Financial Deepening, Trade Openness and Growth: a Multivariate Cointegrated Analysis of the Complementary Effects.

By Sergio Ginebri*, Giacomo Petrioli** and Laura Sabani*
Abstract

A two-way causal relationship between growth and financial development is captured in a number of recent theoretical and empirical works. The basic idea is that the development of financial intermediaries allows the allocation of savings into productive investment opportunities with higher growth-generating potentials. In the same time wealthier economies can afford the costs associated to more sophisticated financial systems. As for the relationship between trade and growth, the effects of trade liberalisation on development seem to depend on the hypothesis about the particular engine of growth at work .In this paper, differently from the bulk of the existing literature, we take the view that financial markets and international good markets are interdependent and an isolated analysis of each market in its effects on growth would impede a clear identifications of the links between trade openness, financial deepening and economic growth. Our starting hypothesis is that potential gains to growth of trade liberalisation depend on the degree to which financial markets and international trade in goods act as complements. In order to verify empirically such hypothesis, in this paper the relationship between financial development, trade openness and economic growth is examined in a vector autoregressive (VAR) framework using two sample countries (Italy and Spain). By estimating and identifying the cointegrating vectors in those two countries, we ascertained the existence of a complementary relationship between financial development and trade openness. Furthermore, by the identification of a structural VAR and the analysis of the Impulse Response Functions, we were able to detect the complementary relation not only in the long period but also in the short-run. In effect, when a positive exogenous shock is given to trade openness, the financial system receives an impulse to grow as well. This means that trade liberalisation not only may encourage growth directly but also could have an “indirect“ effect by fostering financial development with a growth effect of its own.

  • *Università degli Studi di Roma “La Sapienza” , Dipartimento di Economia Pubblica, e CIDEI
  • ** Università di Venezia “Ca’ Foscari” , Master Program

Keywords: endogenous growth, trade openness, financial development, multivariate cointegration analysis.

E-mail: , , .

Financial Deepening, Trade Openness and Growth: a Multivariate Cointegrated Analysis of the Complementary Effects.

1. Introduction

Over the last decade much theoretical and empirical research has been directed to achieve a better comprehension of the effects on welfare levels and rates of GDP growth of trade openness and of financial development and integration[1].

Theoretically the role of global linkages and of financial depth in promoting economic development cannot be properly explained in the context of the standard neoclassical model of growth (Solow ,1965). According to this model technological progress, that is not explained but exogenously assumed, is the unique factor driving growth in the long run

Endogenous growth models provide a more rigorous and convincing conceptual framework for analysing how international trade and capital market development and integration promote growth.

First of all by introducing increasing return to scale of physical capital and/or human capital accumulation[2] into an otherwise neoclassical framework, they show how an economy can sustain indefinite growth in per capita income even in the absence of technological progress ( Rebelo ,1991; Lucas,1988; Azariadis and Drazen,1990). Thus, it becomes theoretically possible to shed light on the determinants of long run growth by looking at the role of trade openness and financial development in affecting the process of accumulation of reproducible factors.

Focusing the attention on human capital accumulation[3], it is possible to think that marginal benefits of the investment in human capital might be increased by demand expansion through foreign market access (scale effects) and that the inflow of new ideas through trade stimulates investment in human capital by increasing its efficiency (Lucas, 1993). As for the role of efficient financial markets, De Gregorio (1996) stresses the fact that, by providing liquidity, they loosen borrowing constraints and help investment in schooling. Differently, Cooley and Smith (1995) argue that efficient financial markets might promote entry in entrepreneurial activity and then human capital accumulation through learning by doing.

A second strand of this new growth theory (Romer, 1990; Grossman Helpman, 1991 and Rivera-Batiz Romer, 1991) has extended the neoclassical model by incorporating market-driven innovations and thus allowing for endogenously driven growth. In these models the engine of growth is technological progress whose level is decided by the investment in R&D activity.

The analysis of the channels through which trade openness and financial development affect investment in innovating activities will thus give important insights on growth. The efficiency of technological knowledge, for example, can be seen as an increasing function of openness due to knowledge spillovers from foreign countries (Grossman Helpman, 1991). Moreover, demand expansion through trade liberalisation encourages the development of innovations and thus the R&D activity (Romer ,1990). Finally, as far as R&D is a risky activity it becomes possible to evaluate the effects on the investment in new and highly risky projects of the presence of efficient and possibly fully integrated financial markets: effective financial institutions allow to improve the process of collecting information on the efficiency of investment projects and /or entrepreneurs’ ability (Greenwood Jovanovic, 1990) and permit, via financial integration, a more adequate diversification of risk (Saint Paul, 1992; Feeney,1994) promoting investment into high yield, highly risky activities.

Against this background, an important point must be stressed: the bulk of this “new growth” literature looks at the growth effects of trade liberalisation and development and/or integration of financial markets separately. However, in our opinion, an isolated analysis of each market in its effects on growth might prevent a clear understanding of the process of economic development by overlooking the analysis of the multicausal linkages among trade openness, financial development and economic growth. In this respect an integrated analysis is needed.

Multicausal linkages among trade, economic growth and financial development emerge from the evidence that not only financial development positively affects growth but the extent of financial activity itself depends positively on growth[4]. The reason is that the cost of financial services carries a fix component that falls with the volume of financial transactions. As a consequence financial markets will develop only when a threshold level of income is reached. But, if financial outcomes are endogenous to an economy there is a question of how greater trade integration affects the state of financial development itself .

Theoretically the question has not been much explored. An exception is Blackburn and Hung (1998): they employ the well-known endogenous growth model of Romer (1990) in which growth is driven by horizontal innovation in intermediate goods. The model embodies a scale effect: by expanding the markets for new goods, through trade liberalisation, for example, horizontal innovations in intermediate goods is encouraged. This means that more firms enter the research sector and more firms look for external financing of risky and independent research projects. This, in turn, helps financial intermediaries to better diversify their portfolios and decreases their default probability reducing the agency cost related to the need for depositors to monitor the intermediary portfolio. The reduction in the agency cost of financial intermediation contributes to higher growth since firms in the research sector start operating at positive profits and this encourages new firms to enter the market increasing the rate at which new process are invented. This in an indirect “financial market” gain from trade: through scale effects trade liberalisation can accelerate innovations and the development of financial markets. Hence, there exists a complementary relationship between trade and financial development .

Feeney (1994a, b) tackles a related question in that she asks whether or not integration in financial markets complements trade. She argues that in an open economy the availability of international financial markets for risk-sharing eliminates uncertainty in aggregate income, due to sector specific productivity shocks, and allows product specialization in the direction indicated by comparative advantage. A higher degree of specialization will be associated to financial integration causing the expected volume of trade to rise. This complementary relationship, however, depends on the initial (autarchic) structure of production and it is possible to identify situations in which international financial assets and goods markets behave instead as substitutes[5].

The object of this paper is to verify empirically the existence of the complementary relationship between trade openness and the extent of financial activity[6]. Focusing the attention on the Italian and the Spanish economies we test such hypothesis by adopting an aggregate time series approach covering the period 1977-1994 for Italy and 1971-96 for Spain. We employ a multivariate cointegrated model and by the identification of a structural VAR and the analysis of the Impulse Response Functions, we are able to detect the complementary relation not only in the long period but also in the short-run. Moreover our result seem to point towards the evidence of a direct nexus between trade and the extent of financial activity. More precisely, while in the Blackburn –Hung modeltrade affects the extent of financial activity through its effect on growth, the results we obtained show the existence of a direct positive link.

The rest of the paper is organised as follows. Section 2 sets out the theoretical background on which we found the empirical analysis; empirical results are presented in section 3 and finally in section 4 results are summarised and policy implications discussed.

2.Theoretical background.

In order to recognize multicausal linkages among trade, financial deepening and economic growth we base our empirical analysis on a simplified version of the paper by Blackburn and Hung (1998) (B-H, hereafter) and on an open economy version of the paper by Cooley and Smith (1995).

The B-H model is a more complex version of Romer (1990): they add financial markets by hypothesising that the design producers need external finance to start their business.

There are two production sectors : the final good and the producers goods sector.

Let us refer to the following production function of the final good

where j indexes the different types of capital goods that can be used in production and Atcaptures the number of capital goods that have been invented as of time t. Ly,t is the amount of labour employed in the final sector.

In each period there are Nt firms engaged in research activity. Once developed a blueprint (or design) shows how to produce the new capital good: one unit of the final good at t gives one unit of the new capital good at t+1 , kj,t+1 [7] .

The success probability of a research project is assumed to be positively related to the stock of knowledge accumulated At and to the amount of labour employed, so the expected flow of new designs can be written as:

The steady state equilibrium can be identified in terms of the constant ratios

where the common constant growth rate of variables is given by

Growth occurs through the increasing variety of capital goods associated with an increasing number of firms engaged in research and development.

Now turn to the issue of trade liberalisation. Consider two economies initially isolated and symmetric. Assume that the designs for these inputs are non tradable so that the production takes place in the original country. Assume that there is no imitation and no redundancy in research. The number of intermediate goods available to firms in each country is now 2At. Designers begin to operate at positive expected profits given the reduction in the labour cost of developing a new design. This encourages new firms to enter research sector intensifying the rate of innovation.

Now let us consider the financial sector. Each designer has got zero wealth and must raise external finance. We assume indirect lending through a financial intermediary. The cost of delegation is represented by the cost of monitoring the intermediary operate. This cost decreases as the probability of bank failure decreases. It is possible to show that the probability of failure tends to zero as the number of independent projects financed by the intermediary grows without bound. Since the direct effect of trade is increasing the number of designers, this means that the delegated monitoring cost will be reduced. This, in turn, reduces the fixed cost in research due to the need to find external finance, promoting growth by encouraging new firms to enter the research sector.

In sum, the B-H model predicts that trade has got a positive effect on growth by encouraging product development, then product development fosters financial deepening through a reduction in the agency cost of external financing. Therefore, trade affects the extent of financial activity indirectly through its effect on growth .

We believe, however, that the existence of such “indirect” nexus does not exclude the presence of a direct link between trade and financial development as we argue in the following basing our discussion on an open economy version of the model by Cooley and Smith (1995).

In an open economy human capital accumulation is influenced by the level of knowledge accumulated not only internally but also in the rest of the world, due to the fact that the ideas developed abroad spread internally through trade in commodities (Parente and Prescott,1991). Thus, human capital accumulation can be represented by the following equation

where W (E) if the level of worldwide knowledge “imported” which is a function of the degree of external openness E. In other words, trade increases the marginal benefits of the investment in human capital since the efficiency of technology is increased by knowledge spillovers. If we intend human capital accumulation as learning by doing we can conclude that trade encourages entrepreneurial development since it is plausible to think that learning by doing is most significant in young enterprises ( Cooley and Smith, 1995). To start their business young entrepreneurs need external finance: we can argue, therefore, for the existence of a direct nexus between trade and the extent of financial activity. This argument leads us to a further observation. According to the Cooley and Smith (1995) model , even when financial markets are free to form (no fixed costs) it is possible to observe equilibria without financial activity . This happens when the rate of return from financial investment is too low. Therefore, by increasing the payoff to financing young entrepreneurs, trade might foster the formation of active capital markets avoiding “low growth trap” equilibria.

4. The results of the empirical analysis

The relationship between financial development, trade openness and economic growth is examined in a vector autoregressive (VAR) framework. We focus our attention on two European countries which have been interested to a process of economic integration with the rest of the world: Spain and Italy. Our empirical analysis aims at shedding light on the relationship among three macro-economic variables in each country: the level of real Gross domestic product (GDP), the level of total credit to the private sector (CRED) and a measure of trade openness of the country, represented by the ratio between the sum of exports and imports and GDP (TRADE)[8]. We first present the results in the case of Italy.

Results on Italian data

We started the analysis on Italian data by estimating a Vector AutoRegressive system in five variables. To the three variables previously mentioned, two other ones were added: the real effective exchange rate (REER) and a measure of international trade openness, represented by the ratio between the sum of exports and imports and the GDP in the whole OECD area (INTRADE)[9].

Data are quarterly; estimation period goes from 1977:4 to 1994:3; following the evidence from Hannan-Quinn’s and Schwartz’ information criteria only one lag was included in the estimated system.

Once estimated the VAR system the cointegration analysis was carried out and the Johansen’s (1995) approach was followed[10]. The cointegration rank was established according to the trace test (see table 1) and just one cointegrating vector was detected. The estimated long run relationship is the following:

TRADE = 0.8816 INTRADE + 0.0793 CRED

Hence, we found empirical evidence on Italian data of the complementary relationship between openness to international trade and development of financial system which we were looking for on the basis of theoretical reasoning.

Italian GDP is not present in the cointegrating equation, but the null of hypothesis of weak exogeneity of GDP is rejected. In other words, growth is affected by the long run relation between international integration and development of the financial system, however the complementary relationship is not directly affected by growth.

The exam of the impulse response functions (IRF) gives us the opportunity to have a better understanding of the interrelations among growth, trade openness and development of financial system. The simulation of the IRFs requires the identification of structural shocks, that is of disturbances idiosyncratic to each endogenous variable in the estimated system but orthogonal among them. Such identification was achieved by imposing arbitrary restrictions on the contemporary links among the endogenous variables. We assumed that international trade openness (INTRADE) and total credit to private sector (CRED) are affected in the short period, i.e. at each time, only by their own, idiosyncratic random shocks. Italian trade openness(TRADE) is affected by either its own random shock and by the shock to international trade openness. Similarly, the real effective exchange rate (REER) is simultaneously influenced by its own random shock and by the shock to Italian trade openness. Finally, all the random shocks in the system impinge on the short run behaviour of Italian GDP.