price chock of the 1970s.In a first step the available sources of unemployment statistics is investigated and compared. The general conclusion is that the labor union reports overestimate the level of economy wide unemployment while being a reasonably good indicator of movements in the rate. While no assertion of absolute levels can be made the conclusion might be drawn that the equilibrium level of unemployment decreased from a higher interwar level down to a substantially lower postwar one. The paper then turns to the overarching question of the possible mainsprings of this development.The concept of the warranted wage - defined as total factor productivity growth divided by the labor share - have been used to explained the development of unemployment in the OECD from the 1970s. The theory pertains that movements in the bargained wage above or below the warranted wage will render movements in the equilibrium unemployment rate. This theoretical framework is used to analyze the Swedish inter- and early postwar experience. The warranted wage in the manufacturing sector and the whole economy is respectively estimated using historical national accounts and growth accounting. The development is then compared to the progress of real labor costs. The conclusion is that the 1920s experienced a negative growth in the warranted wage - and while real labor cost decreased during the period - wages were not cut enough in order to keep profits unchanged for firms. The opposite can be concluded for the succeeding 1930s and 1940s which instead saw a positive evolution of the warranted wage with real labor costs not growing at the same rate. The movements of real labor costs in relation to the warranted wage thus makes this factor a plausible candidate for explaining movements in the unemployment rate during the period understudy.