Stagnation of investment of non-financial firms in developed and underdeveloped EU countries, after the global financial crisis, has returned the issue of investment in the center of interest of scientific and professional community. What are the main determinants of investments, how to increase the investment of non-financial corporations, and how effective the domestic economic policy is in stimulating investment, are the current issues posed by economic theory and empiricism. Drop investments through net worth channel contributed to amplification and persistence of financial shock in Great recession. The main goal of this doctoral dissertation is to explore the fundamental determinants of investments in the Republic of Croatia, and in particular the investment channel of financial frictions. Two types of financial frictions are analyzed: the hypothesis of the existence of a firm financial constraints, and in particular, the hypothesis of the existence of the firm liquidity constraint. The heterogeneity of financial frictions with respect to ownership type and capital source is discussed. In last part of dissertation explains effect of debt forgiveness on investment and employment. For first two hypothesis panel threshold regression is used. The new variables to control aggregate demand I added to threshold regression and the sample include almost the entire population of companies in the Republic of Croatia. First two hypothesis was tested by assessing investment cash flow sensitivities for various degree of financial and liquidity constraints. The hypothesis of a significant degree of financial constraint was rejected when applying approach of Fazzaria et al. (1988). On contrary, the hypothesis of the existence of liquidity constraints was confirmed: there is nonlinearity in the sensitivity of investments to cash flow, as a function of liquidity. Inliquid firms have precautionary demand for cash, so additional cash is used as a buffer to liquidity shock. Firms at margin (liquid firms) shows highest sensitivity of investment to cash flows, while highly liquid firms show low sensitivities. Empirical research has also confirmed the existence of heterogeneity in imperfect capital markets with regard to the form of ownership and the origin of capital. Both group didn’t experience financial and liquidity constraint problems. The last hypothesis has shown the limited effect of the forgiveness policy and the prolongation of debts of non-financial corporations on boosting investment and employment growth. The last hypothesis has the characteristics of quasi-natural experiments which allow overcoming endogenous problems in estimating the investment equation. The results obtained emphasize the importance of designing domestic economic policy in a way that takes into account the significant role of heterogeneity in investment on imperfection of capital markets.