A Panel Analysis of Liquidity Constraints and Firm Investment

Abstract

The issue of financial constraints on company investment is revisited using the U.S. panel data of 561 firms from 1971-1992. A number of economically meaningful factors are discovered to partition firms into relatively homogeneous groups. A mixed fixedand random-coefficients framework is then used to capture unobserved heterogeneity within groups. The prediction criterion is used to selecrthe final specification and evaluate the importance of financial constraints on firm's investment decisions.

Forecasting with panel data

“…Admittedly, these are additional case studies, but they do add to the evidence that simplicity and parsimony in model estimation offered by the homogeneous estimators yield better forecasts than the more parameter consuming heterogeneous estimators. Hsiao and Tahmiscioglu (1997) use a panel of 561 US fi rms over the period 1971-1992 to study the infl uence of fi nancial constraints on company investment. They fi nd substantial differences across fi rms in terms of their investment behavior.…”

How Do Firm Characteristics Influence the Relationship between R&D and Firm Value?

“…Galende and Suárez (1999) find support for the hypothesis that capital intensity increases the probability of firms carrying out R&D activities. Alternatively, there is evidence to suggest that firms with a high level of investment in physical capital face more financial constraints (Fazzari, Hubbard, and Petersen, 1988;Hsiao and Tahmiscioglu, 1997). In their seminal paper, Fazzari, Hubbard, and Petersen, (1988) argue that under financing constraints, capital expenditures are more sensitive to internal funds.…”

Homogeneity pursuit in panel data models: Theory and application

“…Conventional panel data analysis often assumes complete slope homogeneity, which is convenient in practical work and takes full advantage of cross-section averaging. However, homogeneity assumptions are frequently rejected in empirical panel studies, as in Hsiao and Tahmiscioglu (1997), Phillips and Sul (2007), Browning and Carro (2007), and Su and Chen (2013). But if complete slope heterogeneity is permitted, estimation can be imprecise or even impractical when the time dimension is very short, thereby losing a key advantage of working with panel data.…”