Policy, Research, and External Atlalrs WORKING PARERS Financial Policy and Systems Country Economics Department The World Bank April 1991 WPS 654 Reflections on Credit Policy in Developing Countries Its Effect on Private Investment Mansoor Dailami and Marcelo Giugale The joint effect of both the volume of credit and its price - that is, the interest rate- is relevant to firms' investment decisions. So effective credit policy in developing countries must take into account the influence of both the credit supply and the interest rate, not just one or the other. TbePoicy, Researc and External Affaua Complex dis ibutesPRI3 WorkingPapers todisseminatetheftndings of wok in progress and to encoa age the exchange of ideas among Bank staff and al ethcrs intersted in developnent issues. These papers eaTry the namnes of the authors, teflect anly their views, and should be used and cited accordingly. The findings, intepreations, and conclusions are the autshors own. They should not be auributed to the World Bank, its Board of Directa, its managenent, or any of itsmeniber cuties. Policy, Research, and External Affairs Fnanlal Policy and Systems WPS 654 Thispaper-ajolntproductoftheFinancialPolicy and Systems Division, Country Economics Department and the Country Operations Division, Country Department III, Europa, Middle East, and North Africa Regional Office - is part of a larger effort in PRE to analyze the role of financial policy in the growth and adjustment process of developing countries. Copies are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Maria Raggambi, room N9-04 1, etension 37657 (28 pages, with charts, figures, and tables). Previous approaches to credit policy and its role sentiments - that banks consider in making in the stabilization and adjustment of developing lending decisions. These factors affect banks' countries have emphasized either the role of the decisions independent of a borrower's creditwor- availability of credit or the role of its price - thiness. Thus, in times of tight money, firms that that is, the Interest rate. Dailami and Giugale would otherwise have received loans may be argue that effective credit policy in developing denied them and have to postpone or cut back countries must take into account both: terest investment plans. rate and credit channels. Dailami and Giugale use their model to The authors develop their argument in the specify an equation relating aggregate private context of the link between credit policy and investment to aggregate output and to two credit private investment, using a model of firms' market variables - the real interest rate and investment behavior in an economy with aggregate credit. They estimate the equation for exogenous, time-varying borrowing constraints. five developing countries (Brazil, Colombia, The model incorporates a credit ceiling linked to India, Korea, and Turkey) using annual data for the fims' net worth and the state of the credit 1965-85, and test the joint significance of both market. interest rate and credit supply conditions. Their findings show that interest rates and credit The state of the credit market depends on volume exert a joint influence on the behavior of factors - such as credit and interest rate policy, private investment in the countries examined regulatoty and supervisory practices, and market The PRE Working Paper Series disseminates the findings of work under way in the Bank's Policy, Research, and External Affairs Complex. Anobjectiveoftheseries is to getthese fmdingsoutquickly, even if presentations are less than fully polished. The findings. interpretations, and conclusions in these papers do not necessarily represent official Bank policy. Produced by the PRE Dissemination Center Teble of joutents I* Introduction I *0* *******,.. . .. ...................... .......... 1 II. Theoretical Framework ...... ...... ............... *............. ... 4 III. Empirical Findings ...................... .. .........12 IV* Conclusion ......... ....... ... .... ...20 References 2............................................................ 27 List of Tables Table 1 .......... .... .................................................. 18 Table 2 .. . . ... .. ....... ..... .*... 19 List of Charts Chart 1 ................ ................................................ 22 Chart 2 ......................... .................23 C-hart 3 o.o...... .... ................... ........ ........... 24 Chart 4 * .....o .... .........*........... .......... ........o..o.. 25 Chart 5 *....o...... o .*ooo....o. .*o*o o o*.......*...... ............ * o..... 26 2 I. INTRODUCTION Successive developments in the role of credit policy in the stabilization and adjustment strategy of developing countries have been toward greater recognition of the importance of interest rate channel in influencing the pace of financial sector reform,1 in the process of allocative efficiency,3 and in influencing aggregate private spending.' Indeed, the earlier emphasis placed on volume of credit as an instrument of policy, both in development literature and practice, as exemplified most notably in the IMP approach to stabilization [see for instance, Polak and Argy (1971); Keller (1980)1, is now giving way to a new conventional wisdom about the virtues of positive real interest rate policy.' But ironically, as this view is gaining the status of a new orthodoxy there exists a growing body of theoretical research and empirical evidence which proport to return the quantitative channel of credit policy to the forestage. Thus, there exists the important strand of literature on credit rationing which either in its micro context (Hodgman (1960, 1962); Jaffee and Russell (1976); and Stiglitz and Weiss (1981)1 or in its traditional "availability doctrine," context [Roosa (1951); Scott (1957); Lindbeck (1962)] has led to rationalize non-price credit rationing even in environments where interest rates are free of administrative control and regulation.$ Moreover, by shifting the spotlight from the saver/borrower decisions to that of the intermediary, i.e. the 1 On this see, McKinnon (1973); Khatkate (1972); Galbis (1977); and World Bank (1909). 2 See, for instance, Balassa (1982); and Roe (1982). ' See, Taylor (1979, 1981); van Wijnbergen (1983) and Dailami (1989). See Sauv6 (1986) for current state of debate on interest rate policy in developing countries. ' See also, Baltensperger and Devinney (1985) for a recent survey and synthesis of credit rationing theory. 3 lender or the ' Ink, in the debate about the link between credit/monetary policy and the real economy, this literature has provided a rational basis for consideration of the joint influences of interest rate and credit channels. The underlying argument being that an increase in interest rate, for instance, could also be accompanied bv a reduction in the level of bank lending as a result of banks' efforts to adjust their asset portfolio, thereby reinforcing the contractionary influence of higher interest rates. The net outcome could be a more severe slowdown in economic activity then anticipated. Such a possibility is particularly likely in a situation where the economy is under stabilization measures. Following this line of reasoning, this paper argues that effective credit policy in the context of developing countries must acknowledge the influence of both interest rate and oredit channels. The argument is elaborated in the context of the link between credit policy and private investment and is conducted in two stages. In the next section, we draw on the recent literature on credit rationing to introduce a form of bank credit constraint, linked to the firms' net-worth and to the state of credit market, into the standard neo-classical model of firme' investment behavior. While the relevance of borrowers' net-worth, i.e., cred..t worthiness to banks' lending decisions and practices is familiar and has been well documented, both theoretically and empirically,' the introduction of the state of credit market, as a component of the supply of bank loan needs some elaboration. We refer to the state of credit market as a shorthand for a composite of several factors including the level of interest rate, the pace of credit policy, the stringency to which regulation and supervision on the banking system is imposed. These factors are expected to influence banks' lending decisions independently of their impacts on the creditworthiness, i.e., net-worth of individual borrowers. Thus, including the state of credit market in the banks' loan supply function signifies the possibility that credit to a particular See, for instance, Leland and Pyle (1977); Bernake and Gertler (1989, 1990); and Calomiris (1990). 4 firm may chang.e as the result of macro and financial fluctuations, even i# the borrower's creditworthiness has not changed. Thus, in times of tight money, or in "bad times," borrowers that otherwise would have been considered eligible are likely to face a more stringent credit evaluation process or simply to be denied loans. As a consequence, firms' investment plans may have to be postponed or cut back. In Section III we estimate and test for the relevance of credit market condition variables over a set of developing countries over the last two decades; specifically, we have compiled annual time series data on private fixed domestic investment and other relevant variables for Brazil, Colombia, India, Korea and Turkey over the 1965-85 period.' Linear three-stage least squares are used in the estimation to correct both for the endogeneity of aggregate and expected variables (via instrumental variables) and for the shortness of the sample period. Our findings not only show that credit market condition variables as a whole are key determinants of private domestic investment behavior but also (and perhaps more interestingly) that it is their expected future values that really matters. Moreover, the expected volume of aggregate credit appears to be the single most important variable in explaining investment behavior. This implies that credit constraints have actually been binding in the countries in the sample and, presumably, profitable investment opportunities have been lost. In the paper's conclusions, we use those results to offer some policy recommendations. II. THORETICAL FRAHEWORK In this section we develop an optimal model of private investment behavior in the context of an owner-managed representative firm operating in an economy characterized by credit rationing, and administratively set interest rates. To finance its investment outlays, the firm can resort either to its own internally generated source of funds, i.e. retained earnings, or to borrowing from the 5 banking sector.7 Also, siuce the firm is owner-managed, no managerial incentive problems are operative, and the obj. tive of the firm is to maximize the present value of future stream f net cash flows. Formally, let net cash flow in real terms at time t be R(t), and p be the discount rate, then the long-term objective of the firm can be stated as, max f e-PtR ( t) dt (1) By virtue of the assumption that the firm cannot resort to new equity issue through the stock market, its net cash flow (income) can be derived simply as net operating profit (net of taxes and interest) plus new bank loans that the firm can raise in the market, minus debt amortization, and minus new investment expenditures. R(t) = (1-T) t(IKM.)1 - t(1-?)r(t)-x(t)+a]B(t) +b(t) -q(t)I(t) (2) where w(.) - before tax profit, taken to be an increasing and concave function of K, i.e. ->0, aarsO, ' the stock of real capital, B - total debt outstanding, s - the corporate profit tax rate, r - interest rate, a - amortization rate on debt, q - the price of capital goods relative to outpat, I - gross real investment, x - rate of inflation, and all variables refer to time t. If not further constraints were operative, the firm would solve the optimization problem (1), (2) subject to the evolution of the stocks of capital and debt; ' The implicit assumption here is that the firm does not raise capital through issuance of new equity. This assumption is likely to apply to many developing countries where the equity market is not sufficiently developed or broad to serve as an important source of external capital to the firm. For the case that the firm has the option of recorse to equity market, see Dailami (1990). 6 kwr( t) -8K( t) ardh( t) -b I F' -aB( t) where 8 is the rate of depreciation of physical capital. In this case it can be shc,w. that the firm would raise its desired amount of loan at each point in time. At the margin, the rei I cost of equity, p , would be equal to the real effective cost of debt 1(1-v)r-xj , which implies that the only financial variable relevant to the firm's investment decision is the real after-tax rate of interest. In other words, the firm would face an infinitely elastic supply of credit. Figure 1 illustrates for a given point in time the resulting equilibrium for the firm. At point R no further credit is desirable, though it is still available at a fixed cost. This situation, however, hardly describes the reality of loan markets in developing countries, where excess demand for credit and consequently quantity ciedit rationing prevail. To formalize this phenomenon, we introduce a time varying bank credit ceiling imposed on each firm. Such a ceiling, furthermore, will be assumed to be dependent on the firm's net-worth and on the state of credit market as de3cribed by the following equation: b~ ~ ~ ~~~~~f (t)s Xt ;5 )),4o , a 20 (3 ) where v(t) is the firm's net worth at time t, and tb cunction 4 (v, S) defines the bank credit ceiling imposed on a firmv ith net worth v(t), at time t, when the state of credit market is indicated by S(t). Thr state of credit market depends, presumably, on an array of factors including the level of interest rates, the pace of credit policy, the stringency to which regulation and supervision on the banking sector is imposed. Its significance is to emphasize the possibility that credit supply to a particular firm may change as the result of macro and financial fluctuations, independently of the firm's credit worthiness, i.e. its net worth. Thus, in times of tight money, i.e., "bad times," borrowers that otherwise would have been considered eligible for credit are likely to face a more stringent credit evaluation process or simply denied 7 loans. Also, for a given state of the credit market (i.e. for a given value of "S") the position of the credit ceiling will depend on *.he firm's investmevt and borrowing history (i.e. on its current net worth). There is, then, a second reason (apart from future output and profits) to invest now: by doing so. the firm can optimally relax tomorrow's borrowing constraint. A graphical representation of the supply of credit schedule facing the firm under binding credit const-aint is shown in Figure 2. The equilibrium point is shown by point E", where the firm's demand for credit intersects its supply schedule in tho vertical portion. In this case investment will be constrained by r'pply of credit; indeed, we have p>(1-v)r-x , implying that there are investment opportunities with positive net present value which are not undertaken due to lack of availability of credit. To solve for the firm's optimal time path of capital accumulation under credit constraint, we formulate the current value Hamiltonian, H(t) as: Hi(t) a(1-0) [ WO I )- I(1-0)r(t) -x(t) ta B(t) +b(t) -q(t) t (t) -X(t) (1(t)-8K(t))+p(t) (b(Ot)-B(t) (4) where A (t) and p (t) are the associated shadow prices of capital and debt respectively. Taking into account the credit supply constraint (1), we define the Lagrangian L(t) as L(t)-H(t) +0(t) ((V(t);(S)) -b(t)) (5) where Q(t) is the Lagrangian multiplier associated with bank credit ceiling (3). Given the assumption that there exists no uncertainty, the necessary conditions for an optimal solution are the following: . (a) - (p+8) A (S) -[(1-) .;-q(s) 0 (a) (6.a) p (s) - (p 4) it (s) + [ (1-1) r(s) -x(s) +a+Q (a) 4,,j (6.b) abiL4bp (a) -0 (a) -0 (6.C) 8 Figure 1: Flrm's Equilibrium Without Credit Ceilings Colt of \E Supply of Credit Demand for Crd Amount of Credit Figure 2: Firm's Equilibrium With Credit Ceilings Cost Of Crdilt -Supply of Credit Demand for Cedit2 Demand for Credit Amount of Credit 9 +A"-q() fA(a) -0 (6.d) u(ts) ( 4 (v (8) -0 simplifying the above equations, we obtain the following two basic equationst (1-0 ( uq(s) [p+8-4(s)-n (8)* (J7) (s)-'Ip+ * s T a x(s)(7) 4 is the rate of inflation in the price of capital goods. Equation (7) describes the equilibrium condition for the firm's optimum level of real capital. It states that, at equilibrium, the firm's after-tax marginal profitability of capital [left-hand side of (7)1, has to be equal to its cost of capital. tNote that the cost of capital depends not only on the usual depreciation, discount and inflation rate but also on an extra term, which captures the effect of the credit ceiling on the firm's investment behavior; it describes the value of having an extra unit of capital today as a way of relaxing the borrowing constraint in the future (because, ceteris paribus, more capital accumulation today means more net worth tomorrow). Naturally, if the credit ceiling constraint (3) is not binding, Q(W)-0 ; and in that case the supply of credit facing the firm would be horizontal at the current rate of interest, and credit policy could affect investment only through the interest rate channel. But it is not difficult to see that so far a. p>(1-?)r-x , or the real return on equity is greater than the after-tax real cost of debt; the credit constraint (3) must be binding. In that case, Q(s)oo, which from equation (8) can be 10 solved to yield; O(t)fexi -(P+*+.uu)du)|P-(1-T)r(3)-x(sa ds (9) which is the capitalized difference in the real after-tax cost of debt and equity per unit of bank credit. To the extent that p>(1-T)r-x or the real cost of debt falls short of the real coot of equity, a condition which is rather prevalent in most developing countries, debt is subsidized, and a measure of such subsidy is given by D (s) in equation (9). From this equation it is also evident th,at the value of Q(t) the shadow price of credit, depends on the future evolution of both interest rate and credit market conditions. Thus, if credit policy is expected to be tightened in the future, i.e. relatively low values of S(s); for s - t, t+1, t+2..., the firm would prefer tn invest more today;' in this way, it will be holding a larger capital stock (and net worth) tomorrow, allowing easier access to credit when it is more scarce. Likewise, the prospect of rising interest rates accelerates investment today. Finally, substituting fr Q(t) from equation (9) into equation (J) and solvir.g that equation for the firm's optimum level of capital. stock yields a function which would depend on the determinants of the firm's profits, the current and future values of interest rate, and other variables describing the 8 Of course, we are assuming that the "ti8hter" credit conditions do not affect the investment opportunities (and profits) of the firm. Also, notice that we are assuming away time-consistency and credibility problems. The government could pronaise a "loose" credit policy path (i.e. big D(t) for t - 0,1,2,..), get private firms to invest "a lot," and then tighten credit (say, for price stabilization purposes). If private agents know the government's objective function (they probably do over time), firms' investment plans will not be independent of the government's course of credit and, presumably, macro policy 11 future evolution of the state of credit market. Thus, it can be concluded that, as long as the firm's discount rate is higher than its cost of debt, (i.e., profitable investment opportunities are available), the current level of investment depends on the future evolution of both interest rate and credit market conditions. This hypothesis provides the basis for our empirical analysis presented in the next section. 12 III. EMPIRICAL FINDINGS In this section we will examine empirically the relationship between credit market conditions and private investment behavior for a sample of five developing countries, i.e. Brazil, Colombia, India, Korea, and Turkey, using annual data from 1965 to 1985. The main objective is to test for the partial and the joint influence of credit availability and interest rates (i.e. quantity and price constraints) on the behavior of aggregate, private investment in that sample of countries. A preliminary plot (Charts I to V) of actual domestic credit and actual private inv..stment (both as percentages of GDP) does not reveal any obvious link. In fact, aggregate private investment seems to exhibit its own independent pattern. It has declined steadily since the second half of the 7Os in the two Latin-American cases, while domestic credit availability has tended to grow. .rivate investment remained flat both in India and Turkey, in the face of growing, albeit fluctuating, domestic credit supply. Korea is indeed the only country in the sample that exhibits an upward trend in aggregate private investment, although at a much slower pace than domestic credit. Of course, our charts do not control for other factors affecting private investment decisions, most notably growth in real economic activity and expected real interest rate. A more thorough regression analysis is needed. Following our model's conclusions in the previous sections, we will specify a 13 basic estimating equation that expresses the ratio of private fixed investment to GDP (noted by RIGDP), a8 a function of three variables: (1) the rate of growth in real GDP, (RGDPG); (it) the expected future value of the ratio of total domestic credit to GDP (EDCGDP); and (iii) the expected future value of the real rate of interest (ERINT). That is: PIGDP4 - *oW + ,, RGDPG0 + 06 EDCGDP1 + U5 ERINT I + u (10) where a., a1, a, and 4 are parameters to be estimated and u, is an error term, and where the credit and interest rate variables indicate their expected, future values as of time "t". Subscript "i" denotes the country (i- Brazil, colombia,...). There exist several methodological problems with the straightforward, ordinary least squares (OLS) estimation of (10) on a country-by-country basis. First, the use of aggregate data makes the right-hand side variables endogenous, creating simultaneous-equation bias. For instance, in any given year "t", an unexpected macroeconomic shock (imbedded in the error term "u,") is likely to drive not only aggregate private investment (the dependent variable) but also the rate of growth of real GDP (an independent one); in this case, the OLS regression coefficient for RGDPG will not only capture the direct effect of movements in this variable but also the indirect effect of ra&dom macro shocks. In other words, our estimate of the influence of RGDPG on its own right will be (in general) inaccurate. Second, expected series are unobservable as we do not know what expectation 14 formation mechanism (if any) agents employ. We need a way to proxy for those series. Third, although we have succeeded in compiling consistent time series data on aggregate private investment for each country, they are still quite short.' Fourth, the structure of the shocks affecting private investment have probably changed over the last two decades (most notably in Korea), raising the possibility of intra-country heteroskedasticity. For instance, the development of more active stock markets could have changed the structure (i.e. the distribution) of informational shocks affecting investment decisions. To overcome (at least partially) those problems, we have pooled the country eguations (not their data time series) and have estimated expression (10) as a simultaneous equation system (for i-Brazil, Colombia,...) by means of the three-stage least squares technique, constraining cross-country parameters to be zero. Specifically, we have estimated the following system: PIGDPW - Nb + Cab RGDPGb + ca, EDCGDPw + 06 ERINT, + ul PIGDPce - 4c + C18 RGDPG., + Qa, EDCGDPt + af ERINT% + uu PIGDP, - ao + t,, RGDPG,4 + ai, EDCGDPM + Ga ERINTt + u,t (11) PIGDPM - 06 + C11, RGDPG, + 06 EDCGDPM + as ERINTm + u. PIGDPqt - °oq + Glq RGDPGqt + aq, EDCGDPqt + a. ERINTqt + uqt 2/ Twenty-one observations per country, from which lead and lags are to be discounted. 15 where: "bO-Brazil, "c"-Colombia, "n"-India, "k"-Korea, "q"-Turkey. As before, "t" stands for time period. It is worth noting that we have pooled the countries' equatiQna, not their time series; for example, Brazil's and Korea's real GDP growth rates are treated as different variables and, hence, will have different estimated regression coefficient. Accordingly, we have generated a regression coefficient for each variable in each countrv°. . This econometric set-up is particularly useful as it allows us to profit from the existence of cross-country contomporaneous correlation of the disturbances". That correlation has probably been generated by international macroeconomic developments; for example, in each country, the oil shocks affected the observations for the years 1973 and 1979. As for the estimation itself, we have used three-stage least squares; this technique has allowed us to incorporate enough instrumental variables in order to deal with the problem of the endogeneity of the right-hand side 1Q/ Other commonly-used pooling techniques pool variables according to their economic meaning, with no consideration for country differences. Thus, for example, real GDP growth in Brazil is pooled together with Korea's in a single series. Then, only one regression coefficient is generated per "pooled" variable. II/ In a separate exercise, we estimated each country's equation individually through two-stage least squares. The corresponding Breusch-Pagan test on the residuals has shown significant contemporaneous correlation, suggesting the suitable joined estimation of the equations. Also, no statistically significant coefficient was obtained whatsoever; in that sense, the much better performance of the simultaneous-equation approach points to the shortness of the series as the probable cause for the initial lack of precision. 16 variables. The latter feature has been particularly relevant as we have proxied expected values by actual ones, instrumenting them with their first (respect co expectation formation time) lag. At the same time, we have estimated heteroskedasticity-consistent standard errors (of the White- Chamberlain type). Our findings are contained in Panels A, B and C of Table 1. When current (i.e. at time "t"), actual credit and interest rate series are used as proxies (and their first-lags as their instruments) for those variables own future path, the results are mixed. In three out of five countries the expected credit availability variable shows up positive and significant (very much so in Korea and Turkey), while it remains insignificant for the rest. Quite a similar pattern for interest rates: negative and significant in Brazil and Turkey, insignificant in India and Korea, but surprisingly positive and significant in Colombia. (See Panel A, Table 1). The results improve considerably when one-year-forward, actual credit and interest rate series are employed (instrumented by their first lags). The credit variables are now statistically significant, implying a positive role for credit availability in the determination of private investment in all countries. As for interest rates, they are also significant throughout, but with counter-intuitive signs both in Colombia and India (See Panel B, Table 1). In Panel C of Table 1, we combine the proxies that appeared most suitable in the previous Panels. That is, we use one-period-forward credit and 17 c interest rates (both actual ones, and instrumented by their respective first lags). The results are quite conclusive. The credit variables are again positive and (very) significant in all countries, while interest rates are either significantly negative or statistically irrelevant. We have so far concentrated in the statistical significance of individual parameters in our "system" of equations; traditional t-statistics suffice for that purpose."' We next test for the joint significance of credit and interest rate variables in each country. In other words, we want to compare the results from the full estimation of (l:) with those of the estimation of (11) after the credit and interest rate variables of a given countlv are removed (i.e. their regression coefficlents are restricted to be equal to zero). For that purpose, we have compared successively the residual performance of the unrestricted estimation of the whole system with its restricted counter-parts, being the restriction applied to one country's equation at a time. The results were straightforward: for none of the specifications in Panel A through C of Table 1, and for no country, could we accept the hypothesis that the credit Bd interest rate variables are not relevant (See Table 2). In other words, both the credit supply and the interest rate matter, Jointly, for private investment outlays. 12/ See, for instance, Judge et.al. (1985) Chapter 15. >/ See Judge et.al., op.cit., Chapter 15 for the specification of the corresponding )?-dLstributed statistic. 18 Regression of ggg te r vote nvaestment rate on eal, growth of output, real tLerest rate and credit supply (Annual data 1965-1985) PIDPi ' 01 * * RGDPOIC * %J EDCGDP 1V * * ERlzmJC . Ule Parameter Brazil Colombia India Korsa Turkey 1mL a ea 0.144 0.142 0.088 0.165 0.018 (3.90) (7.00) (25.3) (9.22) (1.24) a1 0.249 0.159 0.064 0.029 0.135 (3.44) (1.59) (1.16) (0.42) (4.53) Ga 0.264 -0.091 0.013 0.140 0.142 (2.56) (-1.04) (1.21) (5.60) (4.13) 'as -0.210 0.112 -0.013 -0.001 -0.056 (-2.13) (2.61) (-0.36) (-0.02) (-4.30) Panel B a* 0.080 0.061 0.085 0.122 0.040 (3.48) (2.60) (23.4) (7.52) (3.12) a, 0.054 0.075 0.026 0.089 0.006 (1.41) (0.61) (1.89) (1.31) (0.22) 62 0.409 0.299 0.026 0.220 0.110 (4.42) (3.18) (2.81) (8.8) (3.33) a3 -0.155 0.093 0.029 -0.129 -0.061 (-2.35) (2.15) (3.04) (-2.41) (-5.22) Panel C sO 0.70 0.063 0.083 0.130 0.026 (2.80) (2.76) (23.0) (7.63) (1.63) a1 0.169 0.144 0.040 0.067 0.098 (3.34) (1.37) (1.24) (0.98) (2.91) L2 0.406 0.2.8 0.30 0.207 0.126 (4.02) (2.69) (3.47) (8.21) (3.08) 43 -0.121 0.059 0.006 -0.005 -0.062 (-2.10) (1.28) (0.27) (-0.09) (-3.65) Heteroe0dastLcity-consistent t-statistics in parenten als 19 TABLE 2 Are the Capital Market Variables Jointly Significant in each Coutry? Hx a 0 3' 0 for I - Brazil, Colombia... (Critical Xi at 952 confidence and 2 degrees of freedoms 5.99) PROXIES BRAZIL COLOMBIA I1DIA KOREA TUSY ARE: DCGDt 11.59 6.86 1.93 32.47 25.24 lINTt DCGD., 38.09 21.90 10.85 85.25 33.98 INT.1 DCGDt., 23.31 11.50 13.25 68.61 18.41 RINTt 20 IV. CONCLUSION In their decisions to extend loans, banks are known to be concerned not only with the creditworthiness of the individual borrower, but also with broader issues relating to the stance of credit and interest rate policy, regulatory and supervisory practices, and market sentiments. These factors are often referred to as determining the state of the credit market, and their significance lies in their influence on banks' degree of conservatism and credit rationing. They thus affect the banks' loan supply ca.z-sions independently of the creditworthiness; i.e. net- worth of individual borrowers. Thus, in times of tight money, or in "bad times," borrowers that otherwise would have been considered eligible are likely to face a more stringent credit evaluation process or simply be denied loans. As a consequence, firms' investment plans may have to be postponed or cut back. Motivated by this phonomenon, this paper has developed a simple model of firms' investment behavior in an economy with exogenous, time-varying borrowing constraints. A particular form of credit ceiling linked to the firm's net-worth and to the state of credit market is introduced. The model suffices to illustrate the main theoretica message of the papers It is the joint impact of both the volume of credit an;. its price, i.e. interest rate, which is relevant to firms' investment decisions. Thus, focusing on the impact of interest rate changes in assessing the role of credit policy on investment and growth is misleading to the extent that such changes are accompanied by changes in the volume of credit as well. The model is then used to specify an equation relating aggregate private itvestment to aggregate output, and two credit market variables; real interest rate and aggregate credit. The equation was estimated, using annual data (1965-85) for a sample of five developing countries, i.e. Brazil, Colombia, India, Korea, and Turkey, and the joint significance of both interest rate and credit supply conditions are tested. The evidence strongly supports the joint influence of interest rates and credit volume on the behavior of private investment in these countries. This supports the theoretical conclusion of the paper that effective 21 credit policy in the context of developing countries must ac!'nowledge the influence of both the credit supply and the interest rate channels. This provide, a usoful point of convergence between previous approaches to credit policy and its role In the stabilization and adjustment processes of developing countries, which have either emphasized the role of the availability of credit, or its price; l.e., the interest rate. 22 Chart I BRAZIL 36 34- 32- 30 , 28- 26- 24- 22- 20 - 16- 14 - M 12 1965 I 1967 1969 4 1971 1 1973 1 1975 I 1977 l 1979 I 1981 1 1983 1 1985 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 Year 0 Priv. Invt./GDP + Dom.Cred./GDP 23 Chart 11 COLOMBIA 27 26- 25- 24- 235 22- 21- 20 - 19 18 17 16- 15 14- 13 12- 1 1 10 9 19'65 19'67 19369 j 171 1973 19"75 1977 19'79 19'81 19'83 19`85 1966 1968 1970 1972 1974 1976 1978 1980 19Z2 1984 Yeor 0 Priv.lnvt./GDP + Dom.Cred./GDP 24 Chart III INOIA S55 50 45- 40- 35 _ K 30 _ 25 - 20 - 15 _ 10 _ 1965 I 1967 I 1969 I 1971 I 1973 I 1975 I 1977 I 1979 I 1981 I 1983 I 1985 196'; 1968 1970 1972 1974 1976 1978 1980 1982 1984 Year O Priv.lnvt./GDP + Dom.Cred./GDP 25 Chart IV KOREA 70 60 50 - 40 3u 20 - 1965 1 1967 I 1969 I 1971 I 1973 1 1975 I 1977 I 1979 I 1981 4 1983 4 1985 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 Year O Priv.invt./GDP + Dom.Cred./GDP 26 Chart V TURKEY 55 50 _ 45 - 40- 35- 30 _ 25 - 20 - 10 _ 5 19'65 I 19'67 I 19'69 I 19'71 1 19'73 I 19'75 1 i77 I 1979 19'81 4 1983 19485 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 Year 0 Priv.invt./GDP + Dom.Cred./GDP 27 REFERENCES Balassa, B., (1982), "Disequilibrium Analysis in Developing Countriest, World Bank Development. Baltensperger, E. and T. M. Devinney (1985), "Credit Rationing Theory: A Survey and Synthesis," Zeitschrift fuer die Gosamte Staatswissenschaft, pp. 475-502. Bernanke, B. and M. Gertler (1989), "Agency Costs, Net Worth, and Business Fluctuations," American Economic Review, March, pp. 14-31. (1990), "Financial Fragility and Economic Performance," Quarterlv Journal of Economics, pp. 87-114. Calomiris, C. (1990), "Firm Heterogenity, Internal Finance and Credit Rationing," Economic Journal, pp. 90-104. Dailami, M., (1990), "Financial Policy in Imperfect Capital Markets: The Case of Korea", PRE Discussion Paper, No. 409, The World Bank. Galbis, V. (1977), "Financial Intermediation and Economic Growth in Less Devejoped Countries: A Theoretical Approach", Journal of Develonment Studies. Gertler, M. (1990), "Taxation, Corporate Capital Structure, and Financial Distress," in L.H. Summers, ed., Tax Policy and Economy, vol. 4, Cambridge, Mass. MIT Press. Hodgman, D. (1960), "Credit Risk and Credit Rationing," Ouarterlv Journal of Economics, pp. 258-278. 9 (1962) "Reply", Ouarterlv Journal of Economics, pp. 488-493 Jaffee, D. and T. Russel (1976), "Imperfect Information, Uncertainty and Credit Rationing," Quarterly Journal of Economics, pp. 651-666. Judge, A., W. Griffiths, R. Carter Hill, H. Lutkepohl and T. Lee, The Theory and Practice of Econometrics, (2nd Edition), J. Wiley, London. Khatkate, D.M., (1972), "Analytical Basis of the Working of Monetary Policy in LDC: A Prologue:, IMF Staff Papers. Keller, P.M., (1980), "Implications of Credit Policies for Output and the Balance of Payments", IMF Staff Papers. Leyland, M. E. and D. M. Pyle (1977), "Informational Asymmetrics, Financial Structure and Financial Intermediation," Journal of Finance, pp. 371- 87. Lindbeck, A., (1962), "The New Theory of Credit Control in the United States," Stockholm Economic Studies, 2nd ed. McKinnon, R., (1973), Money and Finance in Economic Develolnment, Washington, D.C., Brookings Institution. Polak, J.J. and V. Arqy, (1971), "Credit Policy and the Balance of Payments:, IMF Staff PaRers, pp. 1-24. ROE, A., (1982), "High Interest Rates: A New Conventional Wisdom for Developmeut Policy? Some Conclusions from Sri Lankan Experience:, World Develoament. Rossa, R. (1951), "Interest Rates and the Central Bank," in Money .Tra4d and Economic Growth: Essays in Honor of Jaohn H. Williamst, New Yorks 28 Sauve, P. (1986), "Aspects of the Debate on INterest Rate Policies in Developing Countries," Canadian Journal of Development Studies, pp. 211-222. Scott, Jr., (1957), "The Availability Doctrine: Development and Implicatidon," Canadian Journal of Economics and Political Science, pp. 532-539. Stiglitz, J. and A. Weiss (1981); "Credit Rationing in Markets with Imperfect Information," American Economic Review, pp. 393-410. Taylor, L. (1979), Macro Models for Developing Countries, Ch. 9, (McGraw- Hill, New York). , (1981), "IS/LM in the Tropics: Diagrammatics of the New Structuralist Macro Critique," in W.R. Cline and S. Weintraub, eds., Economic Stabilization in Developing Countries, (The Brookings Institute, Washington, D.C.). Van Wijnbergen, S. (1983), "Credit Policy, Inflation and Growth in a Financially Repressed Economy, "Journal of Development Economies, vol. 13, August-October, pp. 45-65. World Bank, (1989), World DeveloDment Regort, Oxford University Press. PRF Working Paper Series Contact Autbho fgra"ox WPS632 The Macroeconomics of the Public Carlos Aifredo Rodriguez March 1991 R. Luz Sector Deficit: The Case of Argentina 34303 WPS633 The Macroeconomics of the Public Virabongse Ramangkura March 1991 R. Luz Sector Deficit: The Case of Thailand Bhanupongse Nidhlprabha 34303 WPS634 Trends in Developing Country Bela Balassa March i991 WDR Office Exports, 1963-88 31393 WPS635 Exchange Rates and Foreign Trade Bela Balassa March 1991 C. Cuskelly-Young in Korea 39413 WPS636 Economic Integration in Eastern Bela Balassa March 1991 C. Cuskelly-Young Europe 39413 WPS637 Poverty in Poland, 1978-88 Branko Mllanovic March 1991 WDR Office 31393 WPS638 Researching the Trade-Productivity James Tybout March 1991 D. Ballantyne Link: New Directions 37947 WPS639 The High Cost of Protecting Wendy E. Takacs March 1991 D. Ballantyne Uruguay's Automotive Industry 37947 WPS640 The Impact of Policy in African William Jaeger March 1991 A. Yideru Agriculture: An Empirical Investigation 34663 WPS641 Intertemporal Substitution in Karsten N. Pedersen April 1991 M. Hileman Consumption: Evidence for Some 31284 High- and Middle-Income Countries WPS642 How a Change In Brazils Sugar Brent Borrell April 1991 A. Kitson-Wakters Policies Would Affect the World 33712 Sugar Market WPS643 Regional Integration among Andras Inotal April 1991 D. Ballantyne Developing Countries, Revisited 37947 WPS644 Trade and Payments Arrangements Constantine Michalopoulos April 1991 M. Colinet in Post-CMEA Eastem and Central David Tarr 34698 Europe WPS645 Poverty, Policy, and Industrializa- Ben Polak April 1991 WDR Office tion: Lessons from the Distant Past Jeffrey G. Williamson 31393 WPS646 The Developmental Effectiveness Tony Killick April 1991 S. King-Watson of Aid to Africa 33730 WPS647 Growth Rates and Aggregate Nanak Kakwani April 1991 B. Rosa Welfare: An Intemational Comparison 33751 PRE Workino Paper Series Contact Dtk Author pape WPS648 Who Paid the Bill? Adjustment and M. Louise Fox April 1991 WDR Office Poverty in Brazil, 1980-95 Samuel A. Morley 31393 WPS649 An Observation on the Bias in Margaret E. Grosh April 1991 B. Diallo Clinic-based Estimates of Kristin Fox 30997 Malnutrition Rates Maria Jackson WPS650 Administrative Valuation of Soviet Karen Brooks April 1991 C. Spooner Agricultural Land: Results Using 30464 Lithuanian Production Data WPS651 Taxation of Financlal Assets in Christophe Chamley April 1991 A. Bhalla Developing Countries 37699 WPS652 Demographic Response to Economic Kenneth Hill April 1991 WDR Office Shock 31393 WPS653 The Effects of Option-Hedging on the Donald F. Larson April 1991 D. Gustafson Costs of Domestic Price Stabilization Jonathan Coleman 33714 Schemes WPS654 Reflections on Credit Policy in Mansoor Dailami April 1991 M. Raggambi Developing Countries: Its Effect on Marcelo Giugale 37657 Private Investment WPS655 Interest Rate Policy In Egypt: Its Mansoor Dailaml April 1991 M. Raggambi Role in Stabilization and Adjustment Hinh T. Dinh 37657