Estimating the effects of finance constraints and social learning on inventory investment, trade credit, and capital investment decisions

Conventional theories often cannot explain the observed volatility in aggregate inventories and capital investment. Two types of information-based models may provide alternative explanations: finance/credit constraints, and social learning models. Information asymmetry problems can cause finance constraints whereby credit access is reduced at certain times. This may cause sharp changes in spending by constrained firms, and could account for fluctuations in aggregate spending. Social learning models describe agents learning from others. Widespread social learning can cause actions to be clustered together in time as agents attempt to benefit from watching others before acting. Abrupt changes can result due to small changes in underlying fundamentals if many agents suddenly act in response to others. Thus social learning models can generate large changes in aggregate spending.; The thesis tests three separate models of finance constraints and social learning. Chapters one and two test for finance constraints with inventory and trade credit data. Chapter four investigates social learning and investment by semiconductor plants.; I do not find that finance constraints affect inventory investment in a sample of publicly-traded Canadian manufacturing firms. However, with the much larger sample of Canadian firms used to study trade credit, I find clear evidence of credit constraints. Medium-wealth firms substitute trade credit for bank credit. A one standard deviation decrease in bank credit increases their trade credit use by 0.9%. The wealthy firms' trade credit usage is independent of bank credit, which implies that they are not finance-constrained. The poorest firms are constrained in both trade credit and bank credit. This research suggests that from 1988 to 1998, most Canadian firms faced some degree of credit rationing, producing a procyclical pattern in aggregate trade credit use.; Social learning appears to influence semiconductor plants' decision to adopt new technology. A 1% increase in the number of other plants that have suspended their investment project generates a 3% increase in the probability of suspension by a plant adopting new technology. Firms investing in conventional technology seem to engage in a "war of attrition" since they are less likely to suspend their project if other suspensions occur.