Financial factors in the determination of private fixed capital accumulation: theory and evidence from the Turkish economy

This is an analysis of the investment behaviour of Turkish firms at the aggregate and sectoral levels. Despite the growing literature of empirical private investment studies, a theory-based approach to modelling investment is rare both for developing countries in general and for Turkey in particular. One of the most commonly used modelling strategies is to adapt the main elements of the neoclassical accelerator model subject to additional structural modifications for a developing country. However, many of these studies are eclectic in the sense that they are not based on any specification of the microeconomic optimisation problem of firms. The central purpose of the thesis is to develop econometrically estimateable investment functions based on a sound microeconomic framework. This study theoretically examines both the investment behaviour of firms and the role of financial decisions in investment behaviour in earlier chapters. Subsequent empirical chapters are built upon the main findings of the theoretical chapters, and accommodate three empirical models, all of which originate from an explicit optimisation problem of a representative firm. The first empirical model is derived from the cost minimisation of the firm subject to a given level of demand in the output market and two alternative production technology assumptions (namely putty-putty and putty-clay assumptions). The second model recognises the presence of adjustment costs of the capital stock, and develops an error-correction representation of an investment model from an quadratic cost function minimisation. These first two models analyse the roles of the neoclassical determinants of investment (the accelerator and the relative cost of capital) and credit constraints resulting from imperfections in capital markets in Turkey. However, the inclusion of the financial variable is less sound in these two models. The significance of capital market imperfections in the forms of a rising cost schedule of borrowing and quantitative constraints is the subjects of the third model. A dynamic model is developed through a maximisation of the intertemporal discounted cash flow of a representative firm subject to capital market imperfections, borrowing constraints and capital adjustment costs. An Euler equation for the rate of capital accumulation is derived by re-arranging the first-order condition for capital, which is influenced by the binding borrowing constraint through an unobservable shadow price. These three models are applied to annual aggregate and sectoral data from Turkey. The empirical findings of this thesis suggest that fiscal and monetary policies influence investment behaviour both via the relative cost of capital and via credit availability to the private sector in Turkey. In particular, the empirical results imply that a high interest rate policy in an imperfect capital market imposing extra cost on the market rate through the risk premium may have discouraging effects on investment decisions either through its impact on the user cost of capital or through the risk premium component of cost of borrowing as in the third model. However, high interest rate policies may also affect credit availability as postulated in the McKinnon-Shaw hypothesis