期刊名称：International Journal of Finance and Accounting
印刷版ISSN：2168-4812
电子版ISSN：2168-4820
出版年度：2016
卷号：5
期号：5A
页码：1-29
DOI：10.5923/s.ijfa.201601.01
语种：English
出版社：Scientific & Academic Publishing Co.
摘要：Empirical deadlocks Tobin's “q theory” had confronted initiated the various lines of research which have tried to improve the empirical performance of investment function, such as better measurement of q, structural estimation, and introduction of irreversibility or fixed costs in the adjustment process. We review and argue all of these developments successfully have captured certain aspects of investment behavior that previous theories cannot. However, there is no single model that can explain every aspect of investment alone, mainly because of substantial heterogeneity in investment behavior depending on the type of capital goods or the difference between new acquisition (positive investment) and sale/retirement (negative investment). In the second half of the paper, we estimate a non-linear version of the Multiple q investment function, which can explicitly handle the aforementioned heterogeneity, on the micro data of Japanese listed firms. We confirm our non-linear model dominates the traditional linear Multiple q model and find great dispersion in the range of non-linearity depending on time and the type of capital goods.
关键词：Capital goods heterogeneity; Lumpy investment; Multiple q; Non-linear investment function
