It is a widely held view that imperfect capital markets mean that individuals from poor backgrounds cannot borrow in order to finance educational investments. This view pervades policy formation, and is reflected in the fact that post-compulsory education processes in all countries involve considerable government intervention and large public subsidies. But are the existence of credit constraints an empirical reality? This paper uses unique data to take a new approach to this question. Specifically, the 1995 Canadian Out of Employment Panel (COEP) allows us to explore the financial resources and skill formation choices of a large number of recent job losers. This approach has several advantages, including: a direct test of the role of finances in determining training; the availability of considerable information concerning individual histories; and the fact that the unemployed are a particularly apposite group with which to explore the questions of credit constraints. We find that credit constraints do appear to limit the human capital investments of a significant minority of job seekers. In particular, controlling for a broad range of background characteristics (including past educational investments and labour market outcomes), the possession of liquid assets at the time of job loss is strongly associated with subsequent self-financed training. This basic finding is corroborated with several different kinds of evidence drawn from the survey. The data also allow us to make a rough estimate of the extent to which participation in training would have been increased, had no part of our sample been credit constrained.
