Unemployment Dynamics in Britain

Abstract

We examine the inter-related dynamics of the unemployment flows and stock. Our model allows them to be linked through the standard accounting identity, but also through behavioural relationships. We find that, for the UK, the bulk of the correlation between duration and the unemployment stock arises from the dependence of duration on unemployment, taking careful account of the inter-temporal identity. Outflow shocks have contributed little to the evolution of unemployment since the late 1960s. Given that the correlation of duration and unemployment arises from the behavioural relationship, we show that unemployment dynamics arise mostly from shocks to the inflow. Unemployment remains one of Europe’s more pressing economic and social problems, as highlighted in the EU’s Lisbon Declaration of 2000. A dynamic approach to unemployment starts from the recognition that unemployment is a stock and changes through shocks to the outflow and inflow. The research and policy focus has generally been on the outflow process, or equivalently, unemployment duration. A glance at Figure 1 shows why this might be so: it reveals a very strong co-movement of the unemployment rate and the proportion of the unemployed out of work for a year or more. Analysts have used data like this to argue that the source of unemployment fluctuations has been changes in the outflow process. 1 Possibly as a consequence of this, most policies directed at reducing unemployment focus on improving the search effectiveness and employability of the currently unemployed. The unemployment flows and stock are related in quite complex ways, and in this paper we focus on these inter-related dynamics. The model we use allows them to be linked in two ways. First there is the simple accounting relationship, whereby shocks to the flows accumulate in the stock. Second, search and matching models suggest behavioural relationships of the flows depending on unemployment. The presence of the nonlinear accounting identity in the model makes the analysis of the joint system problematic, so we proceed sequentially, investigating first one link and then the other. We find that, for the UK, the bulk of the correlation between duration and the unemployment stock displayed in Figure 1 arises from the dependence of the former on the latter. We find a strong dependence of the outflow rate on unemployment, taking careful account of the inter-temporal identity. On the other hand, outflow shocks have contributed little to the evolution of unem