buun

Journal of Post Keynesian Economics / Summer 2013, Vol. 35, No. 4 599© 2013 M.E. Sharpe, Inc. All rights reserved. Permissions: www.copyright.com

ISSN 0160–3477 (print) / ISSN 1557–7821 (online)DOI: 10.2753/PKE0160-3477350405

MArcOS rOchA AND JOSé LuIS OrEIrO

Capital accumulation, external indebtedness, and macroeconomic performance of emerging countries

Abstract: This paper aims at presenting a nonlinear post Keynesian growth model to evaluate at the theoretical and empirical levels the relationship between external indebtedness and economic growth in emerging countries. To this end, a post Keynesian endogenous growth model is presented, in which: (1) the desired rate of capital accumulation is assumed to be a nonlinear function of external indebtedness as a share of capital stock; (2) an endogenous country risk premium is assumed to be an increasing (linear) function of external indebtedness (as a share of capital stock); (3) there is a fixed exchange rate regime and perfect capital mobility in the sense of Mundell and Fleming. The main theoretical result of the model is the existence of two long-run equilibrium positions, one of which has a high level of external indebtedness (as a ratio of capital stock) and a low profit rate and the other has a low level of external indebtedness and a high profit rate. This means that “excessive” external indebtedness can result in stagnant growth due to its negative effect on the rate of profit. To test the effects of external indebtedness on the rate of economic growth in emerging economies, a dynamic panel is estimated to evaluate whether external debt has an effective negative influence on economic growth in emerging countries. An empirical test of demand-led growth equations with a dynamic panel for fifty-five emerging countries confirms the potential negative effects of external debt on long-term growth rates in the sample countries.

Key words: External debt, balance of payments, and capital accumulation.

JEL classifications: F3, F4, O2

The proponents of capital account liberalization argue that one of the main benefits of free convertibility is that it gives emerging countries

Marcos rocha is an assistant researcher at the Institute for Applied Economic re-search (IPEA). José Luis Oreiro is a professor in the Economics Department at universidade de Brasília and a researcher of National council for Scientific and Tech-nological Development (cNPq).

600 JOURNAL OF POST KEYNESIAN ECONOMICS

greater access to international capital markets and gives greater inflows of external savings to these countries. An increase in the savings rate (internal + external) will produce an increase in the gross domestic product (GDP) per capita of these economies in the long run, accord-ing to neoclassical models such as Solow–Swan. It is then obvious that emerging countries should open capital accounts as a means to increase the level of income per capita and catching up with income levels of the developed countries.

This view in favor of free convertibility of the current account is mainly founded on the hypothesis that economic growth can be stimulated or induced by external savings. Since external savings are the counterpart of current account deficit, it follows that growth in the developing countries should necessarily be linked to large current account imbalances. The ac-cumulation of current account deficits over time results in growing levels of external debt. This means that the view pro free convertibility of the current account also states the existence of a positive relation between economic growth and external debt (see, among others, Eaton, 1993).

More recently this naive view of external debt and growth has been criticized at the theoretical and empirical levels by a growing body of economic literature about the nonlinear effects of external debt on growth. Pattillo et al. (2002) argue that although for low levels of external debt, economic growth can be stimulated by foreign savings, there are compel-ling reasons to believe that external debt may depress economic growth after some threshold level. One reason for a negative effect of external debt on growth is the “debt overhang theory.” According to this theory, if there is some likelihood that in the future debt will be larger than the country’s ability to repay it, then expected debt service will be an increas-ing function of the country’s output level. In that case, the returns from investing in the country face high marginal taxes from external creditors, and new domestic and foreign investment is discouraged, depressing economic growth (Krugman, 1988; Sachs, 1989).

clements et al. (2003) also argue that for low-income countries a high level of external debt has an indirect but negative effect on economic growth due to the effect of a higher level of debt service over public in-vestment. Indeed, a higher level of debt service reduces the current fiscal position of government in low-income countries, imposing a contraction of public investment in order to restore fiscal balance.

The empirical evidence on the existence of a nonlinear relation between external debt and economic growth is, however, mixed. Patillo et al. (2002) based on a sample of ninety-three developing countries in the pe-riod 1963–98 found a nonlinear effect of external debt on growth, and also

MACROECONOMIC PERFORMANCE OF EMERgINg COUNTRIES 601

that the average effect of debt on per capita growth became negative for debt levels above 160–70 percent of exports and 35–40 percent of GDP. These findings are corroborated by clements el al. (2003), who, based on a sample of fifty-five low-income countries for the period 1970–99, found even lower levels for the threshold level of external debt. Accord-ing to them, external debt had a negative effect on economic growth for debt levels above 30–37 percent of GDP and 115–20 percent of exports. Schclarek and ramon-Bellester (2005) reached different results. Based on a sample of twenty Latin American and caribbean countries for the period 1970–2002, these authors found evidence of a negative effect of total external debt on economic growth, but no evidence whatsoever of a nonlinear relationship between these variables.

Post Keynesian economists are also skeptical regarding the effects of capital account liberalization on long-term growth. Although the balance-of-payments constraint is considered to be the ultimate constraint to output expansion in the post Keynesian literature of demand-led growth (Kaldor, 1977; Thirwall, 1979, 2002), external saving can only produce a temporary increase in the growth rate that is compatible with balance-of-payments equilibrium. In fact, a current account deficit financed by capital flows allows a higher growth rate of imports—and, consequently, a higher rate of growth—than would be the case if the current account was balanced. however, current account deficits increase the level of external debt, increasing the flow of debt services (interest and amortization), thereby reducing the rate of growth that is compatible with equilibrium in the balance of payments. It can be shown that in the long-run equi-librium, where the ratio of current account deficit to domestic income is constant, the balance-of-payments equilibrium growth rate is unlikely to be substantially affected by the growth of capital flows (Mccombie and roberts, 2002, p. 95).

Another issue regarding capital account liberalization is the financial fragility induced by this movement. Since emerging economies are, in general, incapable of borrowing in domestic currency on international capital markets, capital inflows produce a currency mismatch between assets and liabilities of these economies. This problem can worsen if li-abilities are of short-term maturity and assets are of long-term maturity, that is, if short-term capital flows are used to finance long-term assets such as real estate. These problems increase the economy’s level of ex-posure to a currency crisis in the case of a sudden stop of capital flows (Davidson, 2002, ch. 13; Palma, 2002). A sudden stop can be the result of self-fulfilling prophecies if short-term external debt is higher than the country’s international reserves (rodrik and Velasco, 1999). In this


case, an expectation of currency devaluation will trigger larger outflows of capital, producing greater exchange rate depreciation (independent of the exchange rate regime), which will force the central Bank to increase interest rates (mainly in inflation-targeting countries) in order to avoid a huge increase in inflation rates due to the pass-through effect of exchange rate to inflation. The monetary contraction will result in output loss and a substantial reduction in growth rates.

The link between external saving and economic growth was also ana-lyzed by Bresser and Nakano (2003). For them, external financing tends to generate a reduction in the long-term growth rates of emerging coun-tries due to an “excessive” increase in external debt. In fact, an external debt increase makes the emerging countries even more susceptible to a balance-of-payment crisis, which, to resolve, demands significant devalu-ations of nominal and real exchange rates, in turn increasing inflation rates and inducing tight monetary and fiscal policies as a way to control inflation. This dynamic makes a huge trade surplus necessary to “resolve” the balance-of-payments crisis, usually through cutting domestic absorp-tion by means of very tight fiscal and monetary policies.

up to now, however, little effort was made by post Keynesians in order to analyze in a systematic way the relation between external debt and economic growth. In fact, the post Keynesian literature presented above does not show any precise mechanism by which accumulation of external debt through a succession of current account deficits can depress economic growth. We can obtain only some imprecise analysis of the effects of financial liberalization on financial fragility and long-term growth. The post Keynesian literature has developed no formal model to show the effects of external debt on economic growth and the channels by which external indebtedness can depress long-term growth.

The objective of the present paper is to fill this gap and develop a post Keynesian growth model to evaluate, at the theoretical and empirical levels, the hypothesis that external debt may have a negative impact on economic growth after some threshold level of external indebtedness. This nonlinear effect of external debt on long-term growth will be shown to be a consequence of the existence of a kind of “decreasing returns” of capital flows over domestic investment together with a country risk premium that is an increasing function of the level of external debt.

Indeed, for low levels of external debt, an increase in external indebt-edness by means of foreign savings will result in a significant increase in desired investment by capitalists, since investment expenditure is re-stricted, in an open emerging economy, by the capacity to import capital goods from developed economies. This means that external borrowing


can ease the foreign exchange restriction on capital goods imports, stimulating capital accumulation. however, the increase in investment rate will not be proportional to the increase in external indebtedness since capital flows will not be fully used to finance domestic investment, but a share of this increase in external debt will be used, for instance, to finance government deficits or speculation in financial markets. hence, external indebtedness will not result in a proportional increase in the investment rate. On the other hand, the domestic interest rate will be an increasing function of the level of external debt due to its effect on the country risk premium in a setting of free capital mobility in the sense of Mundell (1963) and Fleming (1962).

The combined effects of increasing levels of external debt over invest-ment expenditure and domestic interest rate will produce a hump-shaped relation between the rate of profit (and the rate of capital accumulation) and external debt: for low levels of external debt, an increase in external debt will result in an increase in the rate of capital accumulation and hence in the rate of profit; after some threshold level of external indebt-edness, however, an increase in the level of external debt will produce a decrease in the rate of capital accumulation and in the rate of profit since the positive effect of external debt on investment will be more than offset by the negative effect on domestic investment of the increase in the domestic interest rate.

When this hump-shaped relation between investment (and profit rate) and external debt is combined with the requirement for a nonexplosive ratio between external debt and capital stock (a necessary condition for balanced growth in the model to be developed in this study), then it is possible to show the existence of multiple long-run equilibrium positions for the economic system, one characterized by a low level of external debt (as a ratio to capital stock) and a high rate of profit, and the other characterized by a high level of external debt and a low profit rate. This means that “excessive” external indebtedness—relative to the threshold level of external indebtedness—can result in stagnant growth due to its negative effect on the rate of profit.

To empirically evaluate the predictions of the model about the rela-tion between external debt and economic growth, we will also build an empirical model of growth in the second part of the paper. The empirical test of demand-led growth equations with a dynamic panel for fifty-five emerging countries confirms the potential negative effects of external debt on the long-run growth rates of the sample countries. Indeed, our empirical results show that the threshold level of external indebtedness is negative, so that the “optimal policy” for an emerging country is to


become a net creditor in international markets. In order to do so, emerging countries may adopt policies that allowed them to run persistent current account surpluses.

Capital accumulation and external borrowing

In this section we present a demand-led growth model for a small open economy that operates with a fixed exchange rate and a fully convert-ible capital account. Investment and exports are assumed to be the au-tonomous component of aggregate demand, which is in sharp contrast to the standard balance-of-payments constrained growth models where exports are considered to be the only source of autonomous demand for the economy as a whole (Thirwall, 2002, p. 55).1 For the sake of sim-plicity, there are no government activities, so government expenditures and taxes are set at zero.

The fundamental blocks of the model

We assume an economy in which the firms produce a homogeneous good using labor and imported raw materials. The production technol-ogy of the economy is represented by means of a Leontief technology, in a way that the technical parameters of labor and commodities—that is, the amounts of labor and raw materials required to produce a unit of output—are independent of the production level of the firms. For sim-plicity, we assume an economy without technological progress so that labor productivity—defined as the reverse of the unitary requirement of labor—is constant over time.

Just as is assumed in most post Keynesian growth models, here we assume that the firms of this economy have market power and set the prices of their goods based on a constant markup over the variable unitary costs of production. hence, the price equation of this economy is given by (Taylor, 1989, p. 21):

p = (1+ τ)[wb + ep*0a0], (1)

1 According to Kaldor (1988), in the long run the sources of autonomous demand for the economy as a whole are given only by government expenditures and exports, since investment is supposed to be determined by the harrodian accelerator, so that it is a fully endogenous variable in the long run. According to our understanding of Keynesian investment theory, however, investment has a purely autonomous part, even in the long run, due to its dependence on entrepreneurs’ “animal spirits.” This means that it is theoretically unacceptable to treat investment as a purely endogenous variable, as is assumed in balance-of-payments constrained growth models.


where p is the domestic price level, w is the nominal rate of wages, p* is the international price level, e is the nominal exchange rate, b is the unit requirement of labor, a0 is the unit requirement of imported input commodities, and τ is the rate of markup.

Let r be the profit rate and u the level of capacity utilization. It can be shown that the profit rate is given by:

r u=+τ

τ1. (2)

hence, it can be seen that profitability is an increasing function of markup and of the degree of capacity utilization.2

The goods market is cleared when the aggregate demand equals the production level of the firms:

pC + pI + pE = pX, (3)

where pC is the nominal value of consumption expenditure, pI is the nominal value of investment expenditure, pE is the nominal value of net exports, and pX is the nominal value of the production level.

Additionally, also assumed is the existence of two social classes (capi-talists and workers), which differ by the type of income they earn (profits and wages) and by their propensity to consume out of the disposable income. As in Kaldor (1956), it is assumed that the workers “consume all they get” so their consumption propensity is equal to one (i.e., its propensity to save is equal to zero), and, at the same time, capitalists consume a constant fraction of their income (which is solely made of profits), saving a fraction sc of their income. hence, the nominal value of the consumption expenditure is given by:

pC = wbX + (1 – sc)rpK. (4)

By substituting (4) in (3):

I

K

E

Ks r qa mrc+ − − =0 0, (5)

where qep

p=

*

is real exchange rate and m =+τ

τ1 is profit share of income.

2 The markup rate is taken as given since it depends upon structural factors such as (1) the degree of monopoly of the firms in the economy; (2) the level of entry barriers to new firms in the industry; and (3) the degree of substitution between the products of a given industry.


In this model, the rate of growth of capital stock (I/K = g) that is de-sired by capitalists is a positive function of an autonomous component (α0),3 of the difference between profit rate and interest rate (α1[r – i])4 and of a component that depends upon the external indebtedness g as a share of the capital stock (α2z

ψ).5 It is assumed that an increase in the external borrowing as a share of capital stock will result in a less than proportional increase in I/K, that is, ψ < 1.6

I

Kr i z= + − + < <α α α ψψ

0 1 2 0 1[ ] . (6)

Equation (6) assumes a conventional Keynesian hypothesis that in-vestment decisions are positively influenced by the difference between the current profit rate—which is a proxy of the expected rate of return for new planned investment (Possas, 1987)—and the nominal interest rate.7 The new element in the investment function here is the inclusion of external debt as a share of capital stock. This addition to the investment function is intended to formalize the existence of an external constraint to economic growth and capital accumulation.8 In fact, as emphasized in Bresser and Nakano (2003, p. 14), investment expenditure is restricted, in an open emerging economy, by the capacity to import capital goods from developed economies. According to this line of reasoning, external borrowing can ease the foreign exchange restriction to capital goods

3 This refers, for instance, to the capitalist’s “animal spirits.” To get an understand-ing of this concept, see Keynes (1936, ch. 12).

4 The first two components of the investment demand function are the standard formulation of investment decisions according to the so-called neo (post) Keynesian approach to growth and distribution (Marglin, 1984, pp. 81–95).

5 z = D/K is external indebtedness as a share of capital stock.6 This happens because we are assuming, in accordance with the experience of

emerging countries, that part of foreign capital flows is used to acquire nonreproduc-ible assets such as land.

7 For the sake of simplicity, we assume that the expected rate of inflation is equal to zero.

8 We introduce the external constraint to economic growth in a manner that is dif-ferent from the conventional balance-of-payments constrained growth models, such as that developed by Moreno-Brid (1998–99). In fact, Moreno-Brid analyzed the effect of capital flows on the balance-of-payments constraint, but left open the question of the determinants of effective growth rate. This means that in the long-run equilibrium, it is possible that the effective growth rate—largely determined by the desired rate of capital accumulation—is lower than the balance-of-payments equilibrium growth rate. In this case, capital accumulation and economic growth will be independent of capital flows, which seems to be in contradiction to the empirical evidence of emerging economies.


imports, stimulating capital accumulation. In Equation (6), we use the stock of external debt as a proxy for the inflow of capital to the economy at hand.9

It can also be seen in Equation (6) that the external borrowing effect on the investment rate is nonlinear. More specifically, it is assumed that an increase in external indebtedness as a share of capital stock will generate a less than proportional increase in the investment rate. This hypothesis can be justified by the idea that capital flows are not fully devoted to financing fixed capital investment, but are also used, for instance, to finance government debt or real estate. hence, the external indebtedness will not result in a proportional increase in the investment rate.

Net exports are a positive function of an autonomous component (ε0)10 and a negative function of the level of capacity use, since it is assumed that an increased level of economic activity implies increased imports, which reduces the trade balance. hence, the following equation can be written:11

E

Ku= −ε ε0 1 . (7)

The country risk premium is assumed to be endogenous and given by the following equation:

ρ = ρ0 + ρ1.z (8)

where: ρ0 > 0; e ρ1 > 0.Equation (8) shows that the country risk premium is an increasing

function of the level of the ratio of external debt to capital stock. The implicit idea in this reasoning is that the larger the external indebtedness the greater the flow of external financial commitments of the country, increasing the risk of default on external debt (Oreiro, 2004).

To complete the model, it is necessary to specify the determinants of domestic interest rate. In order to do that, it is assumed that the economy

9 This will happen if lenders in international capital markets desire a constant rate of growth for assets denominated in currencies of emerging economies. In fact, let d be the rate of growth of capital flows for a specific emerging economy. Then d = FKF/D, where FKF is foreign capital flow. This means that D = FKF/d and z = D/K = (1/d)*(FKF/K).

10 This depends upon the real exchange rate, among other variables. Since we as-sume the existence of a pegged currency regime and price rigidity at the domestic and international levels, it follows that the real exchange rate can be treated as a constant and hence be incorporated in the autonomous component of the net export function.

11 In this model, the nominal and real exchange rate are assumed to be constant and captured by the term ε0.


at hand has full convertibility of its capital account—that is, it is assumed that there is free capital mobility in the sense of Mundell (1963) and Fleming (1962). We will also assume the existence of a fixed exchange rate regime. In this setting, the nominal interest rate is determined using the uncovered interest rate parity theorem, which is given for the fol-lowing equation:

i = i* + ρ. (9)

Substituting (8) and (9) into (6), we obtain an equation that defines investment as a ratio to capital stock as a function of the external indebt-edness and profitability of the economy:

I

Kg r i z z= = + − + + +α α ρ ρ α ψ

0 1 0 1 2[ ( )] .* (10)

hence, the relationship between the growth rate of capital stock and the external indebtedness can be seen in Figure 1.

Taking the total derivative of (10), it can be shown that:

∂( )

∂= −−

IKz

zψα α ρψ2

11 1.

(11)

Solving for u in Equation (2) and substituting the resulting equation into (7), we arrive at:

Figure 1 Growth rate as a function of external indebtedness

g = I /K

z* z


E

Kr= −

+( )ε ε

ττ0 1

1. (12)

By substituting (10) and (12) into (5), we define the locus r = 0 by the following equation:

0 1 2 1 0 1 11

0= + + − + + − − − −γ α α α ρ ρ εψr z i z m r s r qa mrc( ) ,* (13)

where: γ ≡ α0 + ε0.Equation (13) shows the value of the rate of profit for which aggregate

demand is equal to the level of production. This is the short-run equilib-rium value for the profit rate.

Solving for r in Equation (13) and taking the first derivatives in regard to r and z, we obtain the following expression:

∂∂

=−( )

+ + −

•=

−r

z

m z

s m qa m mr c0

21

1 1

1 02

1

ψα α ρε α

ψ

. (14)

The sign of

∂∂

•

=

r

z r 0

will depend upon ψα2zψ–1 – α1ρ1, which varies with z. hence, as z be-

comes larger, the sign of ∂r/∂z changes from positive to negative. This behavior characterizes a nonlinear relationship between profitability and external indebtedness. It can also be observed that the threshold level of external debt as a ratio of capital stock is given by:

ψαρ α

ψ2

1 1

1

1

=−

z*.

The relationship between external indebtedness and profitability can be seen in Figure 2.

External indebtedness and multiple equilibrium: the long-run dynamics of the model

The dynamic behavior of external indebtedness is given by the following differential equation (Simonsen and cysne, 1995):

µD = ieD – H, (15)


where D is the stock of total external debt, H is the net resources transfer to abroad, and ie is the nominal interest rate of the external debt. The nominal interest rate of the external debt is given by the following equation:

ie = i* + ρ0 + ρ1z. (16)

By taking the time derivative of z, we get the following expression:

&zD

K

K

K

D

K= −

• •

, (17)

where µK/K is the growth rate of capital stock (g).12

By substituting (15) and (16) into (17), we get the following expression:

&z i g zH

Ke= − −( ) . (18)

The net foreign transfer of resources is nothing more than the net value of exports H/K = E/K.13 By substituting (10) and (12) into (18), we obtain the final expression of external debt dynamics:

z i z r i z z z m r•

−= + + − − − + + −( ) − +* *[ ( )] .ρ ρ α α ρ ρ α ε εψ0 1 0 1 0 1 2 0 1

1 (19)

Figure 2 Profitability as a function of the level of external indebtedness

12 Assuming that the capital–output rate is constant over time, it follows that the rate of real output growth is equal to the rate of growth of the capital stock.

13 We assume that the balance of nonfactor services is equal to zero.

r

z* z


Solving for r in Equation (9) and differentiating with respect to r and z, we get the locus z = 0:

∂∂

=

=+ − −

−

+

−

−r

zz

z z z

zm

mi

& 0 0 1 21

2

1 1

12 0

Φ Φ α ψ αε α

αε

ψ ψ

pp p pz i z i z z

zm

− + +−

+α α αε α

ψ1 1 2

1

1 12

( )

( ),

(20)

where:

Φ0 = i* + ρ0 – α0 – α1r + αi* + α1ρ0 – α1i* – ρ0α1 > 0;

Φ1 = (2ρ1 + 2α1ρ1 – 2α1ρ1) > 0

ip = i* + ρ0 + ρ1z > 0.

Equation (20) shows that the effect of an increase in external borrow-ing over the economy’s profitability depends on the level of external borrowing. Assuming that ε1 – zmα1 > 0; the sign of

∂∂

•

=

r

z z 0

will depend upon the external borrowing. For a low level of z, the slope is positive, whereas for high levels of z, the slope is negative. One of the possible configurations of the locus at hand is shown in Figure 3.

In the steady-state equilibrium, the profit rate and the external indebt-edness are constant over time. This allows us to define the locus r = 0 and the locus z = 0; their slopes are given, respectively, in Equations (14) and (20).

hence, as given in Equations (14) and (20), it can easily be shown that one of the possible configurations of the long-run equilibrium of the economy under study corresponds to what can be seen in Figure 4.

Figure 4 shows the existence of two long-run equilibrium positions. The first one is characterized by the existence of a high profitability (r1) and a low level of external indebtedness (z1). We will call this position “equilibrium with low external indebtedness.” The second position is characterized by a situation of low profitability (r2) and a high level of external indebtedness (z2), which we will call “equilibrium with high external indebtedness.” From this result we can conclude that capital inflows may cause an excessive level of external indebtedness (equilib-rium with a high level of external indebtedness) with a negative effect on profit rates and the degree of capacity utilization, which determines a


situation of economic stagnation. This happens because in this model, an equilibrium in which the profit rate is low configures a situation where the degree of productive capacity utilization is also low due to the supposed constancy of the rate of markup. Since the capital accumulation rate is a positive function of the degree of productive capacity utilization and a negative function—beyond a critical level of external borrowing z*—of external indebtedness, the growth rate of capital stock will be lower than the one that would prevail if the economy were operating with a lower level of external indebtedness.

Figure 3 Equilibrium locus of external debt

Figure 4 Long-run multiple equilibrium

z* z

r

r

r1

r2

z1 z2 zµz = 0

µr = 0


Empirical evidence

We have seen in the theoretical model that the foreign capital inflow can, through accumulating external indebtedness, have negative consequences for performance growth in emerging countries.

The aim of this section is to test the hypothesis that the external debt has any effect on the macroeconomic performance of emerging countries, based on panel analysis of a sample set of data for fifty-five emerging countries,14 according to the World Bank classification of countries by income, covering the period 1980–2000. The sample is constituted by low-middle income and high-middle income countries. The data comes from Penn World Tables of heston et al. (2006) and from the series of World Bank’s World Development Indicators.

Empirical model

The idea here is to build a demand-led empirical growth model in a Keynesian/Kaldorian fashion. In this framework, growth is determined by the investment rate, which, in turn, is dependent not only upon the dynamicity and competitive capacity of the export sector of the economy, but also upon the level of external debt due to the effects presented in the theoretical model. The external dynamicity here is a proxy for the real exchange rate behavior of emerging countries.

To create an index of the competitive capacity of the external sector, a real exchange rate devaluation index is estimated.15 This index is the difference between the actual values of the real exchange rate and its long-run values, which is assessed by the predicted values of a PPP (purchasing power parity) exchange rate equation, corrected for Balassa-Samuelson effects. The properties of growth stimulus of devalued real exchange rates are empirically consolidated. The real exchange devaluation index is as described in rodrik (2008). The author creates an exchange rate misalignment index corrected for Balassa-Samuelson effects. According

14 The sample is composed of Albania, Algeria, Angola, Argentina, Armenia, Belize, Botswana, Brazil, Bulgaria, cameroon, cape Verde, chile, china, colombia, congo, costa rica, croatia, cuba, Djibouti, Dominica, Dominican republic, Ecuador, El Salvador, honduras, India, Indonesia, Iran, Iraq, Jamaica, Kazaquistan, Kiribati, Latvia, Lebanon, Lesotho, Libya, Lithuania, Macedonia, Maldives Island, Mexico, Mongolia, Morocco, Palau, Paraguay, Poland, romania, russia, Saint Lucia, Serbia, South Africa, Sri Lanka, Sudan, Turkey, ukraine, uruguay, and Venezuela.

15 The theoretical vehicles of propagation of the effects of real exchange devaluation on growth can be accessed in detail in Gala (2008) and, in a more conventional fashion, in rodrik (2008).


to rodrik, this measure, which has the advantage of being comparable between countries over time, is obtained in three steps. First, we use the XrAT nominal exchange rate series from the Penn World Tables and the PPP series to calculate the real exchange rate of a country i in period t. Equation (21) shows the equation in logs:

lnθit = ln(eit/PPPit). (21)

When the real exchange rate measured is larger than 1, it indicates that the money value is lower than what is predicted by PPP. however, in practice, nontradable goods are cheaper in poorer countries (Balassa-Samuelson effect), so a correction is needed. hence, in the second step, we regress the real exchange rate over the rGDPch Penn World Tables series, as shown in Equation (22), where ft is the fixed effect for the period of time and uit is the error term.

lnθit = α + β0lnRgDPCHit + ft + uit. (22)

Finally, in the last step we construct the real exchange rate devaluation index measured according to Equation (23) as the difference between the actual exchange rate and the predicted values of the model estimated:

ln(RealDevaluation)it = lnθit – ln¬θit. (23)

This index of devaluation of the real exchange rate can be compared between countries, and it is used in the following empirical model esti-mates as a proxy variable of external competitiveness of the economy. Fixed effect panel regression estimates of the long-run exchange rate are reported in Table 1. The coefficient estimate of beta is –0.133, which suggests a relatively strong Balassa-Samuelson effect: when income increases by 10 percent, the real exchange rate appreciates by approxi-mately 1.3 percent.

The empirical growth equation to be estimated is:

ln( / ) ln( / ) ln( / )

( / ) (

Y P Y P I Y

Debt GDP Debtit it it

it

= + + ++

−α ββ β

1 1

2 3 // )† ( )

, ,

GDP CEX

OPENK SCHOOLING Desvalit it

i t i t

+ ++ + + +

ββ β ψ β ξ

4

5 6 7 iit ,

(24)

where:ln(Y/P)it is the log of GDP per capita based on the Laspeyres index in

1996 from the Penn World Tables 6.2 series;ln(Y/P)it–1 is the log of GDP per capita with the Laspeyres index in 1996

basis, lagged in one period, from the Penn World Tables 6.2 series;


ln(I/Y)it is investment as a share of GDP, from the Penn World Tables 6.2 series;

ln(Debt/gDP)it is a proxy of the foreign savings and external indebt-edness, defined as the net foreign assets position of a country with data drawn from Lane and Milesi-Ferretti (2007);

ln(CEX)it is an index of external competitiveness of a country i, built from XrAT and lnRgDPCH from the Penn World Tables 6.2 series. The construction of the index was detailed above.

OPENKit is an index of the economy’s trade openness (exports + imports/gDP), from the Penn World Tables 6.2 series;

SCHOOLINgit is a human capital index given by primary school enroll-ment, with series drawn from the World Development Indicators 2008;

Desvalit is an index of relative depreciation of the real exchange rate.ψ is a year dummies vector, with one dummy for each period covering

every five-year sequence;ξit is a stochastic residual.

Following Arellano and Bond (1991) and Blundell and Bond (2000) we used the generalized method of moments (GMM) for a dynamic panel. These estimators attempt to deal with unobservable temporal effects by including time-specific intercepts. Dealing with these effects is not a trivial task. As a result, the model is dynamic and may include endogenous regressors that are controlled by means of the instruments of the difference variables.

The instruments that correspond to moment conditions are lagged both in terms of level and difference of the independent and dependent variables. Because moment conditions typically overidentify the model’s regression, the dynamic panel method allows for testing different specifications with Sargan’s or hansen’s test. using Arellano and Bond (1991) estimators, Blundell and Bond (2000) developed a system estimator (System-GMM)

Table 1 Long-run real exchange rate with Balassa-Samuelson correction

Lnθ Coefficient Standard error P > t

ln(RGDPCH) –0.1333 0.0503 0.0080

Constant 1.9592 0.4195 0.0000

Notes: Fixed-effects (within) regression: no. of observations = 1,195; group variable (i): countries: no. of countries = 55.


that uses additional moment conditions. The Arellano–Bond (1991) and Blundell–Bond (2000) estimators were considered suitable for the analy-sis conducted here because they enable dynamic specification (allowing a lagged dependent variable) and because they appropriately instrument potentially endogenous variables. For a more detailed description of these econometric techniques, see Baltagi (2005).

Table 2 presents the results of the empirical growth model. Specifica-tions (1)–(4) test the robustness of the coefficients to different types of variables coordination. The hansen test checks the validity of the overidentification of instruments for SYS-GMM, and is as expected. The Arellano–Bond test indicates first-order correlation, but rejects the null hypothesis for the second order.

The positive and significant coefficients for the lagged dependent variable in all specifications indicate the persistence of the GDP per capita series. The competitiveness index of exports shows a positive and significant relationship with the macroeconomic performance of the sample countries. This result may suggest that the effects of a relatively undervalued real exchange rate are an important matter for the growth of countries in the sample. The investment variable, which is crucial to growth models in the post Keynesian tradition, appears with positive and significant values for its coefficients. The “Trade openness” variable, which is traditionally incorporated in growth regression, also appears positive and significant.

The “Debt/GDP” shows a negative correlation with GDP per capita growth. The negative sign for “Debt/GDP²” signals a hump-shaped re-lationship as indicated in the theoretical model of the previous section. Both effects are confirmed for the sample countries in all specifications. however, the positive effect of external debt on economic growth will occur only for negative values of this variable.16 This means that, accord-ing to the empirical model, any positive level of external debt is harmful for the economic growth of emerging countries. Economic growth can be maximized only for a negative value of external debt—that is, only if a country becomes a net creditor in international markets. This requires running persistent current account surpluses.

This result is in opposition to the results presented in the literature about external debt and growth. In fact, this literature showed that a posi-tive, although small, level of external debt can have positive effects on economic growth. Our empirical model showed that this happens only

16 consider the following equation: y = a – bx – cx2. It can easily be shown that dy/dx > 0 if and only if x < – (b/2c).


for negative values of external debt. But it is important to note that this result is not incompatible with the structure of our theoretical model. In fact, this result can be derived from Equation (10) of the theoretical model if we set ψ = 2 and α2 < 0, which makes the threshold level of external debt, z*, negative.

Table 2 Growth equations

(1) (2) (3) (4)

Ln(GDP/Per Capita) 0.683*** 0.669*** 0.553*** 0.630***

(0.0109) (0.0138) (0.0443) (0.0407)

Debt/GDP –0.0103*** –0.0283*** — —

(0.00394) (0.00203) — —

Debt/GDP² –0.007*** — — —

(0.0016) — — —

Trade openness 0.0020*** 0.0028*** 0.0023*** —

(9.65e–05) (8.19e–05) (0.0002) —

Government expenditure –0.0097*** –0.010*** 0.0023*** —

(0.0006) (0.0004) (0.0002) —

Schooling 0.00230*** 0.00110*** 0.00478* 0.0071***

(0.0001) (0.0001) (0.002) (0.0021)

ln(Investment) 0.00426 0.00411** 0.0793*** 0.0662***

(0.0031) (0.0019) (0.0136) (0.014)

Real devaluation 0.887*** 0.826*** 0.800*** 0.827***

(0.207) (0.142) (0.143) (0.200)

Constant 2.505*** 2.690*** 2.965*** 2.247***

(0.100) (0.126) (0.546) (0.468)

AR(1) test 0.0100 0.0070 0.0070 0.0020

AR(2) test 0.782 0.551 0.576 0.559

Hansen test 0.902 0.314 0.513 0.456

Observations 551 551 551 805

Countries 55 55 55 75

Notes: Standard errors are given in parentheses. ***, **, and * denote significant at the 1 per-cent, 5 percent, and 10 percent levels, respectively. Debt/GDP, Debt/GDP², the index of competitiveness of exports and schooling are treated as endogenous variables in SYS-GMM. Additional moments conditions were used with the level instrument variables foreign direct investment and consumer price index, which prove to be good instruments for capital account variables in several trial specifications.


Final remarks

This paper presented a nonlinear post Keynesian growth model in order to evaluate, at theoretical and empirical levels, the relationship between external indebtedness and economic growth in emerging countries. The main theoretical result of the model is the existence of two long-run equilibrium positions: one has a high level of external indebtedness (as a ratio of capital stock) and a low profit rate, and the other has a low level of external indebtedness and a high profit rate. This means that “excessive” external indebtedness can result in stagnant growth due to its negative effect on the rate of profit.

To test the effects of external indebtedness on the rate of economic growth in emerging economies, a dynamic panel is estimated to evaluate whether external debt has an effective negative influence on economic growth in emerging countries. The empirical test of demand-led growth equations with a dynamic panel for fifty-five emerging countries confirms the potential negative effects of external debt over the long-run rate of growth of the sample countries.

In particular, a hump-shaped relation was detected between external debt and economic growth for the sample countries, but the positive effect of external debt on economic growth occurs only for negative values of this variable. This means that, according to the empirical model, any positive level of external debt is harmful for the economic growth of emerging countries. Economic growth can be maximized only for a negative value of external debt, that is, only if a country becomes a net creditor in international markets. This requires emerging countries to run persistent current account surpluses.

Since persistent current account surpluses are required to foster eco-nomic growth in emerging countries, we can conclude that capital ac-count liberalization and the attraction of foreign savings are harmful for growth performance in these countries.

REFERENCES

Arellano, A., and Bond, B. “Some Test Specification for Panel Data: Monte carlo Evidence and an Application to Employment Equations.” Review of Economic Stud-ies, 1991, 58, 277–297.Baltagi, B. Econometric Analysis of Panel Data. chichester: Wiley, 2005.Blundell, r, and Bond, B. “A Estimation in Dynamic Panel Data Models: Improving on the Performance of The Standard GMM Estimators.” London, Institute for Fiscal Studies, 2000.Bresser, L.c, and Nakano, Y. “crescimento Econômico com Poupança Externa?” [Economic Growth with Foreign Savings?] Brazilian Journal of Political Economy, 2003, 23 (2), 3–27.


clements, B; Bhattacharya, r; and Nguyen, T. Q. “External Debt, Public Investment and Growth in Low-Income countries.” International Monetary Fund, Working Paper WP/03/249, 2003.Davidson, P. Financial Markets, Money and the Real World. Aldershot: Edward Elgar, 2002.Eaton, J. “Sovereign Debt: A Primer.” Word Bank Economic Review, 1993, 7 (2), 137–172.Fleming, J.M. “Domestic Financial Policies under Fixed and Floating Exchange rates.” International Monetary Fund Staff Papers, 1962, 9, 369–379.Gala, P. “real Exchange rate Levels and Economic Development: Theoretical Analysis and Econometric Evidence.” Cambridge Journal of Economics, 2008, 32, 273–288.heston, A.; Summers, r; and Aten, B. Penn World Tables Version 6.2. center for International comparisons of Production, Income, and Prices at the university of Pennsylvania, 2006.Kaldor, N. “Alternative Theories of Distribution.” Review of Economic Studies, 1956, 23, 83–100.———. “Equilibrium Theory and Growth Theory.” In M. Boskin (ed.), Economic and Human Welfare: Essays in Honor of Tibor Scitovsky. New York: Academic Press, 1977, pp. 273–291.———. “The role of Effective Demand in the Short- and Long-run Growth.” In A. Barrére (ed.), The Foundations of Keynesian Analysis. London: Macmillan, 1988, pp. 22–42.Keynes, J.M. The general Theory of Employment, Interest and Money. London: Macmillan, 1936.Krugman, P. “Financing vs. Forgiving a Debt Overhang.” Journal of Development Economics, 1988, 29, 253–268.Lane, P., and Milesi-Ferretti, G. “The External Wealth of Nations Mark II: revised and Extended Estimates of Foreign Assets and Liabilities.” Journal of International Economics, 2007, 73 (2), 263–294.Marglin, S. growth, Distribution and Prices. cambridge, MA: harvard university Press, 1984.Mccombie, J.S.L., and roberts, M. “The role of the Balance of Payments in Economic Growth.” In M. Setterfield (ed.), The Economics of Demand Led-growth. Aldershot: Edward Elgar, 2002, pp. 87–114.Moreno-Brid, J.c. “On capital Flows and the Balance of Payments constraint revis-ited.” Journal of Post Keynesian Economics, 1998–99 21, 283–289.Mundell, r. “capital Mobility and Stabilization Policy under Fixed and Flexible Exchange rates.” Canadian Journal of Economics and Political Science, 1963, 29, 475–485.Oreiro, J.L. “Poupança externa e performance macroeconômica” [Foreign Savings and Macroeconomic Performance]. Brazilian Journal of Political Economy, 2004, 24 (2), 185–200.Palma, G. “Three routes to Financial crisis: The Need for capital controls.” In J. Eatwell and L. Taylor (eds.), International Capital Markets. Oxford: Oxford univer-sity Press, 2002, pp. 12–44.Pattillo, c; Poirson, h; and ricci, L. “External Debt and Growth.” International Mon-etary Fund Working Paper WP/02/69, 2002.Possas, M.L. Dinâmica da Economia Capitalista [Dynamics of the capitalist Econo-my]. São Paulo: Brasiliense, 1987.rodrik, D. “real Exchange rate and Economic Growth: Theory and Evidence.” John F. Kennedy School of Government, harvard university, Draft, July 2008.


rodrik, D., and Velasco, A. “Short Term capital Flows.” NBEr Working paper no. 7364, 1999.Sachs, J. “The Debt Overhang of Developing countries.” In G. calvo et al. (eds.), Debt Stabilization and Development: Essays in Memory of Carlos Diaz Alejandro. Oxford: Basil Blackwell, 1989, pp. 80–200.Schclarek, A., and ramon-Ballester, F. “External Debt and Economic Growth in Latin America.” 2005 (available at www.cbaeconomia.com/Debt-latin.pdf).Simonsen, M.h., and cysne, r.P. Macroeconomia [Macroeconomics]. São Paulo: Atlas, 1995.Taylor, L. Macroeconomia Estruturalista [Structuralist Macroeconomics]. cidade do México: Trillas, 1989.Thirwall, A.P. “The Balance of Payments constraint as an Explanation of Internation-al Growth rate Differences.” Banca Nazionale del Lavoro Quarterly Review, 1979, 128, 45–53.———. The Nature of Economic growth. Aldershot: Edward Elgar, 2002.

Copyright of Journal of Post Keynesian Economics is the property of M.E. Sharpe Inc. and itscontent may not be copied or emailed to multiple sites or posted to a listserv without thecopyright holder's express written permission. However, users may print, download, or emailarticles for individual use.

Documents

buun