A macro-financial theory of corporate financing and forecast modelling of market movements
This thesis identifies a gap in existing theories of corporate finance. This gap is an implication of a Keynesian-Minskian analysis of markets and market-based economies. From a founding theoretical perspective rooted in the view that markets are not reliably efficient the case is developed that past price trend extrapolation is an important factor in corporate financing decisions. At a macro-financial level, companies repurchase equity over periods of strong market rises, while increasing debt at the same time. During periods of sustained, substantial market decline debt is retired and large new equity issues occur. This change in corporate financing is implicitly expensive as relatively low prices are realised for the new stock issued at these times. These factors suggest that conventional theories of corporate financing decisions that rely on corporate rationality and optimisation do not provide a compelling fit with observations in the period 1980-2012. Moreover, inference to Minsky’s (1986) argument that companies are compelled through market declines to shore up their balance sheets provides a better fit with the evidence. These arguments form the basis for the development of the ‘extrapolative expectations’ theory of corporate finance. The second major development in this thesis draws on the theoretical development outlined above to create market movement description and prediction models. These models operate on data drawn from the US Standard & Poors 500 index over the period 1980-2012. Two primary models are developed using binomial logistic regressions. The dichotomous dependent variables are drawn as quarters of market rise (1) or no rise (0), and market falls (1) or no fall (0), respectively for the ‘buy’ model and the ‘sell’ model. Variables tested and those found to add to an explanation of the dependent variables include: corporate debt flows, corporate equity flows, corporate dividend flows, interest rates, market volumes, and historical market levels. Each variable is tested for up to ten lags (two-and-a-half years). Collectively, the variables add to our understanding of those factors influencing (or at the least, signalling) market levels, enabling quarter ahead market forecasts to be made with greater accuracy than arises from an assumption of a random walk. This conclusion crystallises the view that company macro-financial flows and prices are an important cause or signal of future market direction.