Determinants of private investment in LDCs : the case of Colombia

For Staff Use Only DETERMINANTS OF PRIVATE INVESTMENT IN LDCS: THE CASE OF COLOMBIA M'hamed Cherif and CarLos Montes Division Working Paper No. 1986-1 February 1986 Country Analysis and Projections Division Economic .n;dy~s and Pr*jectivns Department Economics and Research Staff The WorLd 3ank Division Working Papers report on work in progress and are circulated Eor Bank staff use to stimulate discussion and comment. The views and interpretations in this document are those of the author(s).  TABLE OF CONTENTS Page No. Introduction .......................................................... 1 1.0 Theoretical Background ....................................... 2 2.0 Empirical Analysis: The Colombia Case ...................... 10 2.1 The Accelerator Variables ................................... 11 2.2 Credit to the Private Sector ................................ 12 2.2.1 Domestic Sources .................................... 12 2.2.2 External Sources .................................... 15 2.3 The Interest Rate.............................................18 2.4 The Real Exchange Rate ...................................21 2.5 Finding the Elasticities .................................... 22 2.6 Discussion of Other VariabLes.................................. 25 Conclusions .....................................................26 List of TabLes Table 1: Key Policy Instruments ............................... 3 TabLe 2: Key Performance Indicators ...........................4 Table 3: Impact of Foreign Capital Inflows ....................16 Table 4: Constant Elasticities ................................21 Table 5: Elasticities from the Linear FormuLation ............23  A B S T A C T This paper aims ac identifying the economic indicacors invoLved in the process of determining the speed of adjustment of actual private investment to notional demand. We test the statisticaL significance of indicators such as domestic and foreign credit, interest rates, exchange rate and government investment. Notional demand is derived here Erom the flexible accelerator principle. In spite of the simplicity of this assumption, econometric results are shown to be statistically significant. The major conclusion is, that during the period considered, privace investment seems to have been constrained by the availability of domestic credit.  Introduct ion The aim of this paper is to assess the importance of vzarious variables in explaining capital formation by the private sector in LDCs. The focus will be on policy instruments developing countries can use to impLement a certain development policy. It is assumed here that governments have a direct control on investments made by the pubLic sector, but can influence private investment only indirectly by resorting to fiscal and monetary policies. The objective of this paper is to highlight the importance of the key policy instruments impacting on private investment. By identifying the significant policy instruments affecting the behavior of investors, it would be possible to design medium-cerm strategies aimed at bringing about the major macroeconomic equilibria and putting the economy back on a steady growth path. The problem that LDCs have faced in recent years is indeed to restore balance of payments and government finance equiLibrium without impeding growth in the medium-term. Experience has shown, however, that countries that succeeded in achieving both stability and medium-term growth are rare. One reasorn behind this failure might be the across-the-board cuts in expenditures and removal of incentives to invest. The trade-off between short-term stability and medium-term growth is without doubt a difficult issue to deal with both from the theoretical and pragmatic points of view. Recent analysis of structuraL adjustments in LDCs has demonstrated the need to anaLyze th.is trade-oCf wichin a general equiLibrium framework. This framework shouLd integrate .he financial and monetary sectors with the reaL side of the economy. 37 doing so, one can analyze che impact of various stabiLization efforts  5CO. g-COLFRME - 2 on the determinants of growth such as efficiency, capacity utilization and capital formation. This present study would feed such a framework and provide an essential Linkage between fiscal and monetary instruments and capital formation by the private sector. The choice of Colombia as a case to illustrate the present analysis stems from, the fact that Colombia experienced diEferent development phases corresponding to different policy stances. Tables 1 and 2 summarize the evolution of key policy instruments and performance indicators over the period 1961-84. The variability in policies should allow us to test the significance of key policy instruments on private investment. Section 1 will be devoted to the theoretical background for estimating investment functicns in LDCs. The analysis wilL be carried out at an aggregated Level because of the lack of sectoral information. However, the discussion will be general in order to allow applicability to other countries and at a sectoral level. In Section 2, we will present empirical results assessing the significance of each of the policy instruments and other variabLes on private investment. Finally Section 3 presents a summary of the major findings and some concluions. 1.0 TheoreticaL Background There are several ways in the economic Literature to approach eCxpLaining the behavior ot investors, i.e., (i) t .e kenestan appru;icn, which emphasizes the roLe of the incerest rate and the capitaL stock;  ECO.R-COLFRAME - 3 - Table 1: KEY POLICY INSTRUMENTS Reserve Growth Rates Requirement Interest Monetary Exchange Claims on Ratio (%) Rate (x) Base Rate Government 1960 13.9 - 1 - 1 11.3 - 13.2 1.0 18.7 2 20.4 - 32.9 3.9 91.3 3 11.3 - 1.6 29.3 12.7 4 16.3 - 60.9 0 26.6 5 18.0 - -2.8 16.4 42.4 6 16.3 - 11.5 28.9 -4.8 7 20.0 - 25.0 7.5 11.1 8 26.5 - 27.5 12.3 -1.5 9 30.7 - 26.7 6.3 -2.3 1970 30.0 13.6 18.8 6.5 -6.4 1 27.3 13.6 7.1 8.1 34.7 2 26.1 13.6 25.5 9.7 1.3 3 28.6 13.6 32.1 8.1 -9.5 4 27.4 26.2 23.7 10.3 61.2 5 28.1 26.2 32.1 18.7 22.8 6 30.8 25.6 41.8 12.2 38.6 7 32.2 25.6 39.6 6.0 28.9 8 49.0 24.4 52.7 6.3 -25.1 9 53.4 25.6 28.3 8.8 -150.9 1980 46.2 36.9 29.7 11.1 138.8 1 41.3 38.6 24.3 15.2 -33.0 2 38.9 35.9 18.5 17.6 -535.1 3 33.5 34.2 21.0 23.t l00.A 4 41.0 86.0 - 2 1L .6  ECOeR-COLFRAME 4 Table 2: KEY PERFORMANCE INDICATORS GDP Current Real Effective Private Growth Inflation Account Exchange Rate 1/ tnvestmenc Rate (%) Rate (%) into GDP (%) (1980 100) Crowth 1961 5.3 7.9 -3 9 2 5.5 5.3 3 .0 3 2.9 21.9 -3.6 4 6.6 14.9 - 5.6 5 2.8 11.9 -10.7 -0.9 6 5.7 11.4 -15.2 82.2 4.8 7 4.2 10.1 -11.3 83.2 2.6 8 6.4 7.8 -2.7 90.5 17.2 9 6.3 7.0 -2.7 91.3 0.5 1970 7.0 7.0 -4.1 96.6 6.2 1 6.0 12.2 -5.8 102.2 3.1 2 7.7 11.7 -2.2 108.2 -3.7 3 6.7 19.2 -0.5 111.6 4.9 4 5.7 26.9 -2.8 111.1 13.2 5 2.3 24.5 -0.8 118.4 -8.0 6 4.7 20.5 1.3 115.8 20.0 7 4.2 27.1 2.3 100.4 -2.4 8 8.5 17.8 1.4 106.8 13.1 9 5.4 26.4 1.7 101.2 5.3 1980 4.1 26.3 0.4 100.0 13.3 1 2.8 25.9 -4.5 92.0 5.6 2 0.9 22.0 -6.5 87. 8 3 0.9 23.4 -6.3 89.2 4 3.1 !/l .6 -5.0 92.1 Decrease impLies an appreciation (increase impLies depreci'atLon).  SCO. R-COLFRAME - 5 - (ii) the accelerator models, simple and flexible, which concentrates on the technological relations between capital and output and aLlowing partial adjustments to achieve the desired capital stock; (iii) the liquidity theory that considers the effect of the cost of internal funds and its availability on the chosen capital/output ratio; and finally (iv) the neoclassical model which is based on the maximization of the present value of the net profits and therefore giving a central role to the production function and the rental cost of capital services. The attempts to determine empirically which theory fits the facts better has led to contradictory results (see for example, Jorgenson, (1971) and Elliot, (1973)). 1/ In the LDCs case there are two elements that suggest the modification of the.neoclassical framework. First, the assumptions this model requires are unrealistic for countries characterized by the absence of developed capital markets. Second, it requires data that are very difficult to find for LDCs, a serious problem for any empirical work. These problems Led us to use an investment model based on a more ecLectic approach, which does not include directly typical neoclassical variables. It is a flexible variant of the accelerator model, (see Blejer-Khan 1984) 2/ where the speed of adjustment to 1/ Jorgeson; "Econometric Studies of Investment Behavior a Survey"; J.E.L. Dec. 1971. ElLiot, J; "Theories of Investment Behavior Revisited", A.E.R. March, 1973. 2/ Blejer, M. and Khan, M. "Government PoLicy and Private Investment in Developing Countries" IMF Staff papers, June 1984 or a simiLar approach by Tun Wai, U. and Wong, C. "Determinants of Private Investmenc in Developing Countries", Journal of Development Studies, October 1982.  sco cR-COLFRAME - 6 - achieve the desired capital (B), is a function of other economic indicators, i.e. (1) AKPt = B (KP KPt ) where KP and KP are desired and actual private capital stock respectively. To simpLify the analysis, let us assume that the desired capital stock is proportional to the desired output level and that investment adjusts to the optimal capital stock, i.e. * * (2) If IPt = [l - (1 - 6) L] KP , and KPt = aY*, then we can derive the flexible accelerator formulation of the demand for gross investment as follows: (3) IPt = B a[l - (1 - 6) L] Yt + (1 - B) IP where 6 is the depreciation rate, L the lag operator, Y the desired output. If we additionally assume that the adaptative expectations scheme for output is formed in the following way: t t- 1 [LI> - 1 (where X is the coefficient of expectations) and that the coefficient 3 is a linear function of other policy variables (Xi), 1/ we get our final equation (5) 2/ c. X. I. it 1/ St b + L lt (IP i Pt-1 t -I 2/ This resuLt is found easier working with the expressions AIPt = B (P Pt -1) and IP =1 - (1 - 6) L ] KP  cco.R-COLFRAME -7- (5) IP l - (1 - i) LI = Xb a - (1 - 6) Y - Ct t I + ( - (1 - \) L (ci Xit + (I b) IPt _ 11 The investment theories usually consider two different steps: the determination of the desired capital stock (K) and its t reaLization through investment. With this approach the change in the reLevant variables (except in the case of the expected output) does not affect directly the desired capital stock, but the velocity at which the agents adjust the discrepancy between KPt and KPt Although these two processes are conceptually different, this does not pose a serious problem for an empirical analysis of the determinants of the private investment. It sh--ild be noted that the relation with the expected output, the accelerator relationship, will tend to be weaker if there is a substantial degree of unused capacity, although this will not be critical depending on the eLasticity of expectations of the private agents. With the equation (5), we have an estimabLe equation, but we still need to specify the factors that affect the veLocity of adjustment to achieve KP ., the costs and constraints involved in the investment decision. Although the importance of these variables (X.i) will vary according to the different characteristics of the LDCs, in what follows we will discuss what we expect to be, taking into account the investment theori,e., the more imoor,anc variabLes. For most LDCs, Liquidity constitutes a severe constraint on investment, the underdeveloped character of capitaL markets limits the issue cf stocks as a source of financing for firms. Furthermore, hne  ECO.R-COLFRAME - 8 - financial repression -- with often negative real interest rates -- constraints the availabilicy of Loanable funds. In this context the firm can only reLy on internal sources of financing (retained earnings and depreciation allowances) and on the restricted external and domestic credit (either from national or foreign institutions). It should be noticed that this system is biased towards the heavy indebtedness of the privileged firms with access to the credic. The specific country cases are certainly differcnt but the roLe of credit is central for most LDCs; a fact that does not necessarily exclude the relevance of the interest rate in the investment decision, either altering the direct cost of borrowing or the opportunity cost of the investment. I/ For the regression analysis, it is difficult to find a proxy for self-financing but since it tends to be correLated with the output it does not constitute a serious probLem in our formulation. 2/ In the case of the credit from financial institutions, the foreign sources should be considered only for countries with consistent external indebtedness. Another characteristic of most LDCs is that they import a significant share of their capitaL goods. For these countries, the availability of the foreign exchange (e.g. reserves,/imports, ) and its real cost [the real exchange rate, (RER) and the import system] should be considered relevant variables. 11 When the internal financing seems to be particuLarl.y relevanc, one can try to incLude the totaL money suppLy. 2/ It can also affect the avaiLabilicy of credic chrough the change in financial savings.  ECO.R-COLFRAME -9 - For our estimation, we must take into account that the availability of foreign exchange tends to have a positive correlation with the changes in the real credit to the private sector and therefore reduces the necessity of incorporating the first variable. 1/ On the other side, the inclusion of the RER variable for countries with a strong dependence on the imports of capital goods is reinforced by its effect on the cost of the firm's foreign borrowing. We are conscious that besides the RER the inclusion of other elements of the import system is important but the data availability will tend to restrict this. There are other eLements which shouLd be considered for specific cases, such as the effect of different kinds of public investment or of tax and incentive policies, but again here the data restrictions are severe. The public investment in infrastructure clearly stimuLates the private investment by increasing the domestic demand and providing a better environment for the private sector but on the contrary, the noninfrastructure public investment will tend to destimulate the private one by physical crowding out and competition. Finally, but most important, nowhere is security and political stability more influential on the private investor's decision than in LDCs. Nowhere it's more true that "... if the animal spirits are dimmed and the spontaneous optimism Ealters, Leaving us to depend upon nothing / this ii not necessariLy the case when the fvre;gn edchan&e C-rIs9s r, accompanied by a decrease in the real credit to che pubLic sec or.  ECO.R-COLFRAME - 10 - but a mathematical expectation, enterprise wilL fade and die..." 1/ Unfortunately, this is not easy to take into account in empirical investigations. 2.0 Empirical Analysis: The Colombia Case In this section we will apply to the CoLombian case the private investment framework that we formulated in the previous section. Since the objective is to show the methodology, we will present and discuss all the results and not only those that are favorable. Following is a list of the variables that we will be using in the testing of the private investment equation, these are available for the 1951-1981 period, which gives us enough degrees of freedom for a statisticalLy significant analysis: IPt : private investment GDP CDP- (1 - 6) GDP 1where 6 has been assumed as 0.05 ACREDF : Change in the real credit to the private sector (from the central bank, commercial banks and other financial institutions, line 52d in IFS) ACRED : Change in the real credit to the private sector (from the central and commercial banks onLy, line 32d in IFS) RER : ReaL exchange rate INTRA : Lending interest rate LX J. M. Keynes, The CeneraL Theory o, .EmpLoyment,