Information about Western Hemisphere Hemisferio Occidental
Journal Issue
Share
Article

Public Investment, Crowding Out, and Growth: A Dynamic Model Applied to India and Korea

Author(s):
International Monetary Fund. Research Dept.
Published Date:
January 1980
Share
  • ShareShare
Information about Western Hemisphere Hemisferio Occidental
Show Summary Details

It is widely believed by policymakers and analysts in developing countries that public investment provides a significant stimulus to private investment and thereby serves as a powerful instrument of stabilization and growth policy. Although this belief is part of the conventional wisdom, it has not been properly tested against empirical evidence. In fact, the set of interrelationships between private and public investment has remained generally unexplored in the empirical literature on developing countries, although the broader topic of the impact of public spending on private investment, especially its financial aspects, has been extensively analyzed for developed countries. 1 The purpose of this paper is to examine critically the relationship between public and private investment in a developing country by postulating a dynamic model of investment, savings, and growth and by testing and simulating it for two countries in which the public sector has played a significant role. The countries selected for this purpose are India and Korea, both with large public sectors, and, despite the differences in philosophy and the composition of public investment, with an important role for it in their growth strategies and stabilization policies. As an interesting sidelight of this analysis, the role of interest rate policy is also examined, with a focus on its effects on investment and economic growth.

The analysis is conducted within the framework of a growth model that is designed to highlight the role of public investment. The model consists of relationships explaining the behavior of private investment, savings, and growth, and it incorporates several channels through which public investment influences private investment. First, public investment competes with the private sector for scarce physical and financial resources and thereby exerts a negative influence on private investment, at least in the short run. Second, to the extent that public investment complements private investment by creating infrastructure and raising the productivity of the private capital stock, private investment requirements per unit of output are reduced. Third, increased public investment raises the demand for the output of the private sector; it thereby influences output expectations and investment requirements of the private sector. Finally, public investment raises aggregate output and savings, supplementing the economy’s physical and financial resources, and thus offsets at least a part of any initial crowding-out effects on private investment. These channels of influence of public investment account for most of its immediate and final effects and are explicitly built into the structure of the model. Within this framework, we address the critical issue of whether the positive effects of public investment are strong enough to offset its negative effects, and, if they are, within what time horizon. This issue is clearly important in assessing the growth effects of stabilizaton programs involving control of public expenditures. By dynamically simulating the model, the time path of the response of key variables to changes in public investment is derived and compared for the two countries.

The specification of the model uses an adaptation of the neoclassical theory of investment behavior as developed by Jorgenson. 2 The Jorgenson version is itself a variant of the flexible accelerator model of investment, with the capital-output ratio allowed to vary with the relative price of capital input. 3 The adaptation of the neoclassical framework has two novel features: the inclusion of the effects of public investment and the incorporation of the constraint on investment imposed by the availability of savings. The crowding-out effect of public investment is specified in a general fashion, which encompasses not only the crowding out in the financial markets—the traditional notion of crowding out—but also the crowding out in the market for real resources. Moreover, in most developing countries, the crowding out of private investment occurs mainly through some nonprice rationing mechanism, rather than through the price mechanism. Also, in these countries, excess demand for investment exists, stemming partly from widespread financial repression; therefore, actual investment in fixed capital is constrained by the availability of savings. These features of developing countries are built into the model by specifying a separate aggregate domestic savings function and by postulating a direct linkage between savings and the speed of adjustment of the fixed capital stock to the desired level.

The effects of interest rates on investment and savings are also included in the model. The nominal interest rate, by affecting the cost of capital as well as the real interest rate, sets in motion a chain of responses influencing the desired level of the capital stock and its productivity, as well as the availability of savings and the consequent speed of adjustment of the actual capital stock to the desired level. Thus, the model provides an integrated framework for analyzing the various effects of interest rates that are emphasized in the classical and modern theories of interest rates.

The background to the analysis of the impact of public investment is described in Section I. The model is derived and discussed in Section II. The estimated equations are discussed in Section III. Simulations of the model are discussed in Section IV. Some concluding comments are presented in Section V.

I. Public Sector in India and Korea

The scope of the public sector in India and Korea is illustrated in Tables 1 and 2. Since the advent of organized planning in both countries, their public sectors, including both general government and public enterprises, have undertaken a major share of investment and production (see Table 1). Official data on general government in India and Korea are comparable, but data on public enterprises are not strictly comparable owing to differences in classification. Moreover, official data on the output of the public enterprise sector in Korea are not available. An official estimate of the share of the public sector in total investment in Korea during 1970-75 was 23 per cent, which was lower than the 41 per cent share in India for the same period. However, comparable estimates for Korea’s public sector provided by Jones (1975) suggest that the share of the public sector in total investment in Korea was almost 50 per cent in 1971-72, considerably higher than the share in India for the same period; the share of the public sector in total output was about the same in both countries.

Table 1.India and Korea: Relative Size of Public Sector(In per cent; period averages)
YearsGeneral Government1Public EnterprisesTotal Public Sector
Share in Total Fixed Investment
India
1960-65133649
1965-70103141
1970-75122941
Korea
1960-65151126
1965-7016824
1970-7517623
Share in Total Gross Domestic Product
India
1960-656612
1965-706814
1970-757916
Korea
1963-64117218
1971-7289217
Sources: India, Central Statistical Organization, National Accounts Statistics, various issues; Bank of Korea, Economic Statistics Yearbook, 1978; and Jones (1975).

Public administration, defense, and education.

Estimates from Jones (1975). Official estimates of public sector output are not available. In view of the economic definition of public enterprises used by Jones, which is more inclusive than the official definition, the shares in GDP are not comparable with the shares in investment, which are based on the official definition. The comparable investment shares of public enterprises given in Jones (1975) are 33 per cent in 1963-64 and 31 per cent in 1971-72.

Sources: India, Central Statistical Organization, National Accounts Statistics, various issues; Bank of Korea, Economic Statistics Yearbook, 1978; and Jones (1975).

Public administration, defense, and education.

Estimates from Jones (1975). Official estimates of public sector output are not available. In view of the economic definition of public enterprises used by Jones, which is more inclusive than the official definition, the shares in GDP are not comparable with the shares in investment, which are based on the official definition. The comparable investment shares of public enterprises given in Jones (1975) are 33 per cent in 1963-64 and 31 per cent in 1971-72.

Table 2.India and Korea: Share of Public Enterprises in Net Domestic Product of Selected Sectors1(In per cent)
IndiaKorea
Sector1964/651975/7619631972
Banking and finance41767687
Construction7945
Manufacturing7151515
Mining11783731
Transport and communications64544730
Sources: India, Central Statistical Organization, National Accounts Statistics, various issues; and Jones (1975).

Indian fiscal years extend from April 1 to March 31; Korean fiscal years are calendar years.

Sources: India, Central Statistical Organization, National Accounts Statistics, various issues; and Jones (1975).

Indian fiscal years extend from April 1 to March 31; Korean fiscal years are calendar years.

The public enterprise sectors in both countries span a wide spectrum of industries, although their evolution has been marked by sharp contrasts (see Table 2). The share of public enterprises in the nonagricultural components of net domestic product is sizable in both countries and shows significant shifts over time. In India, the share in the transport and communications sector declined during 1964-75, while the shares in most other sectors increased. Indeed, the shares in both the manufacturing and banking sectors nearly doubled, and the share in mining increased sevenfold. The changes in the sectoral composition of the public enterprise sector in Korea were not as sharp. The share of public enterprises in manufacturing and construction remained stable; the share in banking rose from an already high level; and the shares in mining and transportation declined. The large share of the public sector in the two countries, however, masks their divergent motives for the expansion of the public sector. In Korea, the motivation was primarily pragmatic, in that the public sector entered areas where private investment was not forthcoming or entered strategic areas where excessive concentration of private power was feared. Although these considerations also played a role in India, they were overshadowed by ideological considerations related to the important role assigned to the public sector in the policy of evolving a socialist pattern of society.

The real output of the public enterprise sector in India increased at an annual rate of 9 per cent during 1965-76, compared with the gross domestic product (GDP) growth of 4 per cent. Output of the public enterprise sector in Korea grew by 15 per cent per annum during 1963-72, compared with GDP growth of 10 per cent. In view of the leading role of the public sector in the two countries, an understanding of the linkages between the public sector and the rest of the economy is clearly important, both in its own right and in enabling one to assess the role of public investment as a policy instrument. In both countries, the public enterprise sector has extremely high forward linkages. The growth of the public sector in India has been predominantly in sectors with high forward linkages (e.g., steel, machinery), while that in Korea has been in sectors with both forward linkages (e.g., machinery) and backward linkages (e.g., consumer goods). In both countries, public enterprises play a significant role in capital-intensive sectors and in sectors exhibiting economies of scale. These and other features of the public sector have implications for the impact of public investment on private investment and on the overall growth of the economy. An understanding of this impact is essential not only in formulating development plans in which both private and public enterprises are expected to play a coordinated role but also in designing stabilization policies to stimulate aggregate investment.

A preliminary view of the scope of public and private investment and their interrelationship in the two countries is provided in Charts 1 and 2, which show time series on real fixed investment in the two sectors, both in absolute terms and as proportions of GDP. The experiences of the two countries reveal significant differences in the relationship between public and private investment. A generally positive association between public and private investment is evident in Korea (see Chart 1). The movements in the two series are broadly in the same direction, suggesting no apparent tendency for one sector to crowd out another or for the public sector to play a countercyclical role by compensating for variations in private investment. However, fairly divergent movements in public and private investment can be seen in India, particularly since 1965 (see Chart 2). A generally positive, albeit weak, association between public and private investment can be seen during 1955-64, giving way to a negative association during most of the decade to 1976. The decline in public investment in the late 1960s in the wake of interruptions in planning was accompanied by an acceleration in private investment. During the subsequent recovery of public investment in the early 1970s, private investment stagnated. Thus, taking the entire period 1965-76, there has clearly been a negative correlation between public and private investment; the partial correlation coefficient between the two variables, after adjusting for the time trend, was significantly negative.

Chart 1.Korea: Public and Private Investment, 1956-76

Chart 2.India: Public and Private Fixed Investment, 1955/56-1976/77 1

1 Indian fiscal years extend from April 1 to March 31.

The observed negative correlation between public and private investment casts some doubt on the view that the slackening of economic growth in India since the mid-1960s has been mainly a result of sluggish public investment and the view that vigorous growth in public investment is a prerequisite for faster overall growth. 4 If a decrease in public investment is offset by an increase in private investment, then overall growth need not decline. In fact, during 1965-70, the ratio of total fixed investment to GDP remained generally stable, despite the fall in the ratio of public investment to GDP. Of course, a fall in public investment might shift the industrial composition of investment in a direction that would be contrary to the overall plan. The private sector activities closely linked with capacity creation in the public sector would suffer, while other sectors would take up the slack in an endogenous response, resulting in a reallocation of resources. In fact, the rate of investment in the portion of the private sector consisting of medium and large enterprises in India followed the pattern set by public investment; it stagnated during 1965-70 and improved slightly in the early 1970s, in sharp contrast to the private sector as a whole. 5 In any case, whether the decline in public sector investment and the consequent shift in the sectoral composition of investment affects aggregate investment and growth depends on many factors in specific situations, and no a priori conclusion seems possible without utilizing the framework of a complete model.

This paper provides a framework for a proper analysis of the growth impulse provided by public sector investment by emphasizing the transmission of the impulse to private investment and the dynamic process that underlies this transmission. 6 The macroeconomic relations emphasized in the model subsume the output and resource allocation effects at the industry level. Although these intersectoral shifts are of interest in their own right, especially in the formulation of a development plan, this model is focused on aggregate relationships in order to highlight the macroeconomic impact of public investment.

II. The Model

The model consists of functional relationships for private fixed investment, aggregate savings, output, and several definitional identities. Public investment is treated as exogenous, and any reaction function that may govern public investment decisions is ignored in order to focus on the public investment multipliers. The notation and the complete list of variables and equations constituting the model are shown at the end of this section.

private investment behavior

A private investment function has been derived by modifying the neoclassical theory of investment in order to incorporate some of the channels through which public investment influences private investment. The neoclassical theory suggests that private investment is positively related to the expected output level and negatively related to the relative price of capital—that is, the user cost of capital relative to the wage rate. The neoclassical model can be thought of as a combination of the traditional flexible accelerator model, which emphasizes the reaction of capital stock to output, and the neoclassical principle that an optimal set of inputs is dependent on their relative prices. The adjustment of capital stock to its desired level is assumed to occur with a lag, as in flexible accelerator models. The novel feature of this study is an attempt to show that private investment also depends on the capital stock of the public sector and the investible funds available to the private sector, and that these variables capture important channels of influence from public investment to private investment.

It is assumed that the private sector determines its desired level of capital by minimizing total cost TC, defined as the discounted present value of future costs including both the costs of production and the cost of acquiring capital. 7 The cost C of producing the planned private sector output QP* is a function not only of the planned output level but also of the plant size represented by private capital stock KP and of the available infrastructure represented by government capital stock KG. The acquisition cost of capital is primarily the value of net investment and replacement investment at current prices; 8 hence it is nothing but the value of gross fixed investment. Thus, the task is to minimize

where R(s) denotes the short-term interest rate, PI denotes the price of capital goods, δp denotes the rate of depreciation of private capital, and 0tR(s)ds denotes the long-term discount rate defined as the integral of short-term rates. The term within square brackets in equation (1) is the sum of the cost of production and the cost of acquiring additional capital. The “dot” above a variable, for example, KP˙ denotes its first derivative. Thus, KP˙ denotes net investment, and KP˙t+δpKPt denotes gross fixed investment. The Euler condition for minimization is given by

Applying this condition, we get

Completing the differentiation and rewriting, we get

This equation can be rewritten to express the first-order condition as

where Ut denotes the user cost of capital, which is given by

Equation (1’) states that capital should be acquired in the current period until the reduction in present and future costs, owing to a unit of additional capital, equals the current user cost of capital. The user cost of capital that is defined in equation (1”) is simply the product of the price of capital goods and the real rate of interest for investors, where the real interest rate is computed as the difference between the nominal short-term lending rate and the rate of increase in capital goods’ prices, plus the rate of depreciation. 9 It is interesting to note that although the long-term rate is used to discount future costs in defining the total cost, in defining the user cost of capital the appropriate interest rate is the short-term rate. 10

From the first-order condition for cost minimization, an expression for the desired capital stock of the private sector can be derived using a specific cost function. If it is assumed that the production function is Cobb-Douglas

where L denotes labor input, and A denotes the effects of shifts in the production function owing to technical change, then the variable cost function 11 can be expressed as

where W denotes the nominal wage rate. Differentiating this cost function with respect to KPt and substituting in the Euler condition, we get

Equation (3) can be rewritten to obtain the desired level of capital stock KP*t that corresponds to the expected or planned output level QP*t

An interesting implication of this equation is that if the private sector productive capacity is improved by an increase in public sector capital stock (α0 > 0), then, ceteris paribus, such an increase will reduce capital requirements of the private sector. Thus, in this formulation, public investment facilitates production in the private sector by lowering the cost of producing private sector output. In other words, public investment provides some of the facilities that the private sector would have to provide for itself in the absence of public investment. Hence, the private sector’s capital requirements are lowered by public investment. Equation (4) also implies that an increase in the rental-wage ratio reduces the desired capital stock owing to capital-labor substitution. However, it is the expected rental-wage ratio (Ut/Wt)* rather than the actual ratio, that governs investment decisions. 12

For purposes of estimation, the following linear approximation to equation (4) is used: 13

The private sector’s desired capital stock is a linear function of the expected rental-wage ratio, the planned level of private sector output, and the public sector capital stock. The determination of the planned output level and the mode of adjustment of the actual capital stock to the desired capital stock need to be specified in order to derive the final form of the investment function. The planned private sector output is assumed to be a function of the current and past levels of output, as well as of the public sector capital stock.

where a(L) is the lag operator. An increase in public sector capital stock raises private sector output expectations, because this increase represents potential additional demand for private sector products when these public investment projects mature. The effects of current demand for private sector output owing to current investment and production activities of the public sector is already subsumed in the private sector output variable in equation (5).

It is assumed that there is only a partial adjustment of the private sector’s actual capital stock to its desired level. 14

Equation (6) states that only a proportion of the gap between the desired capital stock and existing capital stock is closed in a given period.

The speed of adjustment, bt, is assumed to vary in response to the ease with which private investment can be financed. It is specified as

The variable influencing the speed of adjustment stands for the total financing available to the private sector in real terms (S - IG)/PI, relative to the required investment, KP*t — KPt - 1. 15 The financing available to the private sector is nothing but the difference between aggregate savings, S (including foreign savings) and public sector investment, IG; since this difference is merely gross private investment—in both fixed assets and inventories of the private sector—equation (7) also determines the allocation of private domestic investment between plant and equipment on the one hand, and inventories on the other.

Using equations (4’), (5), (6), and (7), and noting that private sector gross fixed investment IPt is

where δp denotes the rate of depreciation, we get the following investment function for the private sector.

The signs of the coefficients B2 and B4 are expected to be positive, while the signs of B1 and B5 are expected to be negative. The sign of B3 is, however, indeterminate. If it is positive, then the positive demand-inducement effect of public investment is larger than the opposite effect, owing to public investment aiding private sector production.

An explanation of the role of the resource availability variable, (S - IG)/PI, is in order. This variable captures important channels through which crowding out of private investment occurs in many developing countries. The view of crowding out represented by this variable is more general than the traditional view that refers to crowding out in the financial markets. It is more general, in that it takes into account both the crowding out that occurs through competition in the markets for real resources, such as cement, steel, and imported materials, as well as the nonprice rationing of financial and real resources. The crowding out can occur both through an increase in prices and interest rates following an increase in public investment and through some nonprice rationing mechanism such as licensing or other controls. The specification chosen here emphasizes the nonprice rationing aspects, since price and quantity controls are pervasive in developing countries. 16 Moreover, in many developing countries, including India and Korea, self-financed investment is important and therefore the availability of self-financing for acquiring fixed capital is a critical aspect. This, as well as the nonprice rationing aspect, is best taken into account by postulating a direct linkage between total resource availability and fixed investment. Thus, in this framework, inventory investment is determined as a residual.

production functions

Private sector output is assumed to be a function of the capital stock in the private and public sectors, and the rental-wage ratio.

This equation has been derived by rewriting equation (3) and linearizing. 17 This specification is consistent with the empirical observation that the possibility of substitution of capital for labor exists only for new capital that has yet to be installed, but the factor intensity implicit in the existing capital stock cannot be readily altered in response to changes in the relative price of capital. Thus, the capital-output ratio is assumed to be fixed ex post; but ex ante, it varies with the rental-wage ratio and the public sector capital stock. Output is determined by the available capital stock and by the currently feasible capital-output ratio. It is assumed that the labor supply is highly elastic. Thus, employment is assumed to be demand-determined—a valid assumption for developing economies with surplus labor.

The public sector output, QG, is assumed to be a linear function of public sector capital stock.

This formulation is valid for the determination of the output of the public enterprise sector; however, a more sophisticated formulation is required to explain the output of the general government. This complication is ignored here.

savings function

Real domestic savings, SD/P, are assumed to be a function of the real interest rate facing consumers and a distributed lag in income.

where D(L) is the lag operator on total output Q, and the real interest rate is computed as the difference between the short-term deposit rate i and the rate of increase in consumer prices, PC˙/PC. The specification of a distributed lag in output is consistent with several alternative models of savings behavior; the specific model best suited for an economy has to be determined on an empirical basis and can be inferred from the estimated lag distribution of the output variable. 18

definitional identities

The model is completed by including the relevant definitional identities. Gross domestic product, Q, is the sum of the output of the private and public sectors.

Aggregate savings is the sum of domestic and foreign savings SF.

Private and government capital stock can be obtained by the perpetual inventory method.

where δp denotes the rate of depreciation of the private capital stock, δG denotes the rate of depreciation of the public sector capital stock, and IGF denotes public gross fixed investment.

Total public sector investment IG is the sum of gross public fixed investment and public inventory investment.

where PIG denotes the deflator for public fixed investment and IGI denotes the public sector’s inventory investment. Equation (17) is necessary for simulating the model to study the effect of an increase in real gross fixed investment of the public sector.

the complete model

The behavioral equations for private investment (9), private and public sector output (10 and 11), and domestic savings (12), together with the definitional identities (13-17), constitute a complete model of savings, investment, and growth. For the sake of convenience, the model and the list of variables are shown in Table 3. The identity equating savings and investment is not explicitly included, because private investment in inventories, which has been left unspecified, is determined residually, ensuring that the identity holds.

Table 3.A Model of Investment, Savings, and Growth1
Behavioral Equations
Private investment
IP=B0+B1(UW)*+B2a(L)QP+B3KG+B4(SIGPI)+B5KPt1
Private sector output
QP=C0+C1KP+C2KG+C3(UW)*
Public sector output
QP=C0+C1KG
Domestic savings
SD/P=D0+D(L)Q+D1(iPC˙/PC)+D2(SD/P)t1
Definitional Identities
Total GDP
Q=QG+QP
Total savings
S=(SD/P)*P+SF
Private capital stock
KP=(1δP)KPt1+IP
Government capital stock
KP=(1δg)KPT1+IGF
Government investment
IG=IGF*PIG+IGI
Endogenous variables (in alphabetical order):
IP=real gross fixed investment by the private sector
KP=private sector capital stock
Q=real GDP
QG=public sector GDP
QP=private sector GDP
S=nominal aggregate savings
SD/P=real domestic savings
Exogenous variables (in alphabetical order):
i=short-term deposit rate
IG=public sector investment at current prices
IGF=real gross fixed investment by the public sector
IGI=inventory investment by the public sector at current prices
KG=public sector capital stock
P=GDP deflator
PI=deflator for private fixed investment
PIG=deflator for public fixed investment
PC/PC=rate of increase in consumer prices
R=short-term lending rate
SF=foreign savings at current prices
U=user cost of capital defined as PI (R + δ – PI/PI)
W=nominal wage rate

The superscript * that appears with the rental-wage ratio UW denotes the expected value of the ratio. Unless otherwise mentioned, all variables are at constant prices; a(L) and D(L) denote lag distributions.

The superscript * that appears with the rental-wage ratio UW denotes the expected value of the ratio. Unless otherwise mentioned, all variables are at constant prices; a(L) and D(L) denote lag distributions.

The working of the model can be illustrated by considering the effects of an increase in public investment, ceteris paribus. 19 An initial one-shot increase in real fixed investment by the public sector raises public sector output, the private sector’s actual and expected output, and aggregate domestic savings, while simultaneously absorbing part of the savings to finance the increased public investment. If there is a negative effect owing to a net reduction in the availability of savings to the private sector (crowding out) that more than offsets the positive effects of increased private sector output and output expectations, then private fixed investment falls; if not, private investment rises. The resulting changes in the capital stock of the private sector, together with the desired capital stock in the next period, determine the level of private investment in the next period. Similarly, changes in savings in the initial period generate adjustments in savings in subsequent periods. In this dynamic framework, the effects of public investment on growth depend critically on the differences in the marginal productivity of capital in the public and private sectors. Assuming that there is a full or partial crowding out of private investment in the initial period, the change in total output will depend on whether the increased public sector output owing to the increase in public investment exceeds or falls short of the reduction in private output owing to a reduction in private investment. If there is no crowding out, the effect on total output is clearly positive.

III. Empirical Results

data

Data for India on GDP, savings, investment, wages, and price deflators for investment are from various issues of National Accounts Statistics, which is published by India’s Central Statistical Organization, and Lal (1977). Data on interest rates and consumer prices are from various issues of the Reserve Bank of India’s Bulletin. The wage rates are derived by dividing wages paid in the organized private sector by the employment in that sector. Employment data are from various issues of the Economic Survey, which is published by the Government of India. The rate on advances by the State Bank of India (a prime rate) is used as the representative lending rate, and the commercial banks’ one-year time deposit rate is used as the representative deposit rate. Data for Korea are from various issues of the Economic Statistics Yearbook published by the Bank of Korea. Wage data are for the industrial sector. The lending rate is the rate on advances for one year, and the deposit rate is the three-month time deposit rate.

The data on capital stock in the private and public sectors are constructed by cumulatively adding the time series on real net fixed investment (at 1960/61 prices for India, at 1970 prices for Korea). Since linear investment functions are used, the unknown initial capital stock can readily be absorbed into the intercept term, thus obviating the need for estimating the initial capital stock. 20

model adaptations

The rate of depreciation required to compute the rental price of capital is obtained by regressing depreciation at constant prices on the initial capital stock. Separate functions for the depreciation of private and public sector capital stock, DP and DG, are estimated.

Since depreciation is thus separately specified, we can redefine capital stock as

These equations, instead of equations (15) and (16), are used in model simulations.

The model specification was slightly altered for Korea because of the lack of separate GDP series for the private sector. Since it is assumed that public sector output is a linear function of the public sector capital stock, private sector output could be derived as the difference between total GDP and a linear function of the public sector capital stock. However, instead of separately estimating private sector output in this way, we use total GDP as an explanatory variable in the private investment function; the public sector capital stock in the private investment function, therefore, reflects the implicit subtraction of public sector output from total output. This modification affects the interpretation of the coefficient of public sector capital stock in the private investment function for Korea. This coefficient reflects not only the effects of public sector capital stock on private costs and private output expectations but also the negative effect owing to the implicit subtraction of public sector output. Therefore the coefficient would be smaller than it would have been if only private sector output, instead of total output, had been used as an explanatory variable. Moreover, for Korea, a production function relating total GDP to public and private capital stocks and the rental-wage ratio is used. The coefficient of the public sector capital stock in this function also reflects not only its effects on private sector output but also its direct effect on public sector output. However, for India, separate production functions are estimated for the output of the private and public sectors.

discussion of the estimates

The estimated behavioral equations are presented in Table 4 for India and in Table 5 for Korea. The regression coefficients have been estimated by the ordinary least-squares method.

Table 4.India: Estimated Behavioral Equations1(Period of fit 1960-76)
Private investment
IP=25.9932(2.67)+0.1854(3.35)QP+0.1398(3.35)QPt1+0.6359(3.02)(SIGPI)49.5378(1.82)(UW)t1168.551(2.74)(UW)t20.1752(2.65)KPt1+0.0430(0.62)KGADJ.R2=0.9526SEE=0.97DQ=2.4(20)
Private sector output
QP=106.593(31.54)+0.4006(26.53)KP+17.7163(0.40)(UW)t1+174.935(3.2)(UW)t27.3321(5.85)DROUGHT2ADJ.R2=0.9884SEE=2.24DQ=2.88(21)
QP=107.073(26.85)+0.4417(2.82)KP+17.2149(0.34)(UW)t1+173.642(3.00)(UW)t20.0367(0.26)KG7.3116(5.55)DROUGHTADJ.R2=0.9870SEE=2.35DW=2.8(21)
Public sector output
QG=6.7392(4.43)+0.1307(14.53)KGADJ.R2=0.9292SEE=2.35DW=0.833(22)
Gross domestic savings
SDP=12.7288(2.23)+0.1484(2.42)Q+0.5969(2.97)(SDP)t1+15.857(2.14)(i+PC/PC1+PC/PC)ADJ.R2=0.9625SEE=2.06DW=1.69(23)
Depreciation of private sector capital
DP=4.5870(23.72)+0.0435(29.51)KPt1ADJ.R2=0.9819SEE=0.3122DW=0.643(24)
Depreciation of public sector capital
DG=0.7898(4.55)+0.0107(9.71)KGt1ADJ.R2=0.8535SEE=0.28DW=0.94(25)

Figures in parentheses below the coefficients are t-statistics. ADJ.R2 denotes the adjusted R2; SEE denotes the standard error of estimate; and D-W denotes the Durbin-Watson statistic.

The variable DROUGHT is a dummy variable that is unity in years of deficient rainfall and is zero in all other years.

Figures in parentheses below the coefficients are t-statistics. ADJ.R2 denotes the adjusted R2; SEE denotes the standard error of estimate; and D-W denotes the Durbin-Watson statistic.

The variable DROUGHT is a dummy variable that is unity in years of deficient rainfall and is zero in all other years.

Private investment and output

The neoclassical framework clearly provides a good fit for output and private investment in both countries. The coefficients of determination of the estimated equations (20), (21), and (21’) of Table 4 and equations (26), (26’), and (27) of Table 5 are high; and all the explanatory variables, except the public sector capital stock in the investment function, are statistically significant and have the expected signs. The major determinants of output are the capital stock and the rental-wage ratio (equations (21), (21’), and (22) of Table 4 and equation (27) of Table 5). The rental-wage ratio, which measures the relative price of capital and labor, has a significant impact on capital-output ratios in both countries. In India, variations in weather conditions, which are measured by a dummy variable, also have a significant effect on output. Moreover, a comparison of the coefficients of capital stocks in the production functions of both countries reveals that the incremental capital-output ratios for Korea, of both the private and public sectors, are considerably smaller than those for India. Private investment is significantly influenced by output (which is a proxy for output expectations), the resources available to the private sector, and the initial private sector capital stock. The rental-wage ratio has a strong negative influence on investment in India, but has no significant influence in Korea.

Table 5.Korea: Estimated Behavioral Equations1(Period of fit 1958-76)
Private investment
IP=320.961+0.3443Q+0.2590(SIGPI)0.16757KPt10.1943KG(4.80)(5.0)(3.36)(2.44)(0.11)ADJ.R2=0.9946SEE=21.03DW=1.67(26)
IP=363.08+0.4192Q+0.2368(SIGPI)0.1847KPt10.1308KG22.730(UW)(4.64)(4.19)(2.97)(2.62)(0.64)(1.04)ADJ.R2=0.9947SEE=20.90DW=1.87(26)
Total output
Q=792.068+0.4362KP+1.0648KG+158.84(UW)(12.63)(2.25)(2.54)(4.18)ADJ.R2=0.9979SEE=53.83DW=1.97(27)
Gross domestic savings
SD/P=79.7483+0.9153Q0.5446Qt10.2443Qt2+72.821(iPC˙/PC1+PC˙/PC)(4.35)(10.35)(4.92)(2.36)(0.73)ADJ.R2=0.9931SEE=24.62DW=1.31(28)
Depreciation of private sector capital
DP=29.9217+0.0996KPt1(7.64)(38.67)ADJ.R2=0.9874SEE=12.70DW=1.68(29)
Depreciation of public sector capital
DP=5.0873+0.0103KGt1(9.10)(13.81)ADJ.R2=0.9100SEE=1.75DW=1.12(30)

Figures in parentheses below the coefficients are t-statistics. ADJ.R2 denotes the adjusted R2; SEE denotes the standard error of estimate; and D-W denotes the Durbin-Watson statistic.

Figures in parentheses below the coefficients are t-statistics. ADJ.R2 denotes the adjusted R2; SEE denotes the standard error of estimate; and D-W denotes the Durbin-Watson statistic.

The immediate influence of government investment in crowding out private investment is measured by the coefficient of the variable, S — IG / PI. As explained earlier, this variable measures the resources available to the private sector and is obtained as the difference in real terms between aggregate savings and total public sector investment. The coefficient of this variable measures the effect of resource availability on the speed of adjustment of the actual capital stock to the desired level of capital stock. A comparison of the coefficients for India (0.63) and Korea (0.24) shows that resource availability had a much stronger influence on the speed of adjustment in India than in Korea. In other words, a larger proportion of an increase in the resources available to the private sector finds its way into fixed capital formation in India than in Korea. This has the implication that the immediate crowding-out effect of public sector investment is likely to be much stronger in India than in Korea. However, the existence of the net crowding-out effect cannot be determined on the basis of this coefficient alone, since public sector investment also has a positive effect on output, output expectations, and savings, and these effects may offset the immediate crowding-out effect operating through the resource constraint variable. Therefore, a complete analysis of the effect of public sector investment on private investment requires the computation of impact and dynamic multipliers of public sector investment. These multiplier effects are discussed in the next section.

A test of the neoclassical framework is given by the statistical significance of the rental-wage ratio. 21 The rental-wage ratio in India has a significantly negative substitution effect on investment, as well as a significantly positive efficiency effect on output; in Korea, the negative substitution effect is weak, while the positive efficiency effect on output is quite strong. Moreover, the relative price effects occurred with a time lag of up to two years in India, but more quickly in Korea. Thus, the evidence suggests that an increase in the interest rate—and hence the rental price of capital—tends to depress investment demand significantly in India (with a lag of up to two years), reflecting the substitution of capital for labor. This substitution effect on investment is negligible for Korea. 22 However, there are theoretical grounds to believe that the rental-wage ratio can have a significant effect on output without having an immediate direct effect on the level of investment. The significant coefficient for the rental-wage ratio in the production functions of both countries can be interpreted as supporting the hypothesis that an increase in interest rates— and hence the rental price of capital—increases the overall efficiency of capital by permitting a shift of resources to more productive sectors and by encouraging more productive use of capital within each sector. 23 This positive effect on output stimulates investment demand and counteracts the negative substitution effect discussed earlier.

An examination of the size of the coefficient of the initial capital stock provides interesting insights. This coefficient, (δp, — b0), is the difference between the rate of depreciation and the speed of adjustment of the capital stock. Although the rate of depreciation of the private capital stock in Korea (0.1) was much higher than in India (0.04), the coefficient of the initial capital stock was almost identical in the two countries (-0.18). Therefore, it follows that the basic speed of adjustment, b0, of the private capital stock to the desired level was faster in Korea than in India. This is consistent with the observation that the speed of adjustment in India is conditioned by the industrial policy environment, which has been less flexible toward the private sector in India than in Korea.

Consistent with the observation that the reactions to relative prices and adjustment of capital stock to the desired level are slower in India than in Korea, the impact of a change in output on investment is also spread out over a longer time period in India than in Korea. A unit increase in private sector output raises private investment by 0.4 units in Korea and by 0.3 units in India; while the increase occurs within the same year in Korea, the increase is spread out over two years in India (0.2 in the first year, and 0.1 in the second).

The public sector capital stock has a negative coefficient in the investment function for Korea and a positive coefficient in the investment function for India. Although the coefficients are not statistically significant, partly because the positive and negative effects offset each other, they were retained for the purpose of simulations. The negative coefficient for Korea merely reflects the implicit subtraction of public sector output that was discussed earlier. Given the fairly large coefficient of public sector capital stock in the output function for Korea (equation (27), Table 5), the fairly small coefficient in the investment function implies a strong effect of public sector capital stock on the private sector’s output expectations. In contrast, the positive coefficient of the public sector capital stock in the investment functions for India shows that the public sector capital stock has some impact on the private sector’s output expectations, although its impact on output is negligible (equation (21’), Table 4).

Savings

The estimated aggregate savings functions provide a good fit of actual data (equation (23) of Table 4 and equation (28) of Table 5). The estimates indicate significant differences in domestic savings behavior in the two countries. In India, the real interest rate had a significantly positive direct effect on total savings; while in Korea, the direct effect of the real interest rate, although positive, was not significant. 24 Although an increase in the interest rate does not have a significant direct influence on the average propensity to save in Korea, it does have a strong and significantly positive impact multiplier effect. 25 This is so because the higher interest rate serves to raise the efficiency of capital and thereby stimulate economic growth, which, in turn, stimulates savings. Thus, a closer examination of the results makes it clear that they are not contrary to the results of the earlier studies, which found a significant positive effect of interest rates on Korean savings. Indeed, the results of this paper throw new light on the channels through which interest rates influence savings, an aspect ignored in earlier studies. The indirect effects of interest rates on savings, operating via the efficiency effects on output, appear more important in Korea than the direct effects. Both the direct and indirect effects seem important in India.

Furthermore, an examination of the lag distributions of income suggests that, in India, savings are related to permanent income, while in Korea, savings are related to permanent income as well as transitory income. This difference in savings behavior can be demonstrated as follows. In India, because of the simultaneous presence of lagged savings and current income in the savings equation, the lag distribution of income is geometric. Therefore, by identifying permanent income as the weighted sum of current and past incomes with geometrically declining weights, the estimated specification for India can be interpreted as a relationship between savings and the level of permanent income. The estimates imply that the marginal propensity to save in India is 0.15 in the short run and 0.37 in the long run. In Korea, the lag distribution of income has a positive coefficient for current income and negative coefficients for one-period and two-period lagged incomes. Moreover, the sum of the coefficients is positive. This form of the lagged distribution is consistent with the hypothesis that savings are related to permanent income YPt and transitory income (Yt — YPt) with different marginal propensities to save out of the two types of income. Defining permanent income, YPt, as a weighted sum of current and past incomes,

and transitory income, YTt, as the deviation of current income from permanent income,

the aggregate domestic savings can be specified as

On substituting the definitions of permanent and transitory incomes in the savings function, we get the following lag distribution for the income variable:

Income VariableYtYt - 1Yt - 2
Coefficientsβ1α2 + β2(1 - α11 - β221 - β23

The sum of the coefficients is β1, which is the marginal propensity to save out of permanent income. While the other coefficients are not identifiable, it is clear that the lag distribution should have a large positive coefficient in the first period and negative coefficients in subsequent periods, with the size of the coefficients declining over time. This is observed for Korea, for which the marginal propensity to save out of permanent income (β1) is estimated at 0.13. 26 The marginal propensity to save out of transitory income is large in Korea, which is consistent with the large negative coefficients for the lagged income variables.

IV. Model Simulations

Static simulations, based on historical values of the lagged variables, showed that the goodness of fit of the model as a whole is excellent for both India and Korea. The model was also dynamically simulated (i.e., lagged variables are those generated by the model itself); the actual and dynamically simulated values of private investment, GDP, and savings are shown in Charts 3 and 4. Dynamic simulations indicate that the model is stable; and the goodness of fit, even in dynamic simulations where errors can cumulate, is indicative of the robustness of the model. These dynamic simulation results also provide a basis for deriving the dynamic effects of changes in public investment.

Chart 3.India: Dynamic Simulations, 1962/63-1976/77 1

(In billions of 1960/61 rupees)

1 Indian fiscal years extend from April 1 to March 31.

Chart 4.Korea: Dynamic Simulations, 1960-76

(In billions of 1970 won)

To derive these effects, the model is dynamically simulated by raising public fixed investment (by Rs 1 billion at 1960/61 prices for India, and by W 1 billion at 1970 prices for Korea) in one period (1962/63 for India and 1960 for Korea), while retaining it at its historical level in all subsequent periods. 27 The results of this one-shot increase in public investment are compared with the original dynamic simulations, and the implied multipliers are derived and shown for India (Table 6) and Korea (Table 7). The dynamic multipliers are compared graphically in Charts 5 and 6.

Table 6.India: Impact and Dynamic Multipliers of Public Sector Fixed Investment1(In billions of 1960/61 rupees)
Private Fixed InvestmentPrivate Sector Real GDPPublic Sector Real GDPTotal GDPReal Domestic Savings
TimeIPt+1IGtQPt+1IGtQGt+1IGtQt+1IGt(Q)S/Pt+1IGt
0–0.657–0.2630.131–0.132–0.020
10.062–0.2270.129–0.098–0.026
20.056–0.1950.128–0.067–0.023
30.055–0.1640.126–0.038–0.021
40.055–0.1350.125–0.010–0.014
50.057–0.1070.1240.017–0.006
60.060–0.0780.1220.0440.003
70.063–0.0490.1210.0720.013
80.065–0.0210.1200.0990.022
90.0680.0070.1190.1260.032
100.0720.0360.1170.1530.042
110.0770.0650.1160.1810.052
120.0780.0940.1150.2090.062
Long-run multiplier0.111–1.0371.5930.5560.012

Dynamic effects of an increase of Rs 1 billion in public fixed investment (at 1960/61 prices) in 1962/63. The long-run multiplier is a total for more than 12 periods and, hence, may not equal the sum of the coefficients shown. Indian fiscal years extend from April 1 to March 31.

Dynamic effects of an increase of Rs 1 billion in public fixed investment (at 1960/61 prices) in 1962/63. The long-run multiplier is a total for more than 12 periods and, hence, may not equal the sum of the coefficients shown. Indian fiscal years extend from April 1 to March 31.

Chart 5.India: Dynamic Multipliers of Increase of Rs 1 Billion in Public Sector Fixed Investment, 1962/63-1976/771

(In billions of 1960/61 rupees)

1 One-shot increase in 1962/63. Public investment returns to its historical value from 1963/64 onward. Indian fiscal years extend from April 1 to March 31.

Chart 6.Korea: Dynamic Multipliers of Increase of W 1 in Public Sector Fixed Investment, 1960-761

(In billions of 1970 won)

1 One-shot increase in 1960. Public investment returns to its historical value from 1961 onward.

dynamic impact on private investment

For India, an increase in public investment by Rs 1 billion initially leads to a net reduction in the resources available to the private sector and causes a decline in private investment of Rs 0.6 billion in the immediate period, but the additional resources generated by the public investment stimulate private investment in all subsequent periods. However, the positive effects in the subsequent periods are so small, relative to the initial negative effect, that it takes nearly a decade for the long-run multiplier effect to become nonnegative. However, the immediate crowding-out effect, though strong, is only partial; it reflects the positive effects of public investment on output and savings that supplement total available resources and partly offsets the negative effects of reduced resource availability to the private sector. Thus, increased public investment does serve to raise overall investment.

For Korea, the multiplier effects on private investment are positive in the immediate period as well as in all subsequent periods. However, reflecting the competition for resources when public investment is raised, the impact multiplier is smaller than the interim multipliers. This is in sharp contrast to the negative impact multiplier observed for India and reflects the much stronger favorable effects of public investment on private sector expectations and on GDP in Korea than in India and the much weaker effects of resource availability on the speed of adjustment than in India.

impact on output and savings

The decline in private sector output in India resulting from reduced private investment is larger than the increase in output resulting from added public investment; this reflects the much larger incremental capital-output ratio in the public sector than in the private sector. Thus, in the first five years following a step-up in public investment, real GDP is lower than it would have been in the absence of additional public investment. Thereafter, real GDP is higher; while public sector output increases, private sector output recovers from the initial crowding-out effects. However, multiplier effects on output for Korea are large because of the absence of a net crowding-out effect. The multiplier effects on domestic savings are weak for India, but they are strong for Korea; this pattern reflects the weak effects of public investment on output in India and the relative strength of these effects in Korea.

The general conclusion that emerges from Table 6 is that the long-run multiplier effects of increased public investment in India are weak. The effect on overall investment is weak, because the negative crowding-out effect is stronger than the positive effects. The effect on overall GDP is weak, because the incremental capital-output ratio in the public sector is considerably larger than in the private sector; and as a result, the increase in public sector output is not adequate to offset the loss in private sector output for a considerable length of time. The results for Korea (Table 7) are strikingly different. The long-run multiplier effects of increased public investment are strong. This is because the negative crowding-out effect is weaker than the positive effects, while the difference between the capital-output ratios of the public and private sectors is not a significant factor.

Table 7.Korea: Impact and Dynamic Multipliers of Public Sector Fixed Investment1(In billions of 1970 won)
Private Gross Fixed InvestmentGross Domestic ProductGross Domestic Savings
TimeIPt+1IGFtQt+1IGFt(SD/P)t+1IGFt
00.2911.1911.090
10.5121.3910.624
20.4401.5390.360
30.4381.6700.350
40.4261.7820.345
50.4071.8740.336
60.3941.9490.328
70.3902.0150.324
80.3832.0700.321
90.3832.1180.319
100.3742.1570.315
110.3712.1900.312
120.3672.2170.309
Long-run multiplier5.17624.1635.333

Dynamic effects of an increase of W 1 billion in public fixed investment (at 1970 prices) in 1960.

Dynamic effects of an increase of W 1 billion in public fixed investment (at 1970 prices) in 1960.

V. Summary and Conclusions

A dynamic model of public investment, private investment, savings, and growth is postulated and applied to India and Korea. The model is designed to highlight the impact of public investment on private investment and growth by incorporating the various channels of influence from public investment to private investment. Public investment exerts a short-term crowding-out effect on private investment; but it also raises the productivity of private capital stock and, by creating demand for the output of the private sector, raises the output expectations and investment requirements of the private sector. It also raises aggregate output and savings, thereby offsetting part of the initial crowding-out effects. The model takes into account these various channels of influence to investigate the crucial issue of the magnitude and the dynamic time path of the effect of public investment on private investment and growth in India and Korea.

The modified neoclassical framework for explaining private investment provides an excellent fit for output and investment in both countries. The immediate crowding-out effect of public investment, which operates through constraining the availability of resources to the private sector and, in turn, lowering the speed of adjustment of the private sector capital stock, is found to be much larger in India than in Korea. However, the normal speed of adjustment of the capital stock in India is slower than in Korea. The response of investment to changes in output is strong in both countries, but it is slower in India than in Korea. The relative cost of capital is found to have a strong positive efficiency effect on capital in Korea but only a weak negative substitution effect on investment; in India, both effects are found to be strong. Domestic savings behavior in the two countries is found to differ significantly. The short-run marginal propensity to save out of income is much higher in Korea than India, while the long-run marginal propensity to save is much lower. In India, the real rate of interest has a strong positive direct effect on the average propensity to save; in Korea, this direct effect is relatively weak.

The dynamic simulation of the model after a one-shot increase in public investment reveals sharply divergent patterns of response of private investment and output in the two countries. In India, there is substantial crowding out in the initial period, but private investment is stimulated in all subsequent periods. These latter effects, however, are weak in each period, so that the initial negative effect is not offset for a considerable period. However, the crowding out is only partial, so that public investment does raise total investment. The effect on aggregate output is also negative in India, reflecting the much larger incremental capital-output ratio in the public sector than in the private sector. In sharp contrast, the effects of public investment on private investment are positive and large, both in the immediate and subsequent periods, in Korea, reflecting its strong positive effect on aggregate output and output expectations of the private sector.

The model incorporates the effects of interest rate changes that operate through their influence on the cost of capital as well as through the real interest rate for savings. These variables, in turn, influence the desired level of the capital stock and its productivity, as well as the availability of savings (and, thereby, the speed of adjustment of the actual capital stock to the desired level). The empirical results indicate that the direct effects of interest rate changes on investment and savings are dissimilar in the two countries. An increase in the interest rate in India has a significant negative substitution effect on investment and a significant positive effect on savings; while in Korea, these direct effects are relatively weak. However, the effects on capital productivity are strong in both countries. While this efficiency effect, which raises the growth of output, yields strong multiplier effects of interest rates in Korea, the multiplier effects are somewhat weakened in India because higher interest rates also dampen investment demand and output growth, counteracting the efficiency effects on output. Thus, the model highlights the channels through which interest rate effects are manifested and makes it possible to test various propositions of the classical and modern theories of interest rates. This, though, is a subject for a separate study.

REFERENCES

    Bank of KoreaEconomic Statistics Yearbook 1978 (Seoul 1978).

    BrownGilbert T.Korean Pricing Policies and Economic Development in the 1960s (Johns Hopkins University Press1973).

    ClarkPeter K.“Investment in the 1970s: Theory, Performance, and Prediction,”Brookings Papers on Economic Activity: 1 (1979) pp. 73-124.

    CoenRobert M.“The Effect of Cash Flow on the Speed of Adjustment,” in Tax Incentives and Capital Spending: Papers Presented at a Conference of Experts Held on November 3 1967ed. byGaryFromm (The Brookings Institution1971) pp. 131-96.

    • Search Google Scholar
    • Export Citation

    EisnerRobert and Robert H.Strotz“Determinants of Business Investment,” in Daniel B.Suits and othersImpacts of Monetary Policy: A Series of Research Studies Prepared for the Commission on Money and Credit (Englewood Cliffs, New Jersey1963) pp. 60-233.

    • Search Google Scholar
    • Export Citation

    HallR. E.“Investment, Interest Rates, and the Effects of Stabilization Policies,”Brookings Papers on Economic Activity:1 (1977) pp. 61-103.

    India Central Statistical OrganizationNational Accounts Statistics (New Delhi) various issues.

    India Government of Economic Survey (Delhi) various issues.

    India Reserve Bank of Bulletin (Bombay) various issues.

    IoannidesYannis M. and Michael C.Caramanis“Capital-Labor Substitution in a Developing Country: The Case of Greece,”European Economic ReviewVol. 12 (February1979) pp. 101-10.

    • Crossref
    • Search Google Scholar
    • Export Citation

    JonesLeroy P.Public Enterprise and Economic Development: The Korean Case (Korea Development InstituteSeoul1975).

    JorgensonDale W. (1963) “Capital Theory and Investment Behavior,”American Economic ReviewVol. 53 (May1963) pp. 247-59.

    JorgensonDale W. (1967) “The Theory of Investment Behavior,” in Determinants of Investment Behavior: A Conference of the Universities-National Bureau Committee for Economic Researched. byRobertFerberNational Bureau of Economic Research (New York1967) pp. 129-55.

    • Search Google Scholar
    • Export Citation

    JorgensonDale W. (1971) “Econometric Studies of Investment Behavior: A Survey,”Journal of Economic LiteratureVol. 9 (December1971) pp. 1111-47.

    • Search Google Scholar
    • Export Citation

    KrishnamurthyKuruganti“An Econometric Model of India, 1948-61” (unpublished doctoral dissertationUniversity of Pennsylvania1963).

    • Search Google Scholar
    • Export Citation

    LalRam N.Capital Formation and Its Financing in India (Bombay1977).

    LeffNathaniel H. and KazuoSato“A Simultaneous-Equations Model of Savings in Developing Countries,”Journal of Political EconomyVol. 83 (December1975) pp. 1217-28.

    • Crossref
    • Search Google Scholar
    • Export Citation

    McLarenK. R.“Equipment Investment: An Alternative Approach” in Three Studies of Private Fixed Investmented. byW. E.NortonReserve Bank of Australia Occasional Paper No. 3E (1971).

    • Search Google Scholar
    • Export Citation

    Pinell-SilesArmandoDeterminants of Private Industrial Investment in India,World Bank Staff Working Paper No. 333 (1979).

    RothschildMichael“On the Cost of Adjustment,”Quarterly Journal of EconomicsVol. 85 (November1971) pp. 605-22.

    RowleyJ. C. R. and P. K.Trivedieds.Econometrics of Investment (New York1975).

    SrinivasanT. N. and N. S. S.Narayana“Economic Performance Since the Third Plan and Its Implications for Policy,”Economic and Political WeeklyVol. 12 (February1977).

    • Search Google Scholar
    • Export Citation

    ThomasD. E. L.“Estimates of the Capital Stock of Welsh Manufacturing, 1949-70,” in The Measurement of Capital: Theory and Practiceed. byK. D.Patterson and KerrySchott (London1979) pp. 159-97.

    • Crossref
    • Search Google Scholar
    • Export Citation

    von FurstenbergGeorge M. and Burton G.Malkiel“The Government and Capital Formation: A Survey of Recent Issues,”Journal of Economic LiteratureVol. 15 (September1977) pp. 835-78.

    • Search Google Scholar
    • Export Citation

    WaiU Tun and Chorng-HueyWong“Determinants of Private Investment in Developing Countries” (unpublishedInternational Monetary FundMarch61979).

    • Search Google Scholar
    • Export Citation

    WilliamsonJeffrey G.“Why Do Koreans Save ‘So Little’?”Journal of Development EconomicsVol. 6 (September1979) pp. 343-62.

SUMMARIES

Domestic Determinants of Net U.S. Foreign Investment—george m. von furstenberg (pages 637-78)

The current account balance known as net foreign investment in the national income and product accounts of the United States is largely determined by domestic factors that have changed the supply of national saving relative to the demand for domestic investment. Thus, the estimated rate of net foreign investment declined by almost half as much as the cyclically adjusted net national saving rate on average over the past two decades and this proportion has been growing. This is established in a quarterly simultaneous-equation model in which the predicted rate of net foreign investment is derived as the balance of the sum of the estimated government, personal, and corporate net saving rates on the one hand and the sum of the estimated domestic fixed and inventory investment rates on the other. Cyclical variables that appear in these equations and allow them to be solved at equilibrium are themselves endogenous. The evidence suggests that measures to raise the national saving rate would contribute as much to the external adjustment of the United States as to its domestic capital stock and thence to its potential gross domestic product. Thus, the openness of the U. S. economy cannot safely be ignored in policy discussions of saving and investment in the United States.

Exchange Rates, Inflation, and Vicious Circles—marian e. bond (pages 679-711)

In order to assess how and why vicious circles can occur and to consider whether certain countries are more prone to them than others, a theoretical framework for analyzing these questions and for discussing the relevant empirical evidence is provided. A general equilibrium framework is used to expand on the theory of the origins of vicious circles and to analyze adjustment from one steady state to another. The origin of vicious circles can be identified as a phase of adjustment in which wage and price inflation and exchange rate depreciation all occur together. In the absence of policies that will accommodate wage and price inflation, this phase will be followed by a period during which wage and price inflation abates as the economy adjusts to its new steady state. The effect of monetary policy on this adjustment process is examined, and an assessment made of whether a shock in the foreign exchange market can cause vicious circles and whether small, open economies are more vulnerable to vicious circles than larger, relatively closed economies.

The empirical evidence focuses on identifying features of individual countries that are responsible for vicious circles as adjustment takes place in foreign exchange markets, and as exchange rate changes are transmitted to the domestic economy. Three types of domestic policies that have an impact on the adjustment process are considered, and it is concluded that the difference between vicious and virtuous circle experiences can primarily be found in the difference in policies of the monetary authorities.

Impact of Inflation on Fiscal Policy in Developing Countries—peter s. heller (pages 712-48)

This paper provides a theoretical and empirical analysis of the factors underlying the relative speed of adjustment of different types of revenue and expenditure to inflation. While there is considerable variability in the relative adjustment rates of total expenditure and revenue for the 24 developing countries under analysis, expenditure tends to adjust more rapidly in 60 per cent of the countries. For any given country, the fiscal response to inflation appears to vary according to the stage of the inflationary process. As the inflation rate rises, there is an accelerated response by expenditure and a lagged response by revenue; once inflation stabilizes at a higher steady-state level, the expenditure and revenue adjustment coefficients then converge to lower levels than in the low inflation rate period. Between countries, there is also evidence that countries experiencing high mean inflation rates have higher overall adjustment coefficients, with expenditure tending to adjust more rapidly than revenue. These fiscal adjustment coefficients also appear fairly insensitive to the rate of acceleration or deceleration in the inflation rate. Finally, the results suggest that expenditures adjust more rapidly than revenues to anticipated inflation; conversely, the opposite result emerges with respect to unanticipated inflation.

Texts on the adjustment properties of individual components of expenditure to inflation suggest that wages and capital expenditure adjust much less rapidly than purchases of other goods and services. Interest payments, capital transfers, and net lending adjust rapidly. Among revenue items, corporate taxes adjust more rapidly than personal income taxes, and, at high inflation rates, domestic sales tax revenues adjust more rapidly than income taxes. In summary, the impact of inflation on the net fiscal position of the public sector is not a priori predictable, being shaped by the discretionary response of budget decision makers, by the country’s particular structure of revenue and expenditure, and by the character of the inflationary environment. If inflation leads to a higher budgetary deficit, it is, in effect, not wholly unplanned.

A General Equilibrium Approach to the Analysis of Trade Restrictions, with an Application to Argentinaandrew feltenstein (pages 749-84)

A disaggregated general equilibrium model is constructed of a small, open economy with ad valorem taxes on domestic production and tariffs on imports. Domestic currency and foreign exchange are included, which allows the existence of current and capital accounts and a general price level. The existence of an equilibrium for the model is demonstrated, and a computational method of solution is developed. The model is implemented with Argentine data and is found to yield a good approximation of the actual outcomes of the fourth quarter of 1978. The model is simulated after assuming a 50 per cent reduction in tariffs, and the effectiveness of monetary and exchange rate policies in stabilizing the balance of payments against the impact of the trade liberalization is examined. It is found that the changes in both domestic credit expansion and the nominal exchange rate that are required if the balance of payments is to be completely neutralized are greater than might be expected. The total government revenue that results from each of the policies considered is estimated, and, via the calculation of utility indices, the welfare implications of these policies are considered. The stability of the model is examined by allowing the simulation to continue for two periods after the initial tariff liberalization, and it is found that the initial shock to the balance of payments damps out rapidly, even in the absence of government policy actions.

Monetary, Financial, and Fiscal Policy Under Rational Expectations—willem h. buiter (pages 785-813)

The paper evaluates the implications for the conduct of monetary, fiscal, and financial policy of the rational expectations revolution in macroeconomics. It concludes that some important policy conclusions derived from conventional eclectic neo-Keynesian models remain valid when rational expectations are introduced: to anticipate policy is not to neutralize it. Structural policies aimed at altering the level and composition of full employment output in the short run and in the long run and stabilization policies aimed at influencing the cyclical departures of output and employment from their full employment levels are considered.

For a given level and composition of real government spending on goods and services, the substitution of bond financing for current (lump-sum) taxes reduces saving in the short run and lowers the capital-labor ratio in the long run. This crowding out persists even if each economic agent allows fully for the future taxes required to service the stock of privately held interest-bearing public debt. It also holds, subject to minor qualifications, when allowance is made for the possibility of bequests and other private intergenerational gifts. The substitution of money financing for tax financing is also not a matter of indifference. The inflation tax is not equivalent to explicit current taxes. In the simple models of the paper, the substitution of money financing for tax financing crowds in saving and raises the capital-labor ratio in the long run.

Anticipated as well as unanticipated monetary policy affects the cyclical behavior of real variables. To obtain the result that anticipated monetary policy does not matter for the behavior of real variables, it is necessary to assume that all prices always and instantaneously assume their competitive market-clearing values, in addition to having private agents with rational expectations based on the same information as is available to the monetary authorities. Even then, anticipated monetary policy will not be neutral unless the “structural” inflation tax channel is also assumed to be ineffective.

There are no grounds, in view of these conclusions, for believing that constraining the conduct of fiscal, monetary, and financial policy by very simple, inflexible rules, such as a constant growth rate for the money supply, a balanced budget, or a constant share of public spending in gross national product, will be optimal or even sensible.

Public Investment, Crowding Out, and Growth: A Dynamic Model Applied to India and Korea—v. sundararajan and subhash thakur (pages 814-55)

A dynamic model of public investment, private investment, savings, and growth is developed and applied to India and Korea. The model highlights the impact of public investment on private investment and growth by incorporating the various channels of influence that public investment has on private investment. Public investment crowds out private investment in the short term; however, it also raises the productivity of the private capital stock and, by creating demand for the output of the private sector, raises the output expectations and the investment requirements of the private sector. It also raises aggregate output and savings, thereby offsetting part of the initial crowding out. Further interesting features of the model include the relationship between the relative cost of capital and the productivity of capital, and the one between savings and the speed of adjustment of the actual capital stock toward a desired capital stock.

The model is used to investigate the magnitude and the dynamic time path of the effect of public investment on private investment and growth in India (1960-76) and Korea (1958-76). The estimates reveal that in India public investment partially crowds out private investment and dampens growth, while in Korea it promotes private investment and stimulates growth. In Korea, the relative cost of capital is found to have a strong positive efficiency effect on capital but only a weak negative substitution effect on investment; in India, both effects have the same sign as they do for Korea and are found to be strong. The dynamic responses after an initial increase in public investment takes place differ sharply between the two countries; a strong initial crowding out is only slowly reversed in India, while in Korea the positive effects on private investment dominate and are large in both the immediate and subsequent periods. The effects of changes in interest rates, operating through their influence on the cost of capital as well as on the real interest rate on savings, are also analyzed.

RESUMES

Facteurs intéerieurs déterminant l’investissement net étranger dans les comptes nationaux des Etats-Unis —george m. von furstenberg (pages 637-78)

Le solde des opérations courantes (synonyme du concept d’investissement net étranger tel qu’il est utilisé dans les comptes relatifs au revenu et au produit national aux Etats-Unis) est déterminé en grande partie par des facteurs intérieurs qui ont fait varier le montant de l’épargne nationale par rapport à la demande d’investissements intérieurs. C’est ainsi que le taux estimatif de l’investissement net étranger a diminué dans une proportion égale à près de la moitié du taux moyen de l’épargne nationale nette, corrigé des variations conjoncturelles au cours des 20 dernières années, et cet écart ne fait que croître. Ce phénomène peut être établi à l’aide d’un modéle trimestriel d’équations simultanées oα le taux prévu de l’investissement net étanger est calculé comme étant la différence entre la somme des taux d’épargne nets estimatifs des administrations publiques, des ménages et des sociétés, d’une part, et la somme des taux d’investissement estimatifs pour la formation intérieure du capital fixe et les stocks, d’autre part. Les variables conjoncturelles figurant dans ces équations et permettant de les résoudre au point d’équilibre sont elles-mêmes endogènes. Les résultats obtenus semblent prouver que les mesures prises pour relever le taux d’épargne nationale contribueraient autant à l’ajustement externe des Etats-Unis qu’ à son stock de capital intérieur et, partant, à son produit intérieur brut potentiel. Par conséquent, on ne saurait sans risque omettre de considérer le degré d’ouverture de l’économie américaine dans les débats portant sur la politique à suivre aux Etats-Unis en matiére d’àpargne et d’investissement.

Taux de change, inflation et cercles vicieux —marian e. bond (pages 679-711)

Pour déterminer comment et pourquoi des cercles vicieux peuvent apparaître et si certains pays sont plus enclins à ce genre de perturbations que d’autres, l’auteur élabore un cadre théorique permettant d’analyser ces questions et d’étudier les observations empiriques qui s’y rapportent. Un cadre conceptuel reposant sur l’équilibre général est utilisé pour développer la théorie relative aux origines des cercles vicieux et analyser l’ajustement qui s’effectue entre un état stable et un autre état stable. On peut identifier l’origine des cercles vicieux comme étant une phase d’ajustement au cours de laquelle interviennent simultanément une inflation des prix et des salaires et une dépréciation du taux de change. En ‘absence d’une politique qui laisse place à l’inflation des prix et des salaires, cette phase sera suivie par une période de recul de l’’inflation des prix et des salaires pendant laquelle l’économie s’ajustera à son nouvel état stable. L’effet exercé par la politique monétaire sur ce processus d’ajustement est étudié et l’auteur cherche à déterminer si un choc sur le marché des changes peut être à l’origine de cercles vicieux et si des àconomies ouvertes et de petite taille sont plus susceptibles d’engendrer des cercles vicieux que des économies plus grandes et relativement fermées.

Les observations empiriques sont centrées sur l’identification de caractéristiques nationales de nature à déclencher des cercles vicieux lorsque l’ajustement s’effectue sur les marchés des changes, et lorsque les variations du taux de change font ensuite sentir leurs effets dans l’économie intérieure. L’auteur examine trois types de politique économique intérieure qui ont une incidence sur le processus d’ajustement et conclut que les expériences diverses en matiére de cercles vicieux ou de cercles vertueux résident essentiellement dans les différences entre les politiques adoptées par les autorités monétaires.

Incidence de l’inflation sur la politique budgétaire dans les pays en développement —peter s. heller (pages 712-48)

L’étude fournit une analyse théorique et empirique des facteurs déterminant la rapidité d’ajustement respective des diverses catégories de recettes et de dépenses vis-à-vis de l’inflation. Bien que les taux d’ajustement respectifs des dépenses totales et des recettes totales accusent une variabilité considérable pour les 24 pays en développement analysés, les dépenses ont tendance à s’ajuster plus rapidement dans 60 % des pays. Pour un pays donné, la réaction budgétaire vis-à-vis de l’inflation semble varier en fonction du stade du processus inflationniste. A mesure que le taux d’inflation augmente, la réaction du côté des dépenses s’accélère, alors que du côté des recettes elle intervient avec un décalage; une fois que l’inflation se stabilise à un niveau plus àlevà, les coefficients d’ajustement des dépenses et des recettes convergent vers des niveaux inférieurs à ceux de la période oα le taux d’inflation était plus faible. D’un pays à l’autre, on constate aussi que les pays qui enregistrent des taux d’inflation moyens élevés ont des coefficients d’ajustement globaux plus forts, les dêpenses ayant tendance à s’ajuster plus rapidement que les recettes. Ces coefficients d’ajustement budgétaires semblent être aussi assez insensibles aux rythmes d’accélération ou de décélération du taux d’inflation. Enfin, les résultats semblent indiquer que les dépenses s’ajustent plus rapidement que les recettes aux taux d’inflation anticipés; en revanche, on observe le résultat inverse dans le cas de taux d’inflation non anticipés.

Les tests concernant les propriétés d’ajustement à l’inflation des composantes individuelles des dépenses tendent à prouver que les salaires et les dépenses en capital s’ajustent bien moins rapidement que les achats d’autres biens et de services. Les versements d’intérêts, les transferts en capital et les prêts nets s’ajustent rapidement. Du côté des recettes, l’impôt sur le revenu des sociétés s’ajuste plus rapidement que l’impôt sur le revenu des personnes physiques et, pour des taux d’inflation élevés, les recettes provenant des taxes sur les ventes intérieures au détail s’ajustent plus rapidement que l’impôt sur le revenu. En bref, l’incidence de l’inflation sur la situation budgétaire nette du secteur public n’est pas prévisible a priori, car elle est déterminée par les réactions délibérées des responsables du budget, par la structure particulière des recettes et des dépenses dans le pays considéré et par la nature du contexte inflationniste. Si l’inflation entraîne un déficit budgétaire plus important, c’est que ce résultat a été, dans une certaine mesure, projeté.

L’analyse des restrictions commerciales au moyen d’un modèle d’équilibre general, avec une application a l’Argentine — andrew feltenstein (pages 749-84)

Un modèle désagrégé d’équilibre général est établi sur la base d’une petite économie ouverte, comportant des taxes ad valorem sur la production intérieure et des taxes sur les importations. Il fait intervenir la monnaie du pays et les devises, ce qui implique la présence de comptes “transactions courantes” et “capitaux”, ainsi qu’un niveau gànàral des prix. L’existence d’un équilibre du modèle est démontrée et une méthode de calcul pour sa solution est àlaboràe. Le modèle est utilisé avec les données relatives à l’Argentine et fournit une bonne approximation des résultats effectifs du quatriême trimestre de 1978. Une simulation est effectuée, après une réduction des droits d’importation de 50 %; l’efficacité avec laquelle la politique monétaire et la politique des taux de change stabilisent la balance des paiements, compte tenu de l’incidence exercée par la libéralisation des échanges, est examinée. On constate que les modifications de l’expansion du crédit intérieur et du taux de change nominal, nécessaires à une neutralisation totale de la balance des paiements, sont plus fortes que ce que l’on pouvait attendre. Les recettes publiques totales procurées par chaque politique considérée sont estimées et les implications de ces politiques du point de vue de l’efficacité économique sont étudiées grâce au calul d’indices d’utilité. On étudie la stabilité du modèle en continuant la simulation pendant les deux périodes suivant la libéralisation initiale, et it apparaît que le choc initial subi par la balance des paiements s’atténue rapidement, même en l’absence de mesures prises par les autorités.

Politique monétaire, financière et budgétaire dans l’hypothèse d’anticipations rationnelles —willem h. buiter (pages 785-813)

L’étude analyse les répercussions que pourrait avoir sur la conduite de la politique monétaire, budgétaire et financière la notion révolutionnaire des anticipations rationnelles en macroéconomie. Elle en déduit que certaines conclusions importantes auxquelles on parvient avec des modèles néo-keynésiens éclectiques classiques demeurent valables lorsqu’on fait intervenir des anticipations rationnelles : anticiper une politique, ce n’est pas la neutraliser. Sont prises en considération les politiques structurelles visant à modifier le niveau et la composition de la production en situation de plein emploi à court et a long terme, de même que les politiques de stabilisation visant à influer sur les écarts cycliques de la production et de l’emploi par rapport à leur niveau de plein emploi.

Pour un niveau et une composition donnés de dépenses publiques réelles au titre des biens et des services, le remplacement des impôts courants (forfaitaires) par un financement sous forme d’obligations réduit l’épargne à court terme et diminue le coefficient capital-travail à long terme. Cet effet de refoulement persiste même si chaque agent économique tient pleinement compte des impôts futurs requis pour assurer le service de la dette publique rémunérée détenue par les particuliers. La situation est analogue, à quelques nuances près, lorsque l’on tient compte d’éventuels legs ou autres donations privées d’une génération à l’autre. La substitution du financement monétaire au financement par l’impôt n’est pas non plus un facteur neutre. L’impôt-inflation n’est pas équivalent à l’impôt courant explicite. Dans les modéles simples abordés dans cette étude, la substitution du financement monétaire au financement par l’impôt augmente l’épargne et accroît à long terme le coefficient capital-travail.

Une politique monétaire, qu’elle soit anticipée ou non, influe sur le comportement conjoncturel des variables réelles. Pour faire en sorte que la politique monétaire anticipée n’influe pas sur le comportement des variables réelles, il faut supposer que tous les prix, à tout moment et instantanément, prennent leurs valeurs compétitives susceptibles d’équilibrer le marché et que les agents privés aient établi des prévisions rationnelles fondées sur les mêmes renseignements que ceux dont disposent les autorités monétaires. Même en pareil cas, une politique anticipée ne sera pas neutre, à moins que le circuit structurel par lequel opére l’impôt-inflation ne soit aussi considéré comme étant sans effet.

A la lumière de ces conclusions, it n’y a pas lieu de penser que les restrictions imposées à une politique budgétaire, monétaire et financière par le moyen de règles inflexibles très simples telles qu’un taux de croissance constant de la masse monétaire, un budget équilibré ou une part constante des dépenses publiques dans le produit national brut, soient optimales ou mêmes judicieuses.

Investissement public, “effet de refoulement” et croissance : un modèle dynamique appliqué à l’Inde et à la Corée —v. sundararajan et subhash thakur (pages 814-55)

Un modèle dynamique de l’investissement public, de l’investissement privé, de l’épargne et de la croissance a été élaboré et appliqué à l’Inde et à la Corèe. Le modéle met en relief l’impact de l’investissement public sur l’investissement privé et la croissance en intégrant les différents circuits par lesquels l’investissement public influence l’investissement privé. L’investissement public évince l’investissement privé à court terme; par ailleurs, it accroît également la productivité du stock de capital privé et, en créant une demande pour la production du secteur privé, augmente les perspectives de production et les besoins d’investissement du secteur privé. Il contribue aussi à augmenter la production globale et l’épargne, compensant ainsi en partie l’effet initial de refoulement. Parmi les autres aspects interessants du modèle figurent la relation entre le coût relatif du capital et la productivité du capital, ainsi que la relation entre l’épargne et la rapidité avec laquelle le stock de capital effectif s’ajuste pour se rapprocher du stock de capital désiré.

Le modèle permet d’étudier l’ampleur et le profil temporel dynamique de l’effet de l’investissement public sur l’investissement privà et la croissance en Inde (1960-76) et en Corée (1958-76). Les estimations montrent qu’en Inde, l’investissement public évince partiellement l’investissement privé et affaiblit la croissance, tandis qu’en Corée it encourage l’investissement privé et stimulela croissance. En Corée, it apparaît que le coût relatif du capital a un important effet positif sur le capital quant à l’efficacité de ce dernier, mais que l’effet négatif de substitution qu’il exerce sur l’investissement est faible; en Inde, les deux effets ont le même signe qu’en Corée mais ils sont très prononcés. Les réactions dynamiques à une augmentation initiale de l’investissement public sont très différentes dans les deux pays; en Inde, un important refoulement initial n’est que lentement neutralisé, tandis qu’en Corée les effets positifs sur l’investissement privé prédominent et sont importants tant dans les périodes immédiates que dans les périodes ultérieures. Les effets des variations des taux d’intérêt, qui se répercutent sur le coût du capital ainsi que sur le taux d’intérêt réel de l’épargne, sont également analysés.

RESUMENES

Factores determinantes internos de la inversión externa neta de Estados Unidosgeorge m. von furstenberg (páginas 638-78)

El saldo en cuenta corriente conocido como inversión externa neta en las cuentas del ingreso nacional y del producto nacional de Estados Unidos queda determinado principalmente por factores internos que han modificado el monto del ahorro nacional en relación con la demanda de inversión interna. Así, la tasa estimada de inversión externa neta se redujo por término medio en casi la mitad de lo que disminuyó la tasa de ahorro interno neto, ajustada respecto al ciclo, durante los dos últimos decenios, y esta proporción ha ido en aumento. Esto se establece mediante un modelo trimestral de ecuaciones simultáneas en que la tasa prevista de inversión interna neta se obtiene como la diferencia entre la suma de las tasas estimadas de ahorro neto público, personal y de las empresas por una parte y la suma de las tasas estimadas de inversión interna en capital fijo y existencias por la otra. Las variables cíclicas que figuran en estas ecuaciones y permiten su solución en equilibrio son endógenas. Las indicaciones son de que las medidas adoptadas para elevar la tasa nacional de ahorro contribuirían tanto al ajuste externo de Estados Unidos como a la masa de su capital interno y, por ende, a su producto interno bruto potencial. Por consiguiente, no se puede, sin correr peligro, hacer caso omiso de la apertura de la economia estadounidense en los debates sobre la política relativa al ahorro y la inversión de Estados Unidos.

Tipos de cambio, inflación y circulos viciosos —marian e. bond (páginas 679-711)

Se presenta un marco teórico mediante el cual se pueda analizar la forma en que se producen los círculos viciosos y las razones de que se produzcan, y se trata de averiguar, mediante la información empírica pertinente, si ciertos países son más propicios que otros a las reacciones denominadas círculos viciosos. Se utiliza un marco de equilibrio general para desarrollar la teoría de los orígenes de los círculos viciosos y analizar el ajuste de un estado estable a otro. El origen de los círculos viciosos puede identificarse como una fase de ajuste en la que se producen al mismo tiempo la inflación de precios y salarios y la depreciación del tipo de cambio. De no tomarse medidas de política para adaptarse a la inflación de precios y salarios, esta fase irá seguida de un período en el que amaine dicha inflación al irse ajustando la economía a su nuevo estado estable. Se examina el efecto de la política monetaria sobre ese proceso de ajuste, y se realiza una evaluación de si una conmoción en el mercado de divisas puede ocasionar círculos viciosos y si las economías abiertas y pequeñas son más vulnerables a los círculos viciosos que las economías grandes y relativamente cerradas.

La informacién empórica se centra en los aspectos identificadores que son causantes de los círculos viciosos en los distintos países al producirse el ajuste en los mercados de divisas y al transmitirse a la economía interna las variaciones del tipo de cambio. Se consideran tres clases de política interna que ejercen un impacto en el proceso de ajuste, y se llega a la conclusión de que la diferencia entre los casos de círculo vicioso y de círculo virtuoso son atribuibles principalmente a las diferencias en la política seguida por las autoridades monetarias.

El impacto de la inflación en la política fiscal de los países en desarrollo—peter s. heller (páginas 712-48)

En este trabajo se presenta un análisis tanto teórico como empírico de los factores que influyen en la rapidez relativa del ajuste de las diversas clases de ingresos y gastos públicos a la inflación. Si bien hay una diversidad considerable en el ritmo relativo de ajuste del gasto e ingreso públicos totales en los 24 países en desarrollo estudiados, tiende a ajustarse más rápidamente el gasto en el 60 por ciento de los países. Para un determinado país, la reacción fiscal ante la inflación parece variar según la fase del proceso inflacionario. Al aumentar la tasa de inflación se produce una reacción acelerada del gasto y una reacción desfasada del ingreso fiscal; al estabilizarse la inflación a un nivel estable más alto, convergen los coeficientes de ajuste del gasto e ingreso púiblicos en un nivel más bajo que en el período con una baja tasa de inflación. Al comparar los paises, hay tambíen pruebas de que los que experimentan una elevada tasa media de inflación tienen un coeficiente global de ajuste más alto, tendiendo a ajustarse el gasto más rápidamente que los ingresos. Estos coeficientes de ajuste fiscal parecen ser bastante insensibles al ritmo de aceleración o desaceleración de la tasa de inflación. Por último, los resultados parecen indicar que el gasto se ajusta más rápidamente que los ingresos a la inflación prevista. En cambio, en lo que se refiere a la inflación no prevista se obtiene el resultado contrario.

Las pruebas efectuadas sobre las propiedades de ajuste de los distintos componentes del gasto a la inflación indican que los gastos públicos de capital y en salarios se ajustan con mucha menos rapidez que las compras de otros bienes y servicios. Los pagos de intereses, las transferencias de capital y los préstamos netos se ajustan con rapidez. Entre los ingresos fiscales, los procedentes de los impuestos a las sociedades se ajustan más rápidamente que los del impuesto sobre la renta de las personas físicas y, con altas tasas de inflación, los ingresos procedentes del impuesto a las ventas internas se ajustan más rápidamente que los del impuesto a la renta. En resumen, el impacto de la inflación en la situación fiscal neta del sector público no es previsible a priori, estando determinado por la reacción discrecional de los que ejercen la facultad de decisión presupuestaria, por la estructura concreta de gastos e ingresos fiscales del país y por el carácter del ámbito inflacionario. Si la inflación ocasiona un déficit presupuestario mayor, es éste un resultado que no estaba totalmente imprevisto.

Análisis de las restricciones comerciales desde el punto de vista del equilibrio general, aplicado a Argentina—andrew feltenstein (páginas 749-84)

Se prepara un modelo desagregado de equilibrio general de una pequeña economía abierta con impuestos ad valórem sobre la producción interna y aranceles de importación. Se incluyen la moneda nacional y las divisas, lo cual hace posible la existencia de una cuenta corriente y una cuenta de capital y de un nivel general de precios. Se demuestra la existencia del equilibrio del modelo y se desarrolla un método de solución mediante computadora. El modelo se aplica con datos sobre Argentina y ofrece una buena aproximación de los resultados reales del cuarto trimestre de 1978. El modelo se simula suponiendo una reducción del 50 por ciento de los aranceles y se examina la eficacia de las politicas monetarias y del tipo de cambio en lo que se refiere a la estabilización de la balanza de pagos después del impacto recibido a causa de la liberalización del comercio. Se llega a la conclusión de que los cambios que son necesarios en la expansión del crédito interno y el tipo de cambio nominal para que la balanza de pagos se vea completamente neutralizada son mucho mayores de lo que podria preverse. Se estima la renta pública total que podría aportar cada una de las políticas consideradas y, mediante el cálculo de índices de utilidad, se considera la repercusión de dichas políticas en el bienestar público. Se examina la estabilidad del modelo permitiendo que la simulación continúe durante dos períodos posteriores a la liberalización inicial de los aranceles y se observa que el impacto inicial recibido por la balanza de pagos disminuye rápidamente, aunque el gobierno no adopte medidas de política.

Las políticas monetaria, financiera y fiscal en condiciones de expectativas racionales—willem h. buiter (páginas 785-813)

En este trabajo se evalúan las implicaciones que tiene para la aplicación de las políticas monetaria, financiera y fiscal la revolución de las expectativas racionales en macroeconomía. Se demuestra que algunas conclusiones importantes de política derivadas de los modelos eclécticos neokeynesianos convencionales siguen siendo válidas cuando se introducen expectativas racionales. El que se prevea Ia política no significa que quede neutralizada. Se consideran las medidas de política estructurales encaminadas a alterar el nivel y la composición del producto de pleno empleo a corto y a largo plazo así como las medidas de estabilización tomadas para influir en las desviaciones coyunturales del producto y el empleo con respecto a sus niveles de pleno empleo.

Para un determinado nivel y composición del gasto real del Estado en bienes y servicios, el financiamiento mediante bonos en lugar de mediante impuestos corrientes (de tanto alzado) reduce el ahorro a corto plazo y la razón capitaltrabajo a largo plazo. Esta exclusiín persiste aun cuando todos los agentes económicos tengan en cuenta los impuestos futuros necesarios para atender el servicio de la deuda pública que devenga intereses y se halla en manos de los particulares. También se cumple, con ciertas condiciones, cuando se tiene en cuenta la posibilidad de legados y otras donaciones privadas entre generaciones. Tampoco es indiferente el que se utilice el financiamiento monetario en lugar del financiamiento tributario. El “impuesto de la inflación” no es equivalente a los impuestos corrientes explicitos. En los modelos sencillos de este trabajo, la utilización de financiamiento monetario en lugar de financiamiento tributario atrae al ahorro y eleva la razón capital-trabajo a largo plazo.

La política monetaria prevista, al igual que la imprevista, afecta al comportamiento coyuntural de las variables reales. Para obtener el resultado de que la política monetaria prevista no afecta al comportamiento de las variables reales, hace falta suponer que todos los precios asumen siempre e instántaneamente su valor competitivo al que se equilibra el mercado, además de tener agentes privados con expectativas racionales basadas en la misma información de que disponen las autoridades monetarias. Aun entonces, la política monetaria prevista no será neutral a menos que se suponga también que es ineficaz el cauce del “impuesto estructural de la inflación.”

En vista de esas conclusiones, no hay justificación para creer que sea óptimo, o incluso sensato, el contener la aplicación de las políticas fiscal, monetaria y financiera mediante reglas sencillas e inflexibles, como por ejemplo la adopción de una tasa constante de crecimiento de la oferta monetaria, un pre-supuesto equilibrado o una proporción constante del gasto pdblico en relación con el producto nacional bruto.

Inversión pública, desplazamiento de la inversión privada y crecimiento: Un modelo dinámico aplicado a India y Corea—v. sundararajan y subhash thakur (páginas 814-55)

Se elabora un modelo dinámico de inversion pública, inversión privada, ahorro y crecimiento, y se aplica a India y Corea. El modelo destaca el impacto de la inversión pública en la inversión privada y el crecimiento incorporando los diversos canales de influencia de la inversión pública en la privada. La inversión pública desplaza a la privada a corto plazo; sin embargo, también eleva la productividad de la masa de capital privado y, al crear demanda para la producción del sector privado, eleva las expectativas de producción y las necesidades de inversión del sector privado. Por otra parte, como acrecienta la producción y el ahorro agregados, contrarresta parcialmente el efecto inicial de desplazamiento. Otras características de interés del modelo son la relación entre el costo relativo del capital y su productividad, y entre el ahorro y la velocidad de ajuste de la masa de capital efectiva a la que se desea obtener.

El modelo se utiliza para investigar la magnitud y trayectoria dinámica en el tiempo del efecto de la inversión pública en la privada y en el crecimiento en India (1960-76) y Corea (1958-76). Las estimaciones ponen de manifiesto que en India la inversión pública desplaza parcialmente a la privada y frena el crecimiento, mientras que en Corea promueve la inversión privada y estimula el crecimiento. Se comprueba que en Corea el costo relativo del capital tiene un fuerte efecto positivo de eficiencia en el capital, pero sólo un efecto negativo débil de sustitución en la inversión; en India, ambos efectos tienen el mismo signo que en Corea, y son fuertes. La reacción dinámica posterior al aumento inicial de la inversión pública es muy distinta en los dos países; en India, el fuerte efecto inicial de desplazamiento sólo cambia de dirección lentamente, mientras que en Corea predominan los efectos positivos en la inversión privada y son importantes en el período inmediato y en los siguientes. También se analizan los efectos de las variaciones de los tipos de interés, que alteran el costo del capital y el tipo de interés real del ahorro.

In statistical matter (except in the résumés and restimenes) throughout this issue,

Dots (…) indicate that data are not available;

A dash (—) indicates that the figure is zero or less than half the final digit shown, or that the item does not exist;

A single dot (.) indicates decimals;

A comma (,) separates thousands and millions;

“Billion” means a thousand million;

A short dash (-) is used between years or months (e.g., 1977-79 or January-October) to indicate a total of the years or months inclusive of the beginning and ending years or months;

A stroke (/) is used between years (e.g., 1978/79) to indicate a fiscal year or a crop year;

Components of tables may not add to totals shown because of rounding.

International Monetary Fund, Washington, D.C. 20431 U.S.A.

Telephone number: 202 477 7000

Cable address: Interfund

*

Mr. Sundararajan, Senior Economist in the Asian Department, received his masters degree from the Indian Statistical Institute and his doctorate from Harvard University.

Mr. Thakur, Senior Economist in the European Department, received his masters degree from the University of Poona, India, and his doctorate from the University of California at San Diego.

1

For a review of the literature on the effects of government expenditure on capital formation in the context of developed countries, see von Furstenberg and Malkiel (1977). For some empirical work on private investment functions for developing countries, with public investment as an explanatory variable, see Wai and Wong (1979).

2

The initial statement of Jorgenson’s theory can be found in Jorgenson (1963); for later surveys, see Jorgenson (1971) and Clark (1979).

3

The adoption of the neoclassical framework for modeling investment behavior in developing countries needs some justification, since it is sometimes argued that several basic assumptions attributed to the neoclassical theory do not hold in these countries and that lack of data on the capital stock prevents observations on the capital stock adjustment process. We believe that these objections are not strong enough to invalidate the many useful insights provided by the neoclassical theory. Many of the assumptions attributed to the neoclassical theory—maximization of rates of return by economic agents; existence of perfect markets for goods, including secondhand capital goods; and well-developed financial markets—are not essential to the propositions of the theory. Moreover, lack of data on the capital stock is not a problem in the commonly used linear models. Although the neoclassical framework is commonly used in analyzing production relations in developing countries, the use of such models in studying the savings-investment process is rare. For a recent exception, see Williamson (1979).

5

See Pinell-Siles (1979). The portion of the private sector consisting of medium and large enterprises accounts for only 15 per cent of total private sector investment in India, the rest being accounted for by the small-scale and household sectors.

6

The influence of public investment on the private sector in developing countries has been studied elsewhere without a complete model of the process. See Wai and Wong (1979), who estimate private investment functions for many countries with government investment as one of the explanatory variables. However, they ignore the crucial role of private sector output and the capital stock as explanatory variables. They also do not make proper allowance for the possibility of crowding out of the private sector; nor do they attempt to trace the dynamic effects of public investment, which requires a complete model of investment and growth.

7

The common assumption that firms maximize profits is not necessary for deriving the neoclassical investment function; the assumption that firms minimize total costs is sufficient. See Hall (1977).

8

The acquisition cost includes, in addition to the purchase cost (the value of investment), other costs such as the costs of installation, which are ignored here for the sake of simplicity.

9

The real interest rates for savers and investors differ not only because of the difference in deposit and lending rates but also because of the differences in the prices relevant for defining the real rates—consumer prices for savers and capital goods prices for investors.

10

The appropriateness of the short-term rate is emphasized in Hall (1977), who derives this result in a discrete time framework.

11

Fixed costs of production are ignored, for the sake of simplicity, by assuming that the addition to fixed costs owing to an increase in the capital stock is small relative to the fixed costs of operating the existing capital.

12

The use of the rental-wage ratio as a determinant of the desired capital stock is not common in the investment literature; exceptions are Jorgenson (1967) and McLaren (1971).

13

Even though equation (4) is linear in logarithms, we have avoided the logarithmic specification, because our capital stock series are net of the initial capital stock and are not suitable for the logarithmic specification.

14

For more on the optimality of the partial adjustment response, see Eisner and Strotz (1963) and Rothschild (1971).

15

This technique of introducing variability in the speed of adjustment is similar to the one used by Coen (1971).

16

If the role of the price mechanism in the crowding-out process is to be emphasized, then the model should be extended to endogenize interest rates and the relative price of capital input. However, we have preferred to focus on the typical case of rigidity in interest rates and the relative price of capital and to treat these variables as exogenous. Also, no account is taken of the crowding out of private investment through increases in the general price level, which results in the transfer of resources to the public sector via an inflation tax.

17

Since the production and investment functions are both derived from the same equilibrium condition, which is shown in equation (3), there are interrelationships between the parameters of equations (9) and (10). Some of the underlying parameters can be estimated, and the implied relationships can be tested. The test results indicate that the implied relationships between the parameters are consistent. These results can be obtained on request from the authors, whose address is International Monetary Fund, Washington, D.C. 20431.

18

Leff and Sato (1975) have argued that savings should depend on the first difference in income. Alternatively, one can specify that savings depend on the first difference in permanent income. Another common specification is that savings depend on permanent income and transitory income, with different marginal propensities for each type of income. All of these formulations can be captured by a general distributed lag in income. The estimated shape of the distributed lag should reveal the correct underlying structure.

19

The ceteris paribus assumption requires that the increased public investment be financed in such a way that the aggregate savings are not immediately affected. This would be the case if there were an offsetting increase in taxation or in borrowings from the public without any change in interest rates and central bank lending.

20

For a discussion of the issues in the estimation of capital stock, see Thomas (1979). Krishnamoorthy (1964) uses an alternative technique that exploits the information on the initial capital stock contained in the regression coefficients.

21

For Korea, the rental price of capital measured without the rate of increase in capital goods prices (i.e., capital gains) provided a better fit than the rental price including these price increases. For India, however, the rental price measured to include the rate of increase in capital goods prices gave a better fit. The measurement and interpretation of the rental price of capital have varied widely among researchers. See Rowley and Trivedi (1975).

22

The method of testing for factor substitution that is used here has some parallels in the development literature. For example, the factor shares are regressed on rental-wage ratios, as in Ioannides and Caramanlis (1979). Their measure of the rental price of capital is the commonly used nonwage income as a proportion of capital stock. This measure is clearly quite different from the one used here. Our measure appears preferable, however, because it is based on explicit dynamic considerations.

23

Moreover, in Korea, the rental-wage ratio rose sharply during 1960-67 and contributed to a decrease in the capital-output ratio; whereas in India, the rental-wage ratio declined during most of the period 1960-71 and contributed to an increase in the capital-output ratio. These developments partly explain the differences noted earlier in the incremental capital-output ratios of the two countries.

24

The real interest rate is measured 1+r1+PC˙/PC1 rather than the usual (rPC˙/PC). The results for Korea appear to be contrary to the findings of Brown (1973), who noted a significant positive relationship between the real interest rate and total private savings. This difference in results may be attributable to the facts that total domestic savings, rather than private savings alone, are explained here and that we have included lagged values of income.

25

This result is obtained by solving the complete model in order to analyze the effects of interest rates. In a dynamic model, it is necessary to distinguish between the direct effects of a policy variable (the coefficients in a single equation) and the impact and long-run multiplier effects (obtained by dynamically solving the complete model). Weak direct effects can coincide with strong multiplier effects.

26

The lag distribution was estimated using a second-degree polynomial without imposing any conditions on the coefficients. The sum of the coefficients (0.13) is significant at the 95 per cent probability level.

27

It is implicitly assumed that the increased public investment is financed either by raising taxes or by borrowing from the public, while interest rates and central bank lending remain unaffected and total savings are not immediately affected. The modifications of the conclusions for deficit financing are readily apparent, since rates of inflation enter the model through the real interest rate variable. The modifications for foreign financing are also straightforward.

Other Resources Citing This Publication