Does financial liberalization reduce financing constraints?

Luc Laeven

I use panel data on a large number of firms in 13 developing countries to find out whether financial liberalization relaxes financing constraints of firms. I find that liberalization affects small and large firms differently. Smallfirms are financially constrained before the start of the liberalization process, but become less so after liberalization. Financing constraints of large firms, however, are low before financial liberalization, but become higher as financial liberalization proceeds. I hypothesize that financial liberalization has adverse effects on the financing constraints of large firms, because these firms had better access to preferential directed credit during the period before financial liberalization.

In this study, I explore the impact of financial market reforms on financing constraints of firms in developing countries. Although financial reform has been one of the most profound policy reforms of many developing countries in recent years and is thought to have a large potential impact on the performance of firms and on economic welfare more generally, there is no professional consensus on the net benefits of financial liberalization.

From a theoretical perspective, financing constraints may arise if there are financial frictions. Under the Modigliani and Miller theorem (1958) (without financial frictions) a firm’s capital structure is irrelevant to its value. In this case, internal and external funds are perfect substitutes and firm investment decisions are independent from its financing decisions. With financial frictions, however, the costs of internal and external finance will diverge. Financial frictions can arise from information asymmetry and from incomplete contracting. (1) Financial frictions lead to a link between net worth, the cost of external financing, and firm investment. Within the neoclassical investment model with financial frictions, an increase in net worth independent of changes in investment opportunities leads to greater investment for firms facing high financial frictions and has no effect on investment for firms facing negligible financial frictions. Firms facing high financial frictions are, thus, expected to face financing constraints. It follows that financial liberalization will reduce financing constraints if it is accompanied by a reduction in financial frictions, such as improved contract enforcement or a reduction of information costs.

Financial reforms have consisted mainly of the removal of administrative controls on interest rates and the scaling down of directed credit programs. Barriers to entry in the banking sector have often been lowered as well and the development of securities markets has been stimulated. In this paper, I focus on the effects of liberalization of banking markets.

Although the main objectives of financial deregulation should be to increase the supply and improve the allocation of funds for investment, the consequence of financial liberalization on the supply of funds for investment is theoretically ambiguous. At the macro level, the effect of financial liberalization on the level of household savings, a major part of the total funds available for investment, is unclear. McKinnon (1973) and Shaw (1973) study the effect of interest rate liberalization, a key component of financial reform, on the supply of household savings. Starting from a repressed financial system in which governments intervene by keeping interest rates artificially low and replace market with administrative allocation of funds, they argue that interest rate liberalization is likely to lead to an increase in interest rates. As higher interest rates on deposits are likely to encourage household savings, they favor interest rate liberalization. Van Wijnbergen (1983), on the other hand, argues that the existence of informal credit markets can reverse the effect of an increase in interest rates on the total amount of savings. The effect of an increase in the deposit rate on the amount of loanable funds depends on whether households substitute out of informal market loans or cash to increase their holdings of time deposits. If time deposits are closer substitutes for informal market loans than for cash, then the supply of funds to firms could fall, given that banks are subject to reserve requirements and informal markets are not. Devereux and Smith (1994) show that financial liberalization may also negatively affect the level of precautionary savings as a result of improved international risk sharing, thereby reducing the overall level of funds available for investment.

The consequence of financial liberalization on the efficiency of the allocation of funds for investment is also theoretically ambiguous. On the one hand, it is often thought that financial reforms improve the allocative efficiency of savings. McKinnon (1973) and Shaw (1973) argue that interest rate ceilings, another common feature of repressed financial systems, distort the allocation of credit and may lead to under investment in projects that are risky but have a high expected rate of return. Similarly, directed credit programs are often associated with a misallocation of funds. The liberalization of financial markets is also thought to lead to an increase in the pool of funds allocated towards risky investment projects as a result of improved risk sharing (Obstfeld, 1994). It is also often thought that financial liberalization creates efficiency gains through increased financial inter-mediation by the formal financial sector. Under the presence of economies of scale in information gathering and monitoring, banks and capital markets are expected to have an advantage over the informal market in allocating investment funds, which should lead to a reduction in the cost of capital. However, Gertler and Rose (1994) claim that financial liberalization has failed to meet these expected efficiency gains in a number of countries, because accompanying a general rise in interest rates was a rise in the cost of capital for a substantial class of borrowers. Gertler and Rose (1994) also argue that the elimination of subsidized credit programs, another common feature of financial reform, could increase the financing constraints of those firms that previously benefited from the directed credit system.

More generally, financial liberalization is thought to reduce the imperfections of financial markets, resulting in a reduction in the cost of capital and an increase in the level of investment (Henry, 2000a,b and Bekaert and Harvey, 2000). Financial liberalization is also often associated with increased financial development (McKinnon, 1973). Increased financial development in turn has been shown to enhance economic growth (King and Levine, 1993 and Beck, Levine, and Loayza, 2000, among others), as well as the growth of industries that depend on external finance (Rajan and Zingales, 1998). Wurgler (2000) finds a direct link between financial development and the efficiency by which capital is allocated towards investment. Bekaert, Harvey, and Lundblad (2001) have found a direct link between financial liberalization and economic growth. The above theories have in common that financial liberalization changes the composition and allocation of savings, but will not necessarily relax financial constraints for all c lasses of firms.

The current empirical findings about the effects of financial liberalization on financing constraints in developing countries have also been inconclusive. A positive effect of financial liberalization on the financing constraints of firms, as measured by a decrease in the sensitivity of firms’ investment to changes in internal cash flow, has been found for Indonesia by Harris, Schiantarelli, and Siregar (1994) and for Chile by Gallego and Loayza (2001). For Mexico, Gelos, and Werner (2002) find that financial constraints were eased during financial liberalization for small firms, but not for large ones. They argue that large firms might have had stronger political connections than small firms and hence better access to preferential directed credit before financial deregulation. For Ecuador, on the other hand, Jaramillo, Schiantarelli, and Weiss (1996) find no evidence of a change in financing constraints after financial reform. This difference in the results may be explained by the fact that financial liberal ization in Ecuador was less profound than in other countries, but could also be due to the fact that financial liberalization in Ecuador benefited only certain firms. The result for Ecuador may also be explained by the use of firm-level data for a relatively short period of time, while the effects of liberalization may only be felt over a longer period of time.

It is clear that there can be distributional consequences to programs of financial liberalization, and whether they relax financing constraints for different categories of firms is ultimately an empirical question. This paper aims to address this question. In particular, I pose the following two questions: 1) does financial liberalization relax the financing constraints of firms, and 2) what types of firms are most likely to benefit from financial liberalization? I contribute to the existing literature by using panel data for a large number of firms in 13 developing countries to analyze the effects of financial liberalization on firm investment and financing constraints, rather than focusing on one single country, as has thus far been done in the literature. The use of data from several countries has the advantage of making my result less dependent on the specifics of a single country.

I find for my sample of firms in 13 liberalizing developing countries that, on average, firms did not face financing constraints over the sample period of 1988 to 1998. I also find that, on average, financial liberalization did not reduce the financing constraints of firms, as firms’ investment remained sensitive to changes in firms’ cash flow after financial liberalization. However, financial liberalization appears to have affected small and large firms differently. Large firms become more financially constrained as financial liberalization progresses, while small firms become less financially constrained. The effects are economically large. Financial liberalization reduces the financing constraints–as measured by the sensitivity of investment to cash flow–of small firms by around 80% on average. The negative effect of financial liberalization is found to be the strongest for the group of very large firms.

By studying the effects of financial liberalization on the financing constraints of firms, I focus on firms’ access to funds for investment, rather than the allocative efficiency of total available funds. Therefore, my results cannot be used to infer whether financial liberalization has a positive impact on the overall efficiency of investment. Although empirical work by Cho (1988) for the case of Korea and by Galindo, Schiantarelli, and Weiss (2002) for several developing countries suggests that financial liberalization has been associated with a substantial improvement in the efficiency of credit allocation in these countries, allocative efficiency is intrinsically hard to measure and I will, therefore, leave this topic to future research.

The paper continues as follows. Section I reviews the literature on financing constraints. Section II presents the structural model of firm investment that I use to estimate the impact of financial liberalization on financing constraints of firms. Section III describes the econometric techniques I employ to estimate my structural model of firm investment. Section IV presents the firm-level data used in my empirical work. Section V presents the results of my empirical work. Section VI concludes.

I. Review of Literature on Financing Constraints of Firms

Following the work of Fazzari, Hubbard, and Petersen (1988), a large body of literature has emerged to provide evidence of financing constraints. I refer to an overview of this literature to the surveys by Schiantarelli (1995); Blundell, Bond, and Meghir (1996); Hubbard (1998); and Bond and Van Reenen (1999). This literature relies on the assumption that external finance is more costly than internal finance due to asymmetric information, agency problems, and contracting costs, and that the premium on external finance is an inverse function of a borrower’s net worth. A firm’s investment is defined to be financially constrained if a windfall increase in the supply of internal funds (i.e., a change which conveys no new information about the profitability of current investment) results in a higher level of investment spending.

Following Fazzari et al. (1988), it is usually assumed that there are cross-sectional differences in effects of internal funds on firms’ investment, so that the investment equation should hold across adjacent periods for a priori unconstrained firms but be violated for constrained firms. This has led to different a priori classifications of firms that have tried to distinguish financially constrained and not-constrained firms. From a theoretical point of view, such sorting criteria should focus on a firm’s characteristics that are associated with information or contracting costs. A number of studies have grouped firms by dividend payouts (Fazzari et al., 1988 and Hubbard, Kashyap, and Whited, 1995); other a priori groupings of firms have focused on the degree of shareholder concentration, group affiliation (Hoshi, Kashyap, and Scharfstein, 1991), size and age (Devereux and Schiantarelli, 1990), the presence of bond ratings (Whited, 1992), or the pattern of insider trading (Oliner and Rudebusch, 1992).

Such a priori classifications are usually assumed to be fixed over the sample period. This may be problematic if the characteristics of firms change over time, or if the criteria used to split the sample are not external to the investment model. In addition, Lamont (1997) and Shin and Stulz (1998) have shown that the finance costs of different parts of the same corporation can be interdependent in such a way that a firm subsidiary’s investment is significantly affected by the cash flow of other subsidiaries within the same firm.

Kaplan and Zingales (1997) question the usefulness of a priori groupings of firms. They divide the firms studied by Fazzari et al. (1988) into categories of financially constrained” to “financially constrained” basedupon statements contained in annual reports, and find no support for the presence of financing constraints. The problem with their analysis is that it is difficult to make such classifications. Fazzari, Hubbard, and Petersen (2000) note that the firms Kaplan and Zingales (1997) classify as most financially constrained are actually firms in financial distress. From a theoretical point of view, the Kaplan and Zingales (1997) critique is limited to the claim that differing cash flow sensitivities reveal different degrees of financing constraints under the alternative hypothesis that these types or firms are both subject to significant financing constraints.

Several empirical investment models have been derived from the investment Euler equation that can be used to test empirically the presence of financing constraints. Most studies on financing constraints since Fazzari et al. (1988) estimate a q-model of investment, pioneered by Tobin (1969) and extended to models of investment by Hayashi (1982). Financial variables such as cash flow are then added to the q-model of investment to pick up capital market imperfections. If markets are perfect, investment should depend on marginal q only. Marginal q is usually measured by average q (see Fazzari et al., 1988; Hayashi and Inoue, 1991; and Blundell, Bond, Devereux, and Schiantarelli, 1992). Hayashi (1982) has shown that only under certain strong assumptions, marginal q equals average q. These assumptions are that the firm is a price-taker in all markets with constant returns to scale in both production and installation. Also, the usual implementation of the q-model assumes that stock markets are strongly efficient. If firms operate in imperfect capital markets, average q will be a poor proxy for investment opportunities. Furthermore, the q-model of investment is not identified if cash flow is a good proxy for future investment opportunities as well. Gomes (2001) shows that in this case significant cash-flow effects can arise even in the absence of financial frictions.

In many cases, the conditions required to equate marginal q to average q do not hold, or stock markets are not strongly efficient. In particular in the case of low-income countries, where capital markets tend to be imperfect, the q-model of investment is likely to be specified incorrectly.

An alternative approach, introduced by Abel and Blanchard (1986), forecasts the expected present value of the current and future profits generated by an incremental unit of fixed capital. Gilchrist and Himmelberg (1995, 1998) have extended this approach by using a vector auto-regression forecasting framework to decompose the effect of cash flow on investment. The advantage of the Abel and Blanchard (1986) model is that it avoids the use of share price data. The disadvantage of this approach is that it needs to assume a certain stochastic process on the variables of the investment model. As Bond and Van Reenen (1999) point out, “how well this can be done using the information available to the econometrician is not entirely clear.”

Several researchers have departed from the strategy of estimating marginal q and estimate the so-called Euler investment equation describing the firm’s optimal investment directly, after using the first-order condition for investment to eliminate the shadow value of capital from the Euler equation. This is the so-called Euler model of investment introduced by Abel (1980), and applied by Hubbard and Kashyap (1992), Whited (1992), Bond and Meghir (1994), and Hubbard, Kashyap, and Whited (1995), among others. The advantages of the Euler approach are that it avoids the use of share price data and can relax the assumption of linear homogeneity of the net revenue function. Perhaps more importantly, this approach avoids the need to parameterize the expectations-formations process as in the case of both the q-model of investment and the Abel and Blanchard (1986) model.

In this paper, I use the Euler equation approach. Given that my sample consists of firms from developing countries with mostly imperfect capital markets, application of the q-model of investment seems problematic. I do not use the Abel and Blanchard (1986) model to avoid the need to parameterize the expectations-formations process.

Empirical work has found that financial variables such as cash flow are important explanatory variables for investment. These findings are attributed to capital market imperfections that arise from informational asymmetries, costly monitoring, contract enforcement and incentive problems. Most studies of financing constraints focus on firms in one country. One of the first cross-country studies is by Bond, Elston, Mairesse, and Mulkay (1997), who study firms’ investment behavior in Belgium, France, Germany and the UK, and find that financial constraints on investment are more severe in the UK than in the three other countries. Mairesse, Hall, and Mulkay (1999) study firms’ investment behavior in France and the US and find significant changes in the investment behavior of French and US firms over the last twenty years. Love (2003) studies the relationship between financial development and financing constraints on a firm-level for a sample of 40 countries. She finds a strong negative relationship between the sen sitivity of investment to the availability of internal funds and an indicator of financial market development, and concludes that financial development reduces the effect of financing constraints on investment. This result provides evidence for the hypothesis that financial development reduces informational asymmetries in financial markets, which leads to an improvement in the allocation of capital and ultimately to a higher level of growth.

In related work, Demirguc-Kunt and Maksimovic (1998) directly identify those firms that are financially constrained by estimating if the firm’s sales grow at a rate that requires long-term external financing. They investigate how differences in legal and financial systems affect firms’ use of external financing to fund growth, and show that in countries with efficient legal systems a greater proportion of firms use long-term financing.

Bond and Van Reenen (1999) point out that, although most of the results of the financing constraints literature appear to be consistent with priors on the extent of financing constraints for some types of firms in some periods, these tests could also be detecting other sources of misspecification in the underlying investment models. My results, therefore, do come with provisos.

II. Methodology to Test for the Presence of Financing Constraints

In this section, I present a model of investment that allows for financial frictions known as the Euler model of investment. I use this model to estimate the financing constraints of firms. I follow closely the setup in Bond and Meghir (1994). For simplicity, I ignore bankruptcy costs and taxes. For a model that incorporates bankruptcy costs and taxes see Bond and Meghir (1994).

I assume that the firm maximizes its present value, which is equal to the expected value of future dividends, subject to the capital accumulation identity and external financing constraints. Let [V.sub.t] be the market value of the firm’s outstanding shares at time t, [K.sub.t], the firm’s capital stock at time t, and [B.sub.t] the firm’s net financial liabilities. Financial frictions are incorporated via the assumption that debt is the marginal source of external finance, and that risk-neutral debt holders demand an external finance premium, [[eta].sub.t] = [eta]([B.sub.t]), which is increasing in the amount borrowed, [partial][eta]/[partial]B > 0, due to agency costs. The idea is that highly indebted Firms have to pay an additional premium to compensate debt holders for increased costs due to information asymmetry problems. I assume that the gross required rate of return on debt s (1 + [r.sub.t])(1 + [eta]([B.sub.t])), where [r.sub.t] is the risk-free rate of return. For debt rather than equity to be :he firm’s ma rginal source of finance, I introduce a non-negativity constraint on dividends, which implies that there is a shadow cost associated with raising new equity due to information asymmetry. Another way to introduce financial frictions would be by limiting the amount of debt that the firm can raise at any point in time as in Whited (1992) or by introducing bankruptcy costs as in Bond and Meghir (1994). The advantage of my approach is that the equation characterizing the optimal path for investment is completely unaffected by the introduction of debt finance into the model, so that I do not need to make any specific assumptions about the firm’s optimal borrowing policy. The profit function is denoted by [[PI].sub.t] = [PI]([K.sub.t], [L.sub.t], [I.sub.t]), where [L.sub.t] is the vector of variable inputs, and [I.sub.t], the firm’s investment at time t. The capital stock accumulation depends on the investment expenditure [I.sub.t] and the depreciation rate [delta]. Dividend paid out to shareholders is denoted by [D.sub.t].

Under these assumptions, the manager’s problem is to maximize:

[V.sub.t] = [E.sub.t] [[summation over ([infinity]/j=0)] [[beta].sup.t.sub.t+j][D.sub.t+j]] (1)

subject to:

[D.sub.t] = [[PI].sub.t] + [B.sub.t] – (1 + [r.sub.t-1])(1 + [eta]([B.sub.t-1]))[B.sub.t-1] (2)

[K.sub.t] = (1 – [delta])[K.sub.t-1] + [I.sub.t] (3)

[D.sub.t] [greater than or equal to] 0 (4)

where [E.sub.t][*] is the expectations operator conditional on time t information,

[[beta].sup.t.sub.t+j] = [[PI].sup.j.sub.i=1][(1 + [r.sub.t+i-1]).sup.-1]

the j-period discount factor for j [greater than or equal to] 1, and [[beta].sup.t.sub.t] = 1. Let [[lambda].sub.t] be the Kuhn-Tucker multiple for the non-negativity constraint on dividends. This multiplier can be interpreted as the shadow cost of internal funds. Substituting Equation (2) into Equation (1) for [D.sub.t], and using the Identity (3) to eliminate [I.sub.t], from the profit function [[PI].sub.t] = [PI]([K.sub.t], [L.sub.t], [I.sub.t]), the first-order condition for capital [K.sub.t] can be calculated as (see Appendix A):

-(1 – [delta])[[beta].sup.t.sub.t+1][E.sub.t][(1 + [[lambda].sub.t+1])[([partial][PI]/[partial]I).sub.t+1]]= -(1 + [[lambda].sub.t])[([partial][PI]/[partial]I).sub.t] -(1 + [[lambda].sub.t])[([partial][PI]/[partial]K).sub.t] (5)

In line with the literature on investment models I refer to Equation (5) as the Euler equation for investment. The Euler equation is an intertemporal condition relating current investment to last period’s investment and marginal product of capital.

The first-order condition for debt requires that (see Appendix A):

1 + [[lambda].sub.t] = [E.sub.t][(1 + [[lambda].sub.t+1])(1 + [[eta].sub.t] + [partial][[eta].sub.t]/[partial][B.sub.t] [B.sub.t])] (6)

It follows from Equation (6) that the marginal cost of debt determines the relationship between [[lambda].sub.t] and [[lambda].sub.t+1].

If [D.sub.t] > 0, the firm generates sufficient net revenue to finance its investment from retained earnings and pay positive dividends. In this case, the Kuhn-Tucker multiplier [[lambda].sub.t], which measures the shadow value of internal finance, is zero, and the Euler investment equation in Equation (5) collapses to the one in the standard model of investment without financial frictions (see Bond and Meghir, 1994 for a derivation):

-(1 – [delta])[[beta].sup.t.sub.t+1][E.sub.t][([partial][PI]/[partial]I).su b.t+1]]= -[([partial][PI]/[partial]I).sub.t] -[([partial][PI]/[partial]K).sub.t] (7)

To obtain an empirical model of investment, I follow Bond and Meghir (1994) and specify the firm’s net revenue function as:

[[PI].sub.t] = [p.sub.t]F([K.sub.t],[L.sub.t]) – [p.sub.t]G([I.sub.t],[K.sub.t]) – [w.sub.t][L.sub.t] – [[p.sup.I.sub.t][I.sub.t], (8)

where F([K.sub.t], [L.sub.t]) is a constant returns to scale production function, G([I.sub.t], [K.sub.t]) is a convex adjustment-cost function of installing [I.sub.t] units of capital, which is assumed to be homogeneous in investment and capital, [p.sub.t] is the price of the firm’s output, [w.sub.t] is the vector of prices for the variable inputs, and [p.sup.I.sub.t] is the price of investment goods. I denote net output as [Y.sub.t] = [F.sub.t], – [G.sub.t] and assume that [Y.sub.t] is homogeneous in both [K.sub.t] and [L.sub.t]. To allow for imperfect competition, I let [p.sub.t] depend on net output [Y.sub.t], with the price elasticity of demand [epsilon] assumed constant and [epsilon] > 1. Differentiating Equation (8) with respect to [I.sub.t] and [K.sub.t] gives (see Appendix B for a derivation):

[([partial][PI]/[partial]I).sub.t] = -[alpha][p.sub.t][([partial]G/[partial]I).sub.t] – [p.sup.I.sub.t] (9a)

and

[([partial][PI]/[partial]K).sub.t] = -[alpha][p.sub.t][[([partial]F/[partial]K).sub.t] – [([partial]G/[partial]K).sub.t]] (9b)

where [alpha] = 1 (1/[epsilon]) > 0.

Following the tradition in the literature since Summers (1981), I specify:

G([I.sub.t], [K.sub.t]) = 1/2b[K.sub.t][[(I/K).sub.t]-c].sup.2]

as a symmetric adjustment-cost function, which is strictly convex in investment, and homogeneous of degree one in investment and capital. With this adjustment-cost function, it follows (see Appendix B for derivations) from Equation (9a) that:

[([partial][PI]/[partial]I).sub.t] = -b[[alpha]p.sub.t] [(I/K).sub.t] + b [[alpha]cp.sub.t] – [[p.sup.I.sub.t] (10a)

and from Equation (9b) that:

[([partial][PI]/[partial]K).sub.t] = [[alpha]p.sup.t] [(Y/K).sup.t] -[[alpha]p.sub.t][([partial]F/[partial]L L/K).sub.t] + [b [alpha]p.sub.t] [(I/K).sup.2.sub.t] -[bc [alpha]p.sub.t][(I/K).sub.t] (10b)

Following Bond and Meghir (1994), I assume that the marginal product of variable factors ([partial]F/[partial]L) can be replaced from the first-order conditions by w/[alpha]p. Substituting Equations (10a) and (10b) into Equation (5) and setting [[lambda].sub.t] and [[lambda].sub.t+1] to zero gives the empirical Euler investment equation under the null of no financial frictions:

[(I/K).sub.t+1] = c(1 – [[phi].sub.t+1]) + (1 + c)[[phi].sub.t+1] [(I/K).sub.t] – [[phi].sub.t+1] [(I/K).sup.2.sub.t] + [[phi].sub.t+1]/b([epsilon] – 1) [(Y/K).sub.t] – [[phi].sub.t+1]/b[alpha][(CF/K).sub.t] +[[phi].sub.t+1]/b[alpha] [J.sub.t] + [v.sub.t+1], (11)

where [[phi].sub.t+1] = (1 + [[rho].sub.t+1])/(1 – [delta]) with [[rho].sub.t+1] = (1 + [r.sub.t+1])([p.sub.t]/[p.sub.t+1])-1 being the real discount rate, [(CF/K).sub.t] = ([p.sub.t][Y.sub.t] – [w.sub.t][L.sub.t])/([p.sub.t][K.sub.t]) is the ratio of cash flow (operating profit) to the capital stock, [J.sub.t] = ([p.sup.I.sub.t]/[p.sub.t]){1 – [p.sup.I.sub.t+1](1 – [delta])/[(1 + [r.sub.t])[p.sup.I.sub.t]} is the user cost of capital, [Y.sub.t] = [F.sub.t] – [G.sub.t] denotes net output, [alpha] = 1 – (1/[epsilon]) > 0 and [v.sub.t+1] is an error term. Note that under perfect competition, the price elasticity of demand goes to infinity, [epsilon] [right arrow] [infinity], and the output term [(Y/K).sub.t] is eliminated from the Euler investment equation. Under imperfect competition, the coefficient on this term is positive and 1 < [epsilon] < [infinity].

I test for the presence of financing constraints by estimating the following empirical specification of Equation (11):

[(I/K).sub.it] = [[beta].sub.t][(I/K).sub.i,t-1] + [[beta].sub.2][(I/K).sup.2.sub.i,t-1] + [[beta].sub.3][(Y/K).sub.i,t-1] + [[beta].sub.4][(CF/K).sub.i,t-1] + [d.sub.t] + [f.sub.i] + [v.sub.it] (12)

where I allow for time-specific effects [d.sub.t] and firm-specific effects [f.sub.i]. The subscript i refers to the company, and t refers to the time period. The testable implications of Equation (12) are as follows. Under the null of no financial frictions, Bond and Meghir (1994) show that [[beta].sub.1] > 1, [[beta].sub.2] < -1, [[beta].sub.3] [greater than or equal to] 0, and [[beta].sub.4] < 0. If any of these restrictions are not met, investment does not follow its optimal path. In particular, when investment responds positively to an increase in cash flow (i.e., when [[beta].sub.4] is positive) the firm is financially constrained in the sense that a windfall increase in cash flow implies that a higher level of investment can be funded before internal finance is exhausted. Therefore, I consider a firm to be more financially constrained if the cash flow coefficient, [[beta].sub.4], is estimated to be more positive. Like Bond and Meghir (1994), I drop the user cost of capital from my empirical specificat ion since it is difficult to measure.

As in Harris et al. (1994); Jaramillo, Schiantarelli, and Weiss (1996); and Gelos and Werner (2002), I test whether small firms are more financially constrained than large firms. In addition, I test whether both small and large firms have become less financially constrained during the process of financial liberalization. Large firms are likely to be less financially constrained than small firms, because lenders are likely to have more information about large firms. Those borrowers also are likely to have relatively more collateral wealth. Size considerations may also affect the access to directed credit programs at subsidized rates, because such programs often favor exporting firms, which are often large firms, and because large firms often have stronger political connections.

III. Estimation Techniques

Dynamic investment models are likely to suffer from an endogeneity problem arising from the presence of unobserved firm-fixed effects. Consider the following model:

[y.sub.it] = [alpha][y.sub.it-1] + [beta]'[x.sub.it] + [gamma]'[f.sub.i] + [u.sub.it], (13)

where:

[u.sub.it] = [[eta].sub.i] + [v.sub.it] (14)

and:

E([v.sub.it] [x.sub.i0],…, [x.sub.iT], [[eta].sub.i]) = 0 (15)

where [f.sub.i] is an observed individual effect and [[eta].sub.i] is an unobserved individual effect. In this model, the existence of unobserved individual effects implies that [y.sub.it-1] is an endogenous variable.

To solve the estimation problem raised by the potential presence of unobserved individual effects, one can estimate the specific model in first-differences. If I remove the unobserved individual effect by first-differencing Equation (13), I obtain:

[DELTA][y.sub.it] = [alpha][DELTA][y.sub.it-1] + [beta]'[DELTA][x.sub.it] + [DELTA][v.sub.it] (16)

The use of Generalized Methods of Moments (GMM) estimation is required, because [DELTA][v.sub.it] is correlated with [DELTA][y.sub.it-1] by construction, and joint endogeneity of the explanatory variables might still be present. For a more detailed introduction to Generalized Methods of Moments (GMM) estimators for dynamic panel data models, I refer to Hansen (1982); Holtz-Eakin, Newey, and Rosen (1988); Arellano and Bond (1991); and Arellano and Bover (1995). Under the assumptions that the error term [v.sub.it] is not serially correlated and the explanatory variables are weakly exogenous, the following moment conditions apply to the lagged dependent variable and the set of explanatory variables:

E([y.sub.it-s][DELTA][v.sub.t]) = 0 [for all]s [greater than or equal to] 2; t = 3,…, T (16)

E([x.sub.it-s][DELTA][v.sub.t]) = 0 [for all]s [greater than or equal to] 2; t = 3,…, T, (17)

so that ([y.sub.it-2], [y.sub.it-3],…, [y.sub.i1]) and ([x.sub.it-2], [x.sub.it-3],…, [x.sub.i1]) are valid instruments. I refer to this estimator as the difference estimator. Arellano and Bond (1991) have shown that under the above assumptions the difference estimator is an efficient GMM estimator for the above model. Blundell and Bond (1998) show that, when the dependent variable and the explanatory variables are persistent over time, lagged levels of these variables are weak instruments for the regression equation in differences.

To assess the validity of the assumptions on which the GMM difference estimator is based, I consider two specification tests suggested by Arellano and Bond (1991). The first is a Sargan test of over-identifying restrictions that tests the validity of the instruments. The second is a test of second-order serial correlation of the error term. The use of endogenous variables dated t – 2 as instruments is only valid if the error term is serially uncorrelated, implying a first-order moving average error term in the differenced model.

IV. Data

To explore the impact of financial reforms on financial constraints of firms, I need a measure of financial liberalization and firm-level data. I construct an index of domestic financial liberalization of the banking sector based upon country reports from various sources. The problem of constructing such an index is that financial liberalization often takes place in various ways and in stages.

I construct the financial liberalization variable as follows. I collect data on the implementation of reform packages related to six different measures. The liberalization variable is simply the sum of six dummy variables that are each associated with one of the six reform measures. The dummy variables take value one in the years characterized by the liberalized regime. Hence, my index of financial liberalization can take values between 0 and 6. The index is not strictly comparable across countries in absolute terms. For example, there is likely to be a significant difference in the initial stage of financial liberalization among the countries in my sample. However, since increases in my index of financial liberalization capture progress in financial liberalization within a country, the index is comparable across countries in relative terms. The six reform measures I focus on are: interest rates deregulation (of both lending and deposit rates), reduction of entry barriers (both for domestic and foreign banks) , reduction of reserve requirements, reduction of credit controls (such as directed credit and credit ceilings), privatization of state banks (and more generally a reduction of government control), and strengthening of prudential regulation (such as independence of the Central Bank or adoption of capital adequacy ratio standards according to the Basle Accord guidelines). These measures correspond to the domestic financial liberalization measures in Bandiera, Caprio, Honohan, and Schiantarelli (2000), who use principal components to construct an index of financial liberalization for eight developing countries.

I focus on a sample of developing countries because progress in financial liberalization has been substantial in these countries over the last decade, and because financial liberalization is expected to have a large impact on the financing constraints of firms in these countries as the financial frictions in these countries are large.

Table I indicates the years in which significant progress been made with respect to one of these six measures. The Annex to this paper describes in more detail what types of progress have been made in these years with respect to one of these six measures. Table II presents the financial liberalization index (FLI) for a number of countries.

A number of clear patterns arise from the evolution of the financial liberalization index. First of all, all developing countries in my sample have made substantial progress in liberalization of their banking sectors. A number of countries had repressed financial systems in the 1980s, but could be considered liberalized in 1996. Secondly, the index suggests that countries liberalize their financial systems gradually and in stages. In most countries, interest rates are liberalized and reserve requirements are reduced in the first stage of liberalization. In a second stage, entry barriers are removed and directed credit systems (and other forms of credit control) are eliminated. Only in the final stage are state banks privatized and is prudential regulation put into place. This sequence of financial liberalization is presented in Table III in more detail.

Williamson and Mahar (1998) have found a similar progress in financial liberalization for these countries. In fact, if I define a country’s financial system to be largely liberalized in the year when significant progress has been made with respect to five of my six measures of financial liberalization (that is when ELI takes value 5), I find a similarity with the years in which Williamson and Mahar (1998) consider a country’s financial system to be largely liberalized. Table IV presents this comparison.

Although I focus on the liberalization of banking markets, the period under consideration has not only been characterized by liberalization of this segment of financial markets. Developing countries have implemented many different types of reform programs during this period under changing political climates. In addition to liberalization of the banking sector, one key component of financial reform in most developing countries has been the liberalization of the stock market. Table IV shows the dates from which point onwards Bekaert and Harvey (2000) consider the stock markets of these countries to be open to foreigners. It follows that stock market liberalization has preceded liberalization of the banking sector in most countries, Chile being the only exception. As stock market liberalization tends to precede banking market liberalization, inclusion of stock market liberalization in my index of financial liberalization does not change my results. Another reason for focusing on the liberalization of banking mar kets is that liberalization of this segment of financial markets is expected to have the greatest impact on financing constraints of firms as bank loans tend to account for the majority of firm debt.

Progress in financial liberalization seems to be strongly correlated with improvements in the political climate of a country. If I use the Political Risk Index of the International Country Risk Guide (ICRG) as a measure of political risk, I find a correlation of as high as 66% between the political risk rating and my financial liberalization index (see Table V). The ICRG political risk index is constructed by the Political Risk Service Group, ranges from zero to 100%, and is decreasing in the level of political risk. This suggests that financial liberalization tends to take place during periods of political stability. Laeven and Perotti (2001) provide a more extensive analysis of the link between political risk and financial liberalization.

I collect firm-level panel data from Worldscope on firms in developing countries for the years 1988-98.1 collect the data from the December 1999 CD-Rom of the Worldscope database. Using panel data has certain advantages: first, it allows it to differentiate across firms. As explained before, it is likely that firms are treated differently in a regime of financial repression (for example, due to directed credit programs). It is also likely that the effects of liberalization differ across firms according to their size and other factors. This is so because, as explained by Schiantarelli, Atiyas, Caprio, Harris, and Weiss (1994), the alternative to a financially repressed system is not a perfect capital market, but a market for funds characterized by informational asymmetries and less than complete contract enforceability, which gives rise to agency problems whose severity varies for different types of firms. Second, the availability of panel data allows it to identify more precisely the effects of financial libe ralization over time, which is attractive since financial reform is often a process over a longer period of time.

I focus on listed firms, since most firms in the Worldscope sample are listed, and because the quality of the accounting data is higher for listed firms. For each company I need several years of data to assess changes over time in the financing structure of the firm. Therefore, I delete firms with less than three consecutive years of data. I also delete firms in transition economies, because soft budget constraints that have been inherited from the socialistic regime may distort the analysis, and I delete firms from countries with data on less than 20 firms. Firms that operate in the financial and service industries (primary US Standard Industrial Classification codes 6, 7, 8 or 9) are also deleted from the sample as their investment and financing behavior is not comparable with firms that operate in other industries. It is, however, necessary to delete more firms, because of outliers in the data. Such outliers can be due to revaluation of assets, divestments, acquisitions, or simply poor quality of the data. I impose a number of outlier rules. First of all, I delete observations with negative or zero fixed capital or investment. Such observations might be due to divestments or revaluations of capital, which would bias my results. Secondly, I delete investment ratios with extreme values. Such values might be due to acquisitions or revaluations of capital. In particular, I delete observations with a ratio of investment to capital that does not take a value between 0.1 and 0.5. Furthermore, I omit observations with extreme values of marginal profitability or cash flow. In particular, I restrict the sales-to-capital ratios to take values between 0.1 and 10, and I restrict the cash flow-to-capital ratios to take values between 0.01 and 1. After deleting firms according to these criteria, I have data for 394 listed firms in 13 countries.(2) Obviously, my sample of firms is non-random. Listed firms, for example, tend to be large in most countries.

For this set of firm-level data, I generate the necessary variables to estimate Equation (12). I assume that flow variables (such as investment and depreciation) during period tare chosen by the firm at the beginning of period t. Since accounting data only provides end-of-period data, I use end-of-period t-1 figures to construct variables at the beginning of period t.

I follow the notation in Section II to define the regression variables. Let [K.sub.t] be net fixed assets at time t, which includes property, plant and equipment net of depreciation; [Y.sub.t] is net sales at time t; [I.sub.t] is net investment at time t, which is calculated as [K.sub.t+l] + [Depr.sub.t] – [K.sub.t](1 + [[pi].sub.t]), where [Depr.sub.t] is depreciation at time t and [[pi].sub.t] is inflation at time t; [CF.sub.t] is operating cash flow at time t, which is calculated as operating income at time t plus depreciation at time t; and [FLI.sub.t] is the financial liberalization index at time t. The financial liberalization index takes values between zero and six, with a higher value indicating a larger degree of liberalization.

To test for a difference in financing constraints between firms of different size, I split my sample according to firm size. As measure of firm size, I use net sales reported in US dollars for comparability across countries. I construct a small size dummy, [Small.sub.t], that takes value of one if net sales at time t is smaller than the sample median of net sales, and zero otherwise. Similarly, I construct a large size dummy that indicates large firms. In an alternative specification, the small size dummy, [Small.sub.t], takes value of one if net sales at time t is smaller than the 75th percentile of sales in the sample, and zero otherwise.

Table VIa presents descriptive statistics of the regression variables for all firms. I have data for 394 firms and for the period 1988 to 1998. The average data coverage for each firm is 4.2 years, with a total number of firm-year observations of 1,645. The period coverage for specific firms ranges from three years of observations to ten years of observations. When comparing the descriptive statistics of small and large firms, I find that large firms have, on average, higher investment-to-capital ratios, higher sales-to-capital ratios, and lower cash flow-to-capital ratios than small firms (not reported). None of these differences are, however, statistically significant. I also find a strong correlation of 61% between operating cash-flow and sales (not reported in the tables). Investment appears to be mostly correlated with cash flow (with a statistically significant correlation of 18%) and less so with sales (with a correlation of 11%). These correlations suggest that firms are financially constrained in the sense that investment responds positively to cash flow. However, since cash flow is correlated with the sales measure, this conclusion may be false. Econometric techniques are needed to determine the exact effect of cash flow on investment.

Table VIb presents the median of the regression variables by country. Given the variation in investment ratios across countries, I allow for the presence of fixed country-effects in my empirical analysis. Table VIc presents the median of the regression variables by industry. The industries are defined according to the Standard Industry Classification (SIC) codes of the US government. I group manufacturing companies in my sample along two-digit SIC codes and the remaining industries along one-digit SIC codes. For my sample of firms, I find significant differences in the variables across the different industries. Some of these differences are not a surprise. For example, cash flow is highest in the tobacco industry–not a surprise given that the tobacco industry is in general considered a “cash cow”. Differences across industries may, however, be partly due to the small sample size for some industries. In my empirical analysis, I allow for the presence of fixed industry-effects.

Table VId presents the median of the variables by year. In general, I see no dramatic changes in the variables overtime. One exception is the level of investment in 1998, which is significantly lower than before. This can be explained by the fact that a number of countries in my sample faced a financial crisis in 1998 which might have reduced the number of investment opportunities for some firms. In my empirical analysis, I include year dummies to correct for such differences over time.

V. Empirical Results

I estimate several specifications of the structural investment model in Equation (11). I estimate all model specifications using the GMM difference estimator described in Section III. Estimation has been carried out using the DPD program developed by Arellano and Bond (1988). I only present two-step GMM estimates, since they are more efficient than one-step estimates, and since the Sargan test of over-identifying restrictions is heteroskedasticity-consistent only if based on the two-step estimates. Further details on the one-step and two-step GMM estimators can be found in Arellano and Bond (1991). All regressions include time dummies. Industry and country dummies were completely insignificant once I controlled for fixed effects. As proxy for net output, [Y.sub.t], I use net sales, and as proxy for cash flow [CF.sub.t], I use operating cash flow, defined as operating income plus depreciation.

First, I estimate Model (12) in differences using the aforementioned GMM difference estimator. The results are presented in Column 1 of Table VII. The instrument set includes the right-hand side variables dated t-2 and t-3. Under the null hypothesis of no financial frictions, it should be that [[beta].sub.1] > 1, [[beta].sub.2] < -1, [[beta].sub.3] [greater than or equal to] 0, and [[beta].sub.4] < 0. However, the first restriction is not met, as the estimated coefficient of [[beta].sub.1] is significantly smaller than one, and therefore investment does not follow its optimal path. Apparently, persistence in investment is not as strong as suggested by my theoretical model. The cash-flow coefficient is not positive and, therefore, I do not find evidence that firms faced financing constraints on average over the whole sample period. I also find that investment reacts positively and significantly to changes in sales. The positive and significant coefficient on net output indicates imperfect competition in the product market. The Sargan test of over-identifying restrictions does not reject the validity of the instruments. The test of second-order serial correlation of the error term indicates that the error term does n ot exhibit second-order serial correlation.

In the remainder of Table VII, I investigate the sensitivity of inverstment to cash-flow further. In Column 2 of Table VII, I test whether financial liberalization has had an impact on the sensitivity of investment to cash flow. For this purpose, I interact the financial liberalization index (FLI) with the right-hand side variables:

[(Y/K).sub.t] and [(CF/K].sub.t]

in Model (12) and estimate the expanded model as before. I do not find that progress in financial liberalization has had a positive impact on the sensitivity of investment to cash-flow for the average type of firm. The FLI-interacted variables are statistically insignificant. As before, I find that the dynamics implied by the model are not met, because the coefficient on the lagged dependent variable is significantly smaller than one. Again, the Sargan test of over-identifying restrictions does not reject the validity of the instruments, and the test of second-order serial correlation of the error term does not indicate that the error term exhibits second-order serial correlation.

Next, I test whether firm size affects the sensitivity of investment to cash-flow and whether financial liberalization has affected small and large firms differently. To this end, I interact the variables in Model (12) with both a small size dummy variable and the financial liberalization index (FLI). I define the small firms as firms with sales below the median of sales in the sample and construct the small size dummy variable accordingly. I estimate the expanded model as before using the GMM difference estimator. The estimation results are presented in Column 3 of Table VII. I find that financial liberalization affects small and large firms differently. While large firms become more financially constrained (as indicated by an increased sensitivity of investment to cash-flow) as financial liberalization progresses, small firms become less financially constrained. On the other hand, before the start of financial liberalization, small firms appear more financially constrained than large firms, although the dif ference is not statistically significant. Again, I find that both the Sargan test of over-identifying restrictions and the test of second-order serial correlation are not rejected.

The estimated effect of cash flow on the investment of firms is economically important. Before financial liberalization, the cash-flow coefficient is–0.36 for large firms and 0.12 for small firms. Once financial liberalization has progressed, say to the median level of the financial liberalization index, or ELI equals 4, the cash-flow coefficient increases to 0.14 for large firms and reduces to 0.02 for small firms. Although these differences are not statistically significant, they point at substantial economic differences in the effect of financial liberalization on the cash-flow sensitivity of investment between small and large firms. Financial liberalization reduces the estimated effect of cash flow on investment from around 12% to 2% for small firms. In other words, financial liberalization reduces financing constraints of small firms by around 80%. On the other hand, the cash-flow sensitivity of large firms increases over the same period from-36% to 14%.

My findings are in line with earlier work that has found that smaller companies are more likely to suffer from financing constraints before financial liberalization (see Schiantarelli, 1995). In particular, Gelos and Werner (2002) find similar results for firms in Mexico, and argue that large firms may have had better access to preferential directed credit before financial liberalization. This might also explain why I do not find that financial liberalization has an overall positive effect on large firms. The positive effect of more efficient financial markets may have been more than offset for large firms by the negative effect of a decreased access to preferential credit.

As robustness check, I assess the sensitivity of my results to the definition of firm size. Using the median of the sample to distinguish between large and small firms may not be representative, since the distribution of firms is highly skewed. Therefore, I also use the 75TH percentile of net sales to distinguish between small (below the 75th percentile) and large firms (above the 75th percentile). The GMM estimates of the model with this alternative size dummy variable are presented in Column 4 of Table VII. The results are qualitatively similar, but stronger from a statistical point of view, to the results based on the sample median. Again, I find that financial liberalization reduces the financing constraints of small firms, but increases the financing constraints of large firms. Comparing the results in Columns 3 and 4 of Table VII, I find that the latter effect is much larger for firms above the 75th percentile of sales than for firms above the median of sales, suggesting that the negative effect of fina ncial liberalization on the reduction of financing constraints is strongest for the group of very large firms (above the 75th percentile).

VI. Conclusions

I have estimated a dynamic investment model using annual panel data for 394 listed firms in 13 developing countries for the period 1988 to 1998. I find that financial liberalization affects small and large firms differently. Before financial liberalization takes place, small firms are found to be more financially constrained than large firms. Financial liberalization then relaxes the external financing constraints of small firms, but increases the financing constraints of large firms. It seems that only small firms in developing countries gain from financial liberalization. I hypothesize that, in many developing countries, large firms had access to preferential (directed) credit during the period before financial liberalization. This form of favoritism is likely to decrease during financial liberalization. In the case of large firms, the efficiency benefits of financial liberalization thus seem to be offset by the adverse effects of loosing access to preferential credit.

Whether financial liberalization improves the allocative efficiency of financial resources more generally, is a question beyond the scope of this paper. Using my financial liberalization index, Galindo et al. (2002) find that there is a strong positive correlation between financial liberalization and improvements in allocative efficiency of investment, although not in all countries.

I also find that countries that have made substantial progress in liberalizing their financial sectors have shown dramatic improvements in their political climate as well. Successful financial liberalization seems to require both the political will and ability to stop the preferential treatment of well-connected firms, firms that often tend to be large.

Appendix A. Mathematical Derivation of Equations (5) and (6)

When I substitute the sources and uses of funds Equation (2) into (1) for [D.sub.t], and use the capital accumulation Identity (3) to eliminate [I.sub.t] from the profit function [[PI].sub.t] = [PI]([K.sub.t],[L.sub.t],[I.sub.t]), I can write the Lagrangian for the model in Equations (1) through (4) as:

[max.sub.{[K.sub.t+j],[L.sub.t+j],[I.sub.i+j],[B.sub.t+j]}] [E.sub.t][[summation over ([infinity]/j=0)] [[beta].sup.t.sub.t+j] {(1 + [[lambda].sub.t+j])([[PI].sub.t+j] ([K.sub.t+j], [L.sub.t+j], [K.sub.t+j] – (1 – [delta])[K.sub.t+j-1]) + [B.sub.t+j] -(1 + [r.sub.t+j-1])(1 + [eta]([B.sub.t+j-1])}] (A.1)

where [[lambda].sub.t+j] is the Kuhn-Tucker multiplier for the non-negativity constraint on dividends. Since [B.sub.t+j] does not depend on [K.sub.t] and only [[PI].sub.t] = [PI]([K.sub.t], [L.sub.t], [K.sub.t] – (1 – [delta])[K.sub.t-1]) and [[PI].sub.t+1] = [PI]([K.sub.t+1], [L.sub.t+1], [K.sub.t] – (1 [delta])[K.sub.t-1]) depend on [K.sub.t], the first-order condition for capital [K.sub.t] can be calculated as:

0 = [partial]/[partial][K.sub.t] {(1 + [[lambda].sub.t])[[PI].sub.t] ([K.sub.t], [L.sub.t], [K.sub.t] – (1 – [delta])[K.sub.t-1]) + [[beta].sup.t.sub.t+1][E.sub.t][(1 + [[lambda].sub.t+1])[[PI].sub.t+1] ([K.sub.t+1], [L.sub.t+1], [K.sub.t+1] – (1 – [delta])[K.sub.t])]} (A.2)

It follows that the first-order condition for capital [K.sub.t] is:

(1 + [[lambda].sub.t])[([partial][PI]/[partial]K).sub.t] +(1 + [[lambda].sub.t])[([partial][PI]/[partial]I).sub.t] -(1 – [delta])[[beta].sup.t.sub.t+1])[([partial][PI]/[partial]I).sub.t+1]] = 0 (A.3)

where [I.sub.t] = [K.sub.t] – (1 – [delta])[K.sub.t-1]. Moving the first two terms in Equation (A.3) to the right-hand sight of the equation gives Equation (5) in the text.

Similarly, it follows from Equation (A.1) that the first-order condition for debt [B.sub.t] can be written as:

0 = [partial]/[partial][B.sub.t] {(1 + [[lambda].sub.t])[B.sub.t] – [E.sub.t][[[beta].sup.t.sub.t+1](1 + [[lambda].sub.t+1])(1 + [r.sub.t])(1 + [eta]([B.sub.t]))[B.sub.t]]} (A.4)

With [[beta].sup.t.sub.t+1] = 1/1+[r.sub.t] it follows that the first-order condition for capital [B.sub.t] is:

1 + [[lambda].sub.t] -[E.sub.t][(1 + [[lambda].sub.t+1])(1 + [[eta].sub.t] + [partial][[eta].sub.t]/[partial][B.sub.t] [B.sub.t])] = 0 (A.5)

which can be re-written as Equation (6) in the text.

Appendix B. Mathematical Derivation of Equations (9a), (9b), (10a), and (10b)

Since [p.sub.t] depends on output Y([K.sub.t],[L.sub.t],[I.sub.t]) = F([K.sub.t], [L.sub.t]) – G([I.sub.t], [K.sub.t]), differentiating Equation (8) with respect to [I.sub.t] gives:

[([partial][PI]/[partial]I).sub.t] = -[p.sub.t][([partial]G/[partial]I).sub.t] + [([partial]p/[partial]Y [partial]G/[partial]I).sub.t] [Y.sub.t] – [p.sup.I.sub.t] = -[1 -[(Y/p [partial]p/[partial]Y).sub.t][p.sub.t][([partial]G/[partial]I).sub.t] – [p.sup.I.sub.t] = -[1 -1/[epsilon]][p.sub.t][([partial]G/[partial]I).sub.t] – [p.sup.I.sub.t] = -[[alpha]p.sub.t][([partial]G/[partial]I).sub.t] – [p.sup.I.sub.t] (A.6)

which equals Equation (9a) in the text.

Similarly, differentiating Equation (8) with respect to [K.sub.t] gives:

[([partial][PI]/[partial]K).sub.t] = [p.sub.t] [[([partial]F/[partial]K).sub.t] -[([partial]G/[partial]K).sub.t]] -[([partial]p/[partial]Y), [[([partial]F/[partial]K).sub.t] -[([partial]G/[partial]K).sub.t]][Y.sub.t] = [1 -1/[epsilon]][p.sub.t][[([partial]F/[partial]K).sub.t] -[([partial]G/[partial]K).sub.t]] = [[alpha]p.sub.t][[([partial]F/[partial]K).sub.t] -[([partial]G/[partial]K).sub.t]] (A.7)

which equals Equation (9b) in the text.

With the adjustment-cost function specified as G([I.sub.t], [K.sub.t]) = 1/2b[K.sub.t][[[(I/K).sub.t] -c].sup.2], I have:

[([partial]G/[partial]I).sub.t] = b[(I/K).sub.t] -bc (A.8)

Substitution of Equation (A.8) in (9a) gives Equation (10a). Under the assumption that [Y.sub.t] is homogeneous of degree one in both [K.sub.t] and [L.sub.t], I have:

F([K.sub.t], [L.sub.t]) = [([partial]F/[partial]K).sub.t] [K.sub.t] + [([partial]F/[partial]L).sub.t] [L.sub.t]

or

[([partial]F/[partial]K).sub.t] = F([K.sub.t], [L.sub.t])/[K.sub.t] – [([partial]F/[partial]L).sub.t] [L.sub.t]/[K.sub.t] (A.9)

From the adjustment-cost function it follows that:

[([partial]G/[partial]K).sub.t] = 1/2b[[[(I/K).sub.t] -c].sup.2] -b[[(I/K).sub.t] -c][(I/K).sub.t] = G([I.sub.t], [K.sub.t])/[K.sub.t] – b[(I/K).sup.2.sub.t] + bc[(I/K).sub.t] (A.10)

Substitution of Equations (A.9) and (A.10) into Equation (9b) and seeting:

[(Y/K).sub.t] = F([K.sub.t], [L.sub.t])/[K.sub.t] – G([I.sub.t], [K.sub.t])/[K.sub.t]

gives Equation (l0) in the text.

Annex. Major Events of Banking Market Liberalization in Various Countries

This Annex presents the major events in the sampled countries during the period 1988 to 1998 related to: 1) interest rates deregulation; 2) a removal of entry barriers; 3) a decrease in reserve requirements; 4) a removal of credit controls; 5) privatization of state banks; and 6) increased prudential bank regulation. Unless otherwise noted, the source of information is Williamson and Mahar (1998). EIU denotes country reports from the Economist Intelligence Unit, IMF denotes IMF Country Reports, and World Bank denotes World Bank Country Reports.

Argentina

1. Elimination of all interest rate controls in 1989 (EIU).

2. Removal of most entry barriers and branching restrictions in 1977 (Lindgren, Garcia, and Saal, 1996).

3. Reserve requirements lowered in 1993 (Galbis, 1993).

4. Credit controls were substantially reduced in 1993.

5. Start to privatize banks in 1995 (Lindgren et al., 1996).

6. Central Bank starts to enforce Basle capital adequacy standards in 1994 (Galbis, 1993).

Brazil

1. Deposit rates are fully liberalized in 1989.

2. Entry barriers are reduced after 1991.

3. Reserve requirements are rationalized after 1988.

4. Start to reduce directed credit especially to agricultural sector in 1994 (IME).

5. Begin of privatization of state-owned banks in 1997 (IME).

6. Central Bank modernizes its supervision practices in December 1997 (IME).

Chile

1. Controls on interest rates are eliminated in 1985 (Gallego and Loayza, 2001).

2. Banks are allowed to expand abroad and to enter new business areas at home since 1997 (EIU).

3. Reserve requirements on both demand and time deposits are reduced in 1980 (Bandiera et al., 2000).

4. Directed credit and credit ceilings are abandoned in 1976 (Bandiera et al., 2000).

5. Banks are re-privatized in 1986 (Bandiera et al., 2000).

6. Revision of banking law to strengthen the supervisory system in 1986 (Bandiera et al., 2000).

India

1. Most interest rates deregulated during the period 1995 to 1996, except those on deposits of less than one year and on small commercial bank loans (IMF).

2. Entry restrictions for banks eased in 1993.

3. After 1992, reserve requirements were reduced in stages (World Bank).

4. Priority credit scheme made more flexible for banks in 1994 (IMF).

5. No major reduction yet in government ownership of public banks (World Bank).

6. New prudential norms in line with Basle Accord become operational in 1996 (World Bank).

Indonesia

1. Most deposit and loan rates freed in 1983.

2. Monopoly of state banks over deposits of state enterprises removed in 1988. Entry of new banks is allowed (Bandiera et al., 2000). Activities of financial institutions broadened in 1988. Foreign banks allowed to establish joint ventures in 1988.

3. Reserve requirements drastically lowered in 1988 (Bandiera et al., 2000).

4. New reform package announced in 1990 which took on the directed credit program; Most of the liquidity credit arrangements for priority loans are eliminated in 1990 (World Bank).

5. Reduction of government ownership of state banks (World Bank).

6. Improved bank supervisory legislation in 1997, including new loan classification and loan loss provisioning rules (World Bank).

Malaysia

1. Interest rate controls completely eliminated in 1991.

2. A two-tier banking framework was introduced for commercial banks in December 1994 (IMF).

3. Reserve requirements were reduced in 1994 (World Bank).

4. The number of priority sectors and the required loan amount is reduced in 1991 (Bandiera et al., 2000).

5. There have been no privatizations of banks. Most large banks have been private since they started operations. Government, however, is majority shareholder in two largest banks (World Bank).

6. New regulation extends and strengthens Central Bank’s supervisory powers (Bandiera et al., 2000).

Mexico

1. Deposit rates liberalized in 1988-89. Loan rates liberalized after 1988, except at development banks.

2. New entry of banks permitted in 1991.

3. Reduction of reserve requirements in 1988-89 (IMF and Bandiera et al., 2000).

4. Abolition of directed lending to preferential sectors in 1989 (IMF). Elimination of the liquidity coefficient requiring that 30% of deposits be invested in T-bills in 1991 (Bandiera et al., 2000).f

5. Authorities nationalized 18 commercial banks in 1982. Nationalized banks re-privatized in 1991-1992.

6. The Central Bank became autonomous in April 1994 (EIU).

Pakistan

1. Most lending rates freed in 1995.

2. Eleven new private banks, including three foreign, established since 1991.

3. No significant reductions in reserve requirements (World Bank).

4. The credit-deposit ratio mechanism, which required banks to keep their credit to the private sector within limits related to their deposits base, was abolished in 1995 (IME).

5. Muslim Commercial Bank privatized in 1991. Allied bank privatized in stages between the period 1991 to 1993. First Women Bank privatized in 1997. Comprehensive reforms in 1997 reduced government interference in public-sector banks.

6. Steps were taking during the period 1993 to 1994 to increase the autonomy of the Central Bank (IMF); Coverage of bank supervision increased in 1994 (IMF).

Peru

1. Interest rate controls abolished in 1991.

2. In December 1996, entry requirements were eased (IMF).

3. Reserve requirements on domestic deposits reduced from 1991 onwards.

4. Subsidized lending eliminated in 1992.

5. All seven public commercial banks liquidated or divested over the period 1991 to 1995.

6. In 1993, the banking law was modified to strengthen prudential regulations that apply to banks (IMF).

Philippines

1. Interest rate controls mostly phased out over the period 1981 to 1985.

2. Restrictions on the entry and operation of banks were eased in 1994 (IMF); Restrictions on foreign bank branching were lifted in 1994 (IMF); Foreign banks were allowed to purchase up to 60% of the equity of local banks in 1994 (IMF).

3. Reserve requirements lowered in 1993.

4. Directed credit partly abolished in 1983.

5. Government reduced stake in PNB to 47% in December 1995.

6. In December 1993, Central Bank was restructured and re-capitalized (IMF).

Rep. of Korea

1. In 1993, deregulation of interest rates on deposits with maturities of two years and on most loans (IMF).

2. Entry barriers are lowered in again in 1989. The establishment of new financial institutions is approved in 1989 (Bandiera et al., 2000).

3. Reserve requirements lowered in 1996 (IMF)

4. Most policy-based lending phased out in 1996; In 1996, the Central Bank removed the restriction on the premium a bank could charge over its prime lending rate, and revised its rules for credit control (IMF).

5. Commercial banks were privatized during the period 1981 to 1983 (IMF).

6. General Banking Act of 1991 introduces new prudential measures and imposes supervisory regulations (Bandiera et al., 2000); In 1992, measures were introduced to increase transparency of regulations and procedures on bank supervision (IMF).

Taiwan

1. Interest rates nominally liberalized in 1989, but prices remained uncompetitive until new banks were established in 1992.

2. Deregulation on the entry of new private commercial banks during the period 1991 to 1992. Establishment of 16 new banks in 1992 (EIU).

3. Directed credit still prevalent. Budgets for subsidized credit continually modified in recent years.

4. No significant reductions in reserve requirements (World Bank).

5. In January 1998, three of the largest commercial banks are partly privatized (EIU).

6. In May 1997, the Central Bank of China (Taiwan) Act was amended to improve bank regulation (IMF).

Thailand

1. Interest-rate ceilings on all types of deposits abolished in 1990. Ceiling on loan rates removed in 1992.

2. Since September 1994, commercial banks are allowed to invest in any business (World Bank); Finance and securities companies are permitted to set up banks outside Bangkok with approval in 1995.

3. Reduction of reserve requirements in 1992 (IMF).

4. Government gradually eliminated directed credit after 1980. The ceiling on commercial bank loans was lifted in 1992. Commercial banks and finance companies were permitted to issues certificates of deposits in 1992. Relaxation of rural credit requirement in 1992 (IMF).

5. No privatization efforts. Most large Thai commercial banks are private, but one of the largest banks, Krung Thai bank, is still public (World Bank).

6. In 1997, banking law was amended to strengthen prudential regulations (IMF).

Table I

Financial Liberalization Dates

This table presents the years in which major progress was made with

respect to six aspects of financial liberlization during the period 1988

to 1998: 1) liberlization of interest rates; 2) removal of entry

barriers; 3) lowering of reserve requirements; 4) removal of credit

controls; 5) privatization of state-owned banks; and 6) introduction of

prudential regulation. “No” indicates that no significant progress has

been made during the last two decades (until 1998) in liberalizing this

particular measure of financial liberalization. My sources include

various IMF country reports and working papers, various World Bank

country reports and working papers, Economist Intelligence Unit Country

Reports, Bandiera, Caprio, Honohan, and Schiantarelli (2000),

Demirgtic-Kunt and Detragiache (1998), Galbis (1993), Gallego and Loayza

(2001), Lindgren, Garcia, and Saal (1996), and Williamson and Mahar

(1998). For more details regarding the sources of data, I refer to the

Annex to this paper.

Interest Entry Reserve Credit

Country Rates Barriers Reqrmnts Controls Privatization

Argentina 1989 1977 1993 1993 1995

Brazil 1989 1991 1988 1994 1997

Chile 1985 1997 1980 1976 1986

India 1996 1993 1993 1994 no

Indonesia 1983 1988 1988 1990 1992

Malaysia 1991 1994 1994 1991 no

Mexico 1989 1993 1989 1991 1992

Pakistan 1995 1993 no 1995 1997

Peru 1991 1996 1991 1992 1995

Philippines 1985 1994 1994 1983 1996

Rep. Korea 1993 1989 1996 1996 1983

Taiwan 1992 1992 no no 1998

Thailand 1992 1995 1992 1992 no

Prudential

Country Reg.

Argentina 1994

Brazil 1998

Chile 1986

India 1996

Indonesia 1997

Malaysia 1989

Mexico 1994

Pakistan 1994

Peru 1993

Philippines 1993

Rep. Korea 1992

Taiwan 1997

Thailand 1997

Notes: In Argentina, interest rates were deregulated in 1987, but

deregulation was reversed in 1988. In 1989, interest rates were

deregulated again. In India, Malaysia and Thailand there were no

significant reductions in state ownership of commercial banks during the

period 1988 to 1998. The largest Indian commercial banks are yet to be

privatized. The major Malaysian commercial banks have been private since

they started operations. Most Thai commercial banks are also privately

held, although one of the largest banks, Krung Thai Bank is still

controlled by the government. In Taiwan, credit controls were still in

place during 1998, and reserve requirements have remained high during

the period 1988 to 1998. Reserve requirements in Pakistan have also

remained high during the period 1988 to 1998.

Table II

Financial Liberalization Index

This table presents a financial liberalization index by year and

country. For each country, the index in a particular year is defined as

the sum of the financial liberalization measures in Table I that have

been implemented in the country up to that year. Thus, the index denotes

changes in the degree of financial liberalization within a country, and

ranges from 0 to 6, with 6 indicating the highest level of financial

liberalization. The index is constructed as of year-end status. In my

empirical work, I use figures at year-end of period t-1 for the level of

financial liberalization at the beginning of period t.

Country 1998 1989 1990 1991 1992 1993 1994 1995 1996 1997

Argentina 1 2 2 2 2 4 5 6 6 6

Brazil 1 2 2 3 3 3 4 4 4 5

Chile 5 5 5 5 5 5 5 5 5 6

India 0 0 0 0 0 2 3 3 5 5

Indonesia 3 3 4 4 5 5 5 5 5 6

Malaysia 1 2 2 4 4 4 6 6 6 6

Mexico 0 2 2 3 4 5 6 6 6 6

Pakistan 0 0 0 0 0 1 2 4 4 5

Peru 0 0 0 2 3 4 4 5 6 6

Philippines 2 2 2 2 2 3 5 5 6 6

Rep. Korea 1 2 2 2 3 4 4 4 6 6

Taiwan 0 0 0 0 2 2 2 2 2 3

Thailand 1 1 1 1 4 4 4 5 5 6

Country 1998

Argentina 6

Brazil 6

Chile 6

India 5

Indonesia 6

Malaysia 6

Mexico 6

Pakistan 5

Peru 6

Philippines 6

Rep. Korea 6

Taiwan 4

Thailand 6

Table III

Sequence of Financial Liberalization

The numbers in this table indicate the sequence of financial

liberalization with respect to each of the six different measures in

Table I. The number 1 indicates that a particular type of liberalization

was the first measure (out of a total number of six measures) to come

into effect in the country. The number 2 refers to the second measure to

come into effect, etc. Averages are used when measures are implemented

in the same year

Interest Entry Reserve Credit

Country Rates Barriers Requirements Controls Privatization

Argentina 2 1 3.5 3.5 6

Brazil 2 3 1 4 5

Chile 3 6 2 1 4.5

India 4.5 1.5 1.5 3 –

Indonesia 1 2.5 2.5 4 5

Malaysia 3.5 5.5 5.5 3.5 1

Mexico 1.5 5 1.5 3 4

Pakistan 3.5 1 – 3.5 5

Peru 1.5 6 1.5 3 5

Philippines 2 4.5 4.5 1 6

Rep. Korea 4 2 5.5 5.5 1

Taiwan 1.5 1.5 – – 4

Thailand 3 5 3 3 1

Average 2.5 3.4 2.9 3.2 4.0

Median 2 3 2.5 3.3 4.8

Prudential

Country Regulation

Argentina 5

Brazil 6

Chile 4.5

India 4.5

Indonesia 6

Malaysia 2

Mexico 6

Pakistan 2

Peru 4

Philippines 3

Rep. Korea 3

Taiwan 3

Thailand 6

Average 4.2

Median 4.5

Table IV

Comparison of Financial Liberalization Dates

This table compares the dates of substantial banking market

liberalization with the dates of stock market liberalization. FLI5

indicates the year when the banking sector became largely liberalized

and is defined as the year when the financial liberalization index (FLI)

in Table I reaches value of 5. LLI is an alternative measure that

indicates the year when the banking system became largely liberalized

and is taken from Williamson and Mahar (1998). The stock market

liberalization dates are from Bekaert, Harvey, and Lundblad (2001).

Country Largely Liberalized Stock Market

Banking Sector Liberalization

FLI5 LLI

Argentina 1994 1993 Nov 1989

Brazil 1997 Not until 1996 May 1991

Chile 1986 1985 Dec 1988

India 1996 Not until 1996 Oct 1992

Indonesia 1992 1989 Sep 1989

Malaysia 1994 1992 Dec 1988

Mexico 1993 1992 May 1989

Pakistan 1997 Not until 1996 Feb 1991

Peru 1995 1993 Jun 1993

Philippines 1994 1994 Oct 1989

Rep. Korea 1996 Not until 1996 Feb 1992

Taiwan Not until 1998 Not until 1996 Jan 1991

Thailand 1995 1992 Dec 1988

Table V

Financial Liberalization and the Political Climate

This table presents the Political Risk Index of the International

Country Risk Guide (ICRG) from the Political Risk Services Group for the

years 1988 and 1998 (corresponding to the first and last year of my

sample period, respectively), and the correlation between the ICRG

political risk index and my financial liberalization index (FLI) over

the sample period 1988 to 1998. The ICRG Political Risk Index is a broad

measure of the level of political risk in a country. The index ranges

from 0 to 100, and is decreasing in the level of political risk. I do

not have ICRG data for Pakistan and Peru.

Country ICRG Political Risk Index FLI and ICRG

Political Risk Index

Correlation During

1988 1998 1988 to 1998

Argentina 57 76 0.89

Brazil 67 66 0.29

Chile 56 73 0.49

India 45 59 0.85

Indonesia 39 42 0.45

Malaysia 58 66 0.77

Mexico 66 69 0.16

Philippines 40 74 0.90

Rep. Korea 64 75 0.75

Taiwan 77 82 0.81

Thailand 58 70 0.86

Average 57 68 0.66

Table VI

Descriptive Statistics

Table VIa to VId present descriptive statistics of the key variables of

the empirical part of the paper. Table VIa presents the descriptive

statistics for the whole sample of firms. Table VIb presents the median

values of the key variables by country. Table VIc presents the median

values of the key variables by industry. Table VId presents the median

values of the key variables by year. The industry codes in Table VIb

follow the classification of the 1987 US SIC. Industry codes are at

two-digit SIC levels for the manufacturing industries (SIC codes 20-39)

and at one-digit SIC levels for other industries. I exclude the

following sectors: finance, insurance and real estate; services;

government; and others. In other words, I exclude SIC codes 6 to 9. I/K

denotes the investment-to-capital ratio, Y/K denotes the

sales-to-capital ratio and CF/K denotes the cash flow-to-capital ratio.

The total number of firm-year observations is 1,645. The sample period

is 1988 to 1998, and the source of the firm-level data is Worldscope.

Panel A

Descriptive Statistics for All Firms in the Sample

I/K Y/K CF/K

Mean 0.189 2.002 0.280

Median 0.170 1.534 0.247

Maximum 0.500 9.859 0.967

Minimum 0.010 0.147 0.012

Standard deviation 0.125 1.621 0.155

Observations 1,645 1,645 1,645

Panel B

Median Statistics Categorized by Country

Country I/K Y/K CF/K Observations

Argentina 0.157 0.716 0.203 51

Brazil 0.140 0.570 0.167 30

Chile 0.161 1.075 0.252 149

India 0.163 2.004 0.288 300

Indonesia 0.176 2.385 0.310 55

Malaysia 0.147 1.369 0.242 335

Mexico 0.194 1.378 0.217 99

Pakistan 0.154 2.995 0.379 17

Peru 0.148 1.569 0.288 10

Philippines 0.250 0.939 0.219 32

Rep. Korea 0.194 1.876 0.220 249

Taiwan 0.121 1.392 0.211 114

Thailand 0.177 1.728 0.302 204

All 0.170 1.534 0.247 1,645

Panel C

Median Statistics Categorized by Industry

SIC

Industry Code Industry Name I/K Y/K CF/K

1 0 Agriculture, forestry, and fishing 0.144 0.974 0.288

2 1 Mining 0.146 0.871 0.203

3 2 Construction 0.208 0.898 0.201

4 4 Transportation, communication, 0.175 0.616 0.211

electric, gas and sanitary services

5 5 Wholesale trade and retail trade 0.218 2.137 0.239

6 20 Food and kindered products 0.153 2.038 0.254

7 21 Tobacco manufactures 0.190 2.042 0.500

8 22 Textile mill products 0.154 1.414 0.221

9 23 Apparel and fabrics-based products 0.192 3.049 0.363

10 24 Lumber and wood products 0.108 0.848 0.189

11 25 Furniture and fixtures 0.135 2.690 0.419

12 26 Paper and allied products 0.158 1.387 0.242

13 27 Printing and publishing 0.163 1.589 0.368

14 28 Chemicals and allied products 0.157 1.541 0.279

15 29 Petroleum refining 0.187 3.141 0.275

16 30 Rubber and plastics products 0.114 2.026 0.218

17 31 Leather and leather products 0.120 4.724 0.301

18 32 Stone, clay, glass, and concrete 0.171 1.017 0.223

19 33 Primary metal industries 0.123 1.440 0.209

20 34 Fabricated metal products 0.168 1.800 0.302

21 35 Machinery, except electrical 0.277 2.250 0.278

22 36 Electrical and electronic 0.217 2.232 0.376

machinery, equipment and supplies

23 37 Transportation equipment 0.181 2.496 0.268

24 38 Instruments; photographic, medical 0.276 1.595 0.251

and optical goods; clocks

25 39 Miscellaneous manufacturing 0.170 1.681 0.302

All industries 0.170 1.534 0.247

SIC

Industry Code Industry Name Observations

1 0 Agriculture, forestry, and fishing 26

2 1 Mining 54

3 2 Construction 26

4 4 Transportation, communication, 161

electric, gas and sanitary services

5 5 Wholesale trade and retail trade 148

6 20 Food and kindered products 185

7 21 Tobacco manufactures 12

8 22 Textile mill products 70

9 23 Apparel and fabrics-based products 44

10 24 Lumber and wood products 36

11 25 Furniture and fixtures 8

12 26 Paper and allied products 63

13 27 Printing and publishing 20

14 28 Chemicals and allied products 201

15 29 Petroleum refining 45

16 30 Rubber and plastics products 38

17 31 Leather and leather products 9

18 32 Stone, clay, glass, and concrete 139

19 33 Primary metal industries 106

20 34 Fabricated metal products 36

21 35 Machinery, except electrical 30

22 36 Electrical and electronic 72

machinery, equipment and supplies

23 37 Transportation equipment 85

24 38 Instruments; photographic, medical 10

and optical goods; clocks

25 39 Miscellaneous manufacturing 21

All industries 1,645

Panel D

Median Statistics Categorized by Year

Year I/K Y/K CF/K FLI Observations

1989 0.237 2.028 0.168 1 13

1990 0.278 1.761 0.174 2 36

1991 0.197 1.802 0.163 2 60

1992 0.173 1.821 0.234 3 100

1993 0.181 1.532 0.221 4 165

1994 0.160 1.503 0.264 4 212

1995 0.186 1.538 0.275 4 286

1996 0.169 1.565 0.314 4 321

1997 0.169 1.460 0.313 5 275

1998 0.117 1.434 0.337 6 177

All 0.170 1.534 0.285 4 1,645

Table VII

Regression Results Based on Estimation of the Investment Model

This table presents the main regression results based on the estimation

of several specifications of the structural investment model presented

in Equation (12). The dependent variable in all regressions is the

investment-to-capital ratio of a specific firm at time t, [(I/K).sub.t]

Y/K denotes the sales-to-capital ratio and CF/K denotes the cash

flow-to-capital ratio. In Regressions (1) to (3), Small, is a dummy

variable that takes value one if sales of the firm is smaller than the

median sales in the sample at time t, and zero otherwise. In Regression

(4), Small, is a dummy variable that takes value one if sales of the

firm is smaller than the 3/4 quantile of sales in the sample at time t,

and zero otherwise. [FLI.sub.t] is the financial liberalization index at

time t. All models are estimated in first differences using instrumental

variables. I present two-step GMM estimates of all regressions as

developed by Arellano and Bond (1991). I estimate all model

specifications in first-differences. All regressions use variables at

t-2, t-3 as instruments and assume that [Small.sub.t], and [FLI.sub.t]

are exogenous variables. All regressions include year dummies, but these

are not reported. Heteroskedasticity consistent standard errors are

between parantheses

(1) (2) (3)

Variable Basic FLI Small

Constant -0.0078 ** -0.0158 *** -0.0185 ***

(0.0031) (0.0042) (0.0040)

[(I/K).sub.t-1] 0.5660 *** 0.5316 *** 0.5385 ***(0.1382) (0.1417) (0.1428)

[[(I/K).sup.2].sub.t-1] -0.9431 *** -0.8219 *** -0.8038 ***(0.2771) (0.2808) (0.2832)

[(Y/K).sub.t] 0.0871 *** 0.0635 *** 0.0703 *(0.0266) (0.0235) (0.0397)

[(CF/K).sub.t] -0.0217 -0.0484 -0.3613(0.1288) (0.1670) (0.2827)

[FLI.sub.t] * [(Y/K).sub.t] 0.0027 -0.0038(0.0036) (0.0061)

[FLI.sub.t] * [(CF/K).sub.t] 0.0343 0.1253 **(0.0314) (0.0538)

[Small.sub.t] * [(Y/K).sub.t] -0.0198(0.0416)

[Small.sub.t] * [(CF/K).sub.t] 0.4771(0.3277)

[Small.sub.t] * [FLI.sub.t] * 0.0140 *[(Y/K).sub.t][Small.sub.t] * [FLI.sub.t] * (0.0080)

[(Y/K).sub.t][Small.sub.t] * [FLI.sub.t] * -0.1485 **

[(CF/K).sub.t](0.0654)

Specification Tests(p-values):

a. First-Order Serial 0.0000 *** 0.000 *** 0.000 ***

Correlation:

b. Second-Order Serial 0.209 0.246 0.303

Correlation:

c. Sargan Test; 0.200 0.295 0.452

Instruments: t-2, t-3 t-2, t-3 t-2, t-3

Number of Observations: 857 857 857

Number of Firms: 394 394 394

(4)

Small 75th

Variable Percentile

Constant -0.0195 ***

(0.0043)

[(I/K).sub.t-1] 0.5543 ***(0.1465)

[[(I/K).sup.2].sub.t-1] -0.7832 ***(0.2941)

[(Y/K).sub.t] 0.0282(0.0592)

[(CF/K).sub.t] -0.5615(0.4075)

[FLI.sub.t] * [(Y/K).sub.t] -0.0034(0.0084)

[FLI.sub.t] * [(CF/K).sub.t] 0.2440 ***(0.0774)

[Small.sub.t] * [(Y/K).sub.t] 0.0520(0.0597)

[Small.sub.t] * [(CF/K).sub.t] 0.5913(0.4305)

[Small.sub.t] * [FLI.sub.t] * 0.0074[(Y/K).sub.t][Small.sub.t] * [FLI.sub.t] * (0.0095)

[(Y/K).sub.t][Small.sub.t] * [FLI.sub.t] * -0.2465 ***

[(CF/K).sub.t](0.0816)

Specification Tests(p-values):

a. First-Order Serial 0.000 ***

Correlation:

b. Second-Order Serial 0.310

Correlation:

c. Sargan Test; 0.493

Instruments: t-2, t-3

Number of Observations: 857

Number of Firms: 394

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

(1.) There exists an extensive theoretical literature on potential causes of financial frictions. Information asymmetry can create adverse selection (Akerlof, 1970), and may cause credit rationing in the loan market (Stiglitz and Weiss, 1981). Myers and Majluf (1984) discuss the information asymmetry problems of equity markets. Information asymmetry can also give rise to principal-agent problems, such as incentive problems and costly monitoring. Jensen and Meckling (1976) show that in the presence of limited-liability debt agency costs can arise between equity holders and debt holders as equity holders may have the incentive to opt for excessively risky investment projects that are value destroying. Townsend (1979) discusses how debt markets are affected by costly state verification. Hart (1995) gives an overview of how incomplete contracting can affect firms’ financial structure. Modigliani and Perotti (1997) and La Porta, Lpez-De-Silanes, Shleifer, and Vishny. (1997) discuss how contract enforcement affects the choice between internal or external finance.

(2.) I also created a larger set of firms by applying less strict outlier rules. This set includes firms from Colombia, Sri Lanka, Turkey, and Venezuela. My empirical results for this larger set of firms are similar to the results I present here (not reported).

References

Abel, A., 1980, “Empirical Investment Equations: An Integrative Framework,” in K. Brunner and A. Meltzer, Eds., On the State of Macroeconomics, Carnegie-Rochester Conference Series 12,39-93.

Abel, A. and O. Blanchard, 1986, “The Present Value of Profits and Cyclical Movements in Investment,” Econometrica 54, 249-274.

Akerlof, G.A., 1970, “The Market for Lemons: Quality Uncertainty and the Market Mechanism,” Quarterly Journal of Economics 84, 488-500.

Arellano, M. and S.R. Bond, 1988, “Dynamic Panel Data Estimation Using DPD–A Guide For Users,” Institute for Fiscal Studies Working Paper 88/15, London.

Arellano, M. and S.R. Bond, 1991, “Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations,” Review of Economic Studies 58, 277-297.

Arellano, M. and O. Bover, 1995, “Another Look at the Instrumental Variable Estimation of Error-Components Models,” Journal of Econometrics 68, 29-51.

Bandiera, O., G. Caprio, P. Honohan, and F. Schiantarelli, 2000, “Does Financial Reform Raise or Reduce Savings?” Review of Economics and Statistics 82, 239-263.

Beck, T., R. Levine, and N. Loayza, 2000, “Finance and the Sources of Growth,” Journal of Financial Economics 58, 261-300.

Bekaert, G. and C.R. Harvey, 2000, “Foreign Speculators and Emerging Equity Markets,” Journal of Finance 55, 565-613.

Bekaert, G., C.R. Harvey, and C. Lundblad, 2001, “Does Financial Liberalization Spur Growth?” NBER Working Paper 8245.

Blundell, R. and S.R. Bond, 1998, “Initial Conditions and Moment Restrictions in Dynamic Panel Data Models,” Journal of Econometrics 87, 115-143.

Blundell, R., S.R. Bond, M. Devereux, and F. Schiantarelli, 1992, “Investment and Tobin’s Q: Evidence from Company Panel Data”, Journal of Econometrics 51, 233-257.

Blundell, R., S.R. Bond, and C. Meghir, 1996, “Econometric Models of Company Investment,” in L. Matyas and P. Sevestre, Eds., The Econometrics of Panel Data: A Handbook of the Theory with Applications, Advanced Studies in Theoretical and Applied Econometrics, Vol. 33, Kluwer Academic Publishers.

Bond, S.R. and C. Meghir, 1994, “Dynamic Investment models and the Firm’s Financial Policy,” Review of Economic Studies 61, 197-222.

Bond, S.R., J. Elston, J. Mairesse, and B. Mulkay, 1997, “Financial Factors and Investment in Belgium, France, Germany, and the UK: A Comparison Using Company Panel Data,” NBER Working Paper 5900.

Bond, S. and J. Van Reenen, 1999, “Microeconomic Models of Investment and Employment,” in J.J. Heckman and E.E. Leamer, Eds., Handbook of Econometrics, Vol. 6, Elsevier Science, North-Holland, Amsterdam (Forthcoming).

Cho, Y.J., 1988, “The Effect of Financial Liberalization on the Efficiency of Credit Allocation: Some Evidence from Korea,” Journal of Development Economics 29, 101-110.

Demirguc-Kunt, A. and Maksimovic, 1998, “Law, Finance, and Firm Growth,” Journal of Finance 53, 2107-2137.

Demirguc-Kunt, A. and E. Detragiache, 1998, “Financial Liberalization and Financial Fragility,” World Bank Policy Research Working Paper 1917.

Devereux, M.B. and F. Schiantarelli, 1990, “Investment, Financial Factors, and Cash Flow from UK Panel Data,” in G. Hubbard, Ed., Information, Capital Markets, and Investment, Chicago, IL, University of Chicago Press.

Devereux, M.B. and G.W. Smith, 1994, “International Risk Sharing and Economic Growth,” International Economic Review 35, 535-551.

Fazzari, S., G. Hubbard, and B. Petersen, 1988, “Financing Constraints and Corporate Investment,” Brookings Papers on Economic Activity 1, 141-195.

Fazzari, S., G. Hubbard, and B. Petersen, 2000, “Investment-Cash Flow Sensitivities are Useful: A Comment on Kaplan and Zingales,” Quarterly Journal of Economics 115, 695-705.

Galbis, V., 1993, “High Real Interest Rates under Financial Liberalization: Is There a Problem?” IMF Working Paper 93/7.

Galindo, A., F. Schiantarelli, and A. Weiss, 2002, “Does Financial Liberalization Improve the Allocation of Investment?” Boston College Working Paper 503.

Gallego, F. and N. Loayza, 2001, “Financial Structure in Chile: Macroeconomic Developments and Microeconomic Effects,” in A. Demirguc-Kunt and R. Levine, Eds., Financial Structure and Economic Growth, Cambridge, MA, MIT Press, 299-346.

Gelos, G. and A. Werner, 2002, “Financial Liberalization, Credit Constraints, and Collateral: Investment in the Mexican Manufacturing Sector,” Journal of Development Economics 67, 1-27.

Gertler, M. and A. Rose, 1994, “Finance, Public Policy, and Growth,” in G. Caprio, I.

Atiyas, and J. Hanson, Eds., Financial Reform: Theory and Experience, New York, NY, Cambridge University Press.

Gilchrist, S. and C. Himmelberg, 1995, “Evidence on the Role of Cash Flow for Investment,” Journal of Monetary Economics 36, 541-572.

Gilchrist, S. and C. Himmelberg, 1998, “Investment, Fundamentals and Finance,” in Ben S. Bernanke and Julio Rotemberg, Eds., NBER Macroeconomics Annual, Cambridge, MA, MIT Press.

Gomes, J.F., 2001, “Financing Investment,” American Economic Review 91, 1263-1285.

Hansen, L., 1982, “Large Sample Properties of Generalized Method of Moments Estimators,” Econometrica 50, 1029-1054.

Harris, J.R., F. Schiantarelli, and M.G. Siregar, 1994, “The Effect of Financial Liberalization on the Capital Structure and Investment Decisions of Indonesian Manufacturing Establishments,” World Rank Economic Review 8, 17-47.

Hart, O., 1995, Firms, Contracts, and Financial Structure, Clarendon Lectures in Economics, Oxford, UK, Oxford University Press.

Hayashi, F., 1982, “Tobin’s Marginal q and Average q: A Neoclassical Interpretation,” Econometrica 50, 213-224.

Hayashi, F. and T. Inoue, 1991, “The Relationship Between Firm Growth and Q with Multiple Capital Goods: Theory and Evidence from Panel Data on Japanese Firms,” Econometrica 59, 731-753.

Henry, P.B., 2000a, “Do Stock Market Liberalizations Cause Investment Booms?” Journal of Financial Economics 58, 301-334.

Henry, P.B., 2000b, “Stock Market Liberalization, Economic Reform, and Emerging Market Equity Prices,” Journal of Finance 55, 529-564.

Holtz-Eakin, D., W. Newey, and H.S. Rosen, 1988, “Estimating Vector Autoregressions With Panel Data,” Econometrica 56, 1371-1396.

Hoshi, T., A. Kashyap, and D. Scharfstein, 1991, “Corporate Structure, Liquidity and Investment: Evidence from Japanese Industrial Groups,” Quarterly Journal of Economics 106, 33-60.

Hubbard, G., 1998, “Capital Market Imperfections and Investment,” Journal of Economic Literature 36, 193-225.

Hubbard, G. and A. Kashyap, 1992, “Internal Net Worth and the Investment Process: An Application to US Agriculture,” Journal of Political Economy 100, 506-534.

Hubbard, G., A. Kashyap, and T. Whited, 1995, “Internal Finance and Firm Investment,” Journal of Money, Credit, and Banking 27, 683-701.

Jaramillo, F., F. Schiantarelli, and A. Weiss, 1996, “Capital Market Imperfections Before and After Financial Liberalization: An Euler Equation Approach to Panel Data for Ecuadorian Firms,” Journal of Development Economics 51, 367-386.

Jensen, M. and W. Meckling, 1976, “Theory of the Firm: Managerial Behavior, Agency Costs, and Ownership Structure,” Journal of Financial Economics 1, 305-360.

Kaplan, S. and L. Zingales, 1997, “Do Investment-Cash Flow Sensitivities Provide Useful Measures of Financing Constraints?” Quarterly Journal of Economics 112, 169-215.

King, R.G. and R. Levine, 1993, “Finance and Growth: Schumpeter Might Be Right,” Quarterly Journal of Economics 108, 717-737.

La Porta, R., F. Lopez-De-Silanes, A. Shleifer, and R.W. Vishny, 1997, “Legal Determinants of External Finance,” Journal of Finance 52, 1131-1150.

Laeven, L. and E.C. Perotti, 2001, “Confidence Building in Emerging Stock Markets,” CEPR Discussion Paper 3055, November.

Lamont, O., 1997, “Cash Flow and Investment: Evidence from Internal Capital Markets,” Journal of Finance 52, 83-109.

Lindgren, C.J., G. Garcia, and M. Saal, 1996, Bank Soundness and Macroeconomic Policy, Washington, DC, International Monetary Fund.

Love, I., 2003, “Financial Development and Financing Constraints: International Evidence from the Structural Investment Model,” Review of Financial Studies (Forthcoming).

Mairesse, J., B. Hall, and B. Mulkay, 1999, “Firm-level Investment in France and the United States: An Exploration of What I Have Learned in Twenty Years,” NBER Working Paper 7437.

McKinnon, R.I., 1973, Money and Capital in Economic Development, Washington, DC, Brookings Institution.

Modigliani, F. and M.H. Miller, 1958, “The Cost of Capital, Corporate Finance, and the Theory of Investment,” American Economic Review 48, 261-297.

Modigliani, F. and E.C. Perotti, 1997, “Protection of Minority Interest and the Development of Security Markets,” Managerial and Decision Economics 18, 519-528.

Myers, S.C. and N.S. Majluf, 1984, “Corporate Financing and Investment Decisions When Firms Have Information That Investors Do Not Have,” Journal of Financial Economics 13, 187-221.

Obstfeld, M., 1994, “Risk Taking, Global Diversification and Growth,” American Economic Review 84, 1310-1329.

Oliner, S.D. and G.D. Rudebusch, 1992, “Sources of the Financing Hierarchy for Business Investment,” Review of Economics and Statistics 74, 643-654.

Rajan, R. and L. Zingales, 1998, “Financial Dependence and Growth,” American Economic Review 88, 559-586.

Schiantarelli, F., 1995, “Financial Constraints and Investment: A Critical Review of Methodological Issues and International Evidence,” in J. Peek and E. Rosengren, Eds., Is Bank Lending Important for the Transmission of Monetary Policy?, Boston, MA, Federal Reserve Bank of Boston.

Schiantarelli, F., I. Atiyas, G. Caprio, J. Harris, and A. Weiss, 1994, “Credit Where It is Due? A Review of the Macro and Micro Evidence on the Real Effects of Financial Reform,” in G. Caprio, I. Atiyas, and J. Hanson, Eds., Financial Reform: Theory and Experience, New York, NY, Cambridge University Press.

Shaw, E.S., 1973, Financial Deepening in Economic Development, New York, NY, Oxford University Press.

Shin, H.H. and R.M. Stulz, 1998, “Are International Capital Markets Efficient?” Quarterly Journal of Economics 113, 531-552.

Stiglitz, J.E. and A. Weiss, 1981, “Credit Rationing in Markets with Imperfect Information,” American Economic Review 71, 393-410.

Summers, L., 1981, “Taxation and Corporate Investment: A q-Theory Approach,” Brookings Papers on Economic Activity 1, 67-127.

Tobin, J., 1969, “A General Equilibrium Approach to Monetary Theory,” Journal of Money, Credit, and Banking 1, 15-29.

Townsend, R., 1979, “Optimal Contracts and Competitive Markets with Costly State Verification,” Journal of Economic Theory 21, 265-293.

Van Wijnbergen, S.J.G., 1983, “Credit Policy, Inflation, and Growth in a Financially Repressed Economy,” Journal of Development Economics 13, 45-65.

Whited, T., 1992, “Debt, Liquidity Constraints, and Corporate Investment: Evidence from Panel Data,” Journal of Finance 47, 1425-1460.

Williamson, J. and M. Mahar, 1998, “A Survey of Financial Liberalization,” Essays in International Finance No. 211, Department of Economics, Princeton University, Princeton, NJ.

Wurgler, J., 2000, “Financial Markets and the Allocation of Capital,” Journal of Financial Economics 58, 187-214.

I am grateful to Lemma Senbet and Alex Thiantis (the Editors) and an anonymous referee for suggestions that helped improve the paper considerably. 1 would also like to thank Thorsten Beck, Geert Bekaert, Jerry Capria, Stijn Claessens, Gaston Gelos, Jtm Hanson, Inessa Love, Pieter van Oijen, Fabto Schiantarelli, Sweder van Wijnbergern, and seminar parttcipants at the I 17th Annual Congress of the European Economic Assoctatton tn Venice for valuable comments, and Ying Lin for help with the data. The views expressed in this paper are those of the author and should not be interpreted to reflect those of the World Bank or its affiliated institutions.

Laeven Luc *

* Luc Laeven is with the World Bank.

COPYRIGHT 2003 Financial Management Association

COPYRIGHT 2003 Gale Group