حساسیت های جریان وجوه نقد – سرمایه گذاری و ناپدید شدن سرمایه Investment-Cash Flow Sensitivities and Capital Misallocation

1 Introduction Inputs misallocation across heterogeneous production units lowers aggregate total factor productivity (TFP). A new and growing literature, as surveyed in Restuccia and Rogerson (2013), Önds that di§erence in allocative e¢ ciency may be an important explanation to the large and persistent cross-country income di§erences. Among various sources of misallocation, perhaps the single most studied mechanism is through Önancial frictions. Quantifying how much the observed capital misallocation can be accounted for by Önancial frictions is the central theme of a recent literature. 1 While modelling details and estimated magnitudes di§er, these studies share a common methodology: they develop theoretical models and gauge the size of TFP loss, by calibrating model parameters to match the distribution and dynamics of output across production units. In this paper, we propose an alternative accounting framework to estimate TFP loss due to Önancial frictions, using investment-cash áow sensitivity. Investment-cash áow sensitivity arises from a large body of empirical literature, which aims to test the presence of Önancial frictions. Following Fazzari et al. (1988), this literature adds a cash áow variable to a standard Q model of investment, and investigates the sensitivity of investment to cash áow across di§erent sub-samples of Örms. A common Önding is that there is a stronger correlation between investment and cash áow for sub-samples that are considered more likely to face Önancing constraint. This Önding has often been cited as evidence of signiÖcant capital market imperfections. Though investment-cash áow sensitivity is frequently used as an indicator of Önancing constraint, and Önancing constraint is one of the major sources of capital misallocation, there has not been any research, to our knowledge, that connects capital misallocation directly to investment-cash áow sensitivity. This paper attempts to Öll in this gap by providing a simple yet consequential theoretical model, which links the heterogeneity in investment-cash áow sensitivity, a common indicator of Önancing constraint, to the dispersion of marginal revenue product of capital (MRPK), a direct measure of allocative ine¢ ciency. We then apply this accounting framework to a panel of Chinese manufacturing Örms and calculate the aggregate TFP loss implied by the investment-cash áow sensitivities estimated from various sub-samples.