ArticlePDF Available

Do External Trade Promote Financial Development?

August 2005

August 2005

Source
RePEc

Authors:

Yongfu Huang

United Nations University (UNU)

Jonathan Temple

Several recent papers have argued that trade and financial development may be linked, either for political economy reasons, or because foreign competition and exposure to shocks lead to changes in the demand for external finance. In this paper we use the cross-country and time-series variation in openness to study the relationship between trade and finance in more detail. Our results suggest that increases in goods market openness are typically followed by sustained increases in financial depth.

The increase in openness

…

Trade and finance

…

Partial scatter plots

…

. Descriptive Statistics: 1990-2001 =============================================================== A. Summary Statistics for Openness and financial development measures

…

. External trade and financial development (whole sample), 1990-2001

…

Figures - uploaded by Yongfu Huang

Content may be subject to copyright.

Content uploaded by Yongfu Huang

Content may be subject to copyright.

Does external trade

promote financial development?

Yongfu Huang

Jonathan Temple

Discussion Paper No. 05/575

July 2005

Department of Economics

University of Bristol

8 Woodland Road

Bristol BS8 1TN

Does external trade

pr omote ﬁnancial developm ent?

∗

Yo ngfu Hu ang

Department of Economics, University of Bristol

8 Woodland Road, Bristol BS8 1TN, UK

Jonathan Temple

†

Department of Economics, University of Bristol

8 Woodland Road, Bristol BS8 1TN, UK and CEPR

July 7, 2005

Abstract

Several recent papers hav e argued that trade and ﬁnancial development may

be linked, either for political economy reasons, or because foreign competition

and exposure to shocks lead to changes in the demand for external ﬁnance. In

this paper we use the cross-country and time-series variation in openness to study

the relationship between trade and ﬁnance in more detail. Our results suggest that

increases in g oods market openness are typically followed by sustained increases

in ﬁnancial depth.

Keywords: Openness, Trade, Financial Development

JEL Classiﬁcation: F13; O16

∗

We thank Stephen Bond, Edmund Cannon, David Demery, David Winter, seminar participants in

Bristol, and especially Chris Bowdler for helpful comments and discussion. The usual disclaimer ap-

plies. Temple also thanks the Leverhulme Trust for ﬁnancial support under the Philip Leverhulme Prize

Fellowship scheme.

†

Corresponding author. Telephone +44 117 928 8430. Fax number +44 117 928 8577. Email

jon.temple@bristol.ac.uk

1 Introduction

Many of the world’s economies are becoming more open to international trade. Not

surprisingly, this has led to renewed interest in the consequences of openness for com-

petition, technology transfer, and productivity. In this paper, we ask whether openness

promotes ﬁnancial dev elopment. We will use cross-section methods to examine the

long-run relationship, and recently dev eloped panel data methods to examine Granger-

causality. Our panel data estimates suggest that, for lower -income countries, increases

in goods market openness are typically follo wed by sustained increases in ﬁnancial

dev elopment.

Trade and ﬁnance have been connected in the literature in at least two dif ferent

ways, which can be loosely characterized as supply-side and demand-side. In an im-

portant paper , Rajan and Zingales (2003) emphasize the supply-side role of interest

groups, and especially the vested interests of incumbent industrialists and ﬁnancial in-

termediaries. Incumbents, worried by the threat of entry, hav e strong incentives to resist

ﬁnancial development. These incentives are weakened if a country becomes more open

to foreign competition or to international ﬂowsofcapital. Inthisview,goodsmarket

openness can i mprove the supply of external ﬁnance, because it aligns the interests of

the economically powerful more closely with ﬁnancial development.

In contrast, Sv aleryd and Vlachos (2002) emphasize the role of risk diversiﬁcation.

To the extent that openness is associated with greater risks, such as increased expo-

sure to external demand shocks or foreign competition, it will create new demands for

external ﬁnance. Firms will need credit in order to overcome short-term cashﬂow prob-

lems and adverse shocks. In this view, the effects of trade on ﬁnance are likely to work

primarily through the demand side.

These long-run effects may be complemented by other, more short-lived consider-

ations. Although our empirical work examines the long-run relationship between trade

and ﬁnance, it acknowledges that short-run effects will also play a role. For example,

when developing countries liberalize trade on a major scale, such as Mexico under the

NAFTA, restructuring and investment are likely to increase the demand for external

ﬁnance. If trade liberalizations are followed by investment and lending booms, there

could be a strong association between openness and ﬁnance in the short run, whether

or not there is an association in the long run.

Giavazzi and Tabellini (2004) ﬁnd that

trade liberalizations are indeed often followed by higher inv e stment.

The Rajan and Zingales view assigns importance to ﬁnancial openness as well as goods market open-

ness. We discuss the role of ﬁnancial openness in more detail in section 2 below.

One mechanism here may be information asymmetries. To the extent that performance in export

markets is a useful and observable index of a company’s productivity, increased openness can reduce

information asymmetries among different banks. Dell’Ariccia and Marquez (2005) present a formal model

in which reduced information asymmetries can generate a lending boom.

When examining the effects of external trade on ﬁnancial depth, disentangling

cause and effect will not be straightforward, partly because ﬁnance may inﬂuence trade

as well as vice versa. Our empirical work will address this problem in two ways. First,

we follow Rajan and Zingales (2003) and Stulz and Williamson (2003) in using the

measure of “natural openness” due to Frankel and Romer (1999). We ﬁnd evidence

that openness and ﬁnance are strongly associated for higher-income countries, but not

for lower-income. Perhaps contrary to some of the ideas put forward in Rajan and Zin-

gales, we ﬁnd that openness has a much stronger association with bank-based ﬁnance

than with stock market de velopment.

Our second approach, which forms the heart of the paper, is to use panel data meth-

ods. These allow us to exploit the substantial time series variation in goods market

openness. For example, in a subset of countries, openness has risen quite sharply. To

show this we use a standard measure of openness, namely imports plus exports, rel-

ative to GDP, all in current prices. For a sample of 82 countries, the median extent

of openness rose from 43% in 1960-64 to 66% in 1995-99. There are 59 countries in

which openness rose by at least ten percentage points over this period, and 34 in which

it rose by at least 25 percentage points. To show this pattern in more detail, ﬁgure 1

is a smoothed (kernel density) plot of the distributions for 1960-64 and 1995-99. The

rightward shift in the distribution of openness ov er the forty-year period is clear.

Panel data methods allow us to examine the consequences of increased openness for

ﬁnancial dev elopment, and to see whether timing patterns are consistent with a causal

effect. In a forty-year data set for 88 countries, we examine whether changes in open-

ness precede (Granger-cause) changes in ﬁnancial depth. From the estimated models,

we ﬁnd strong effects of openness on ﬁnancial development in the whole sample, and

for lower-income countries, but not for higher income countries. There is some evi-

dence that these effects persist into the long run, and do not simply reﬂect temporary

booms in bank lending.

It is important to note that, for low er -income countries, the panel data results in-

dicate stronger positiv e effects of trade than we ﬁnd in the cross-section. Increases in

openness are systematically associated with subsequent increases in ﬁnancial depth. It

is possible that the (conﬂicting) cross-section estimates are contaminated by the pres-

ence of individual ef fects that are correlated with openness. For the higher-income

countries, the situation is rev ersed: the panel data results are weaker than the cross-

section. The imprecise panel data estimates may arise partly because of the small

cross-section dimension of the higher-income panel.

Note that this shift is not universal, and not all countries are more open now than twenty or thirty

years ago. See Dollar and Kraay (2004) for more details on trends in openness. Their analysis conﬁrms

that the time series variation in the trade share can be substantial, helping to motivate our emphasis on a

panel data approach.

The remainder of the paper is organized as follo ws. Section 2 discusses the litera-

ture on ﬁnancial dev elopment in more detail. Section 3 describes the data sources, and

the measures of ﬁnancial development that we adopt. Section 4 focuses on the cross-

section methods and results. In section 5, we discuss the methods and tests used in the

panel data analysis. Section 6 reports the panel data estimates, and can be seen as the

heart of the paper. In section 7 we brieﬂy summarize our conclusions.

2Tradeandﬁnance

In this section, we sketch the theoretical background to the paper, and expand on the

arguments in the introduction. We discuss the links between ﬁnance and growth, the

role of trade and other determinants of ﬁnancial development, and the possibility that

ﬁnance, in turn, can inﬂuence the extent and structure of trade.

Conventionally, trade is seen as a way to beneﬁt from specialization and scale

economies. But if external trade promotes ﬁnancial development, this offers a more

complex route by which trade may raise productivity and living standards. Beginning

with King and Levine (1993), the study of ﬁnance and growth has ﬂourished, and the

evidence that ﬁnancial sector dev elopment plays a role in growth is increasingly per-

suasive; Rajan and Zingales (1998) is one inﬂuential contribution. The literature on

ﬁnance and growth is reviewed by Levine (1997, 2005).

There has been less work on the determinants of ﬁnancial development, and our

work contributes to this emerging line of research. The central question is why en-

trepreneurs and ﬁrms appear to have easier access to external ﬁnance in some countries

than others. As discussed in the introduction, Rajan and Zingales (2003) argue that

such differences arise because of political economy considerations.

Incumbent indus-

trial ﬁrms, and perhaps domestic ﬁnancial intermediaries, will wish to block entry to

their sectors: they have a direct interest in maintaining an underdeveloped ﬁnancial sec-

tor. These incentiv es may weakened by openness, ho w ever. A combination of foreign

competition through external trade, and openness to capital ﬂo w s, will tend to align the

interests of incumbents more closely with ﬁnancial dev elopment. Although Rajan and

Zingales give most emphasis to arms-length ﬁnance (especially stock markets) some of

the arguments can be applied to bank-based external ﬁnance as well.

In support of their arguments, Rajan and Zingales document surprising long-term

swings in ﬁnance, using data for 1913 and decades since. Some countries have seen

long-term declines in the importance of external ﬁnance relative to GDP, before a resur-

gence in the 1980s and 1990s. An attractive feature of the Rajan and Zingales theory

For a general discussion of the political economy of ﬁnancial development, see Pagano and Volpin

(2001a).

is that it can explain these rev ersals in political economy terms. Openness to trade was

relatively lo w in the wake of the Great Depression, and ev en the adv anced industrial

economies maintained capital controls for sev eral decades after 1945, motiv ating the

familiar point that the economy was less globalized in mid-century than at the begin-

ning and end of the twentieth century. These reductions in openness, perhaps helping to

create a political constituency opposed to ﬁnancial sector development, could explain

the U-shaped path that has been taken by the importance of external ﬁnance relative to

GDP.

Other mechanisms that could link trade and ﬁnance do not rely as strongly on po-

litical economy, and give more emphasis to the demand side. As discussed in the in-

troduction, Svaleryd and Vlachos (2002) argue that openness may be associated with

greater risks, including exposure to external shocks and foreign competition. This will

encourage the development of ﬁnancial markets that can be used to diversify such risks,

and that allow ﬁrms to overcome short-term cashﬂow problems or adverse shocks.

Openness and ﬁnancial de velopment may also be linked in simpler ways. The

cross-country study of Levine and Renelt (1992) identiﬁed a robust correlation between

openness and the share of investment in GDP, and if trading economies are also high

investment economies, this could promote ﬁnancial dev elopment. Openness may also

inﬂuence the demand for external ﬁnance through the nature of specialization and sec-

toral structure, or through the pace of innovation and technology transfer , activities that

are likely to make intensive use of external ﬁnance.

Our central focus is goods market openness, but previous research also points to

the importance of ﬁnancial openness: for example, the e xtent of restrictions on equity

o wnership by foreign investors. Alessandria and Qian (2005) use a formal model to

argue that capital account liberalization has ambiguous effects on the efﬁciency of do-

mestic ﬁnancial intermediaries, while Chinn and Ito (2005) and Law and Demetriades

(2005) consider the empirical evidence. There are at least three good reasons for com-

plementing these studies with a consideration of goods market openness. First, goods

market openness will be more relevant for the many developing countries where domes-

tic equity markets are absent or small in size. Second, there is often richer time series

v ariation in goods market openness than in measures of capital controls.

Third, goods

A more ambitious argument is that political forces lead either to a ‘corporatist’ political settlement,

with low in vestor protection and high employment protection, or a non-corporatist (Anglo-Saxon) outcome

with ﬂe xible labour markets and greater ﬁnancial depth (Pagano and Volpin 2001b). Commentary in

the media often assumes, perhaps wrongly, that increased trade and globalization has made a corporatist

settlement harder to sustain, providing another way to link trade and ﬁnance.

The evidence for this should not be overstated. Using stock market data for e merging markets, Li

et al. (2004) ﬁnd that goods market openness is associated with greater market-wide variation, but not

greater ﬁrm-speciﬁc variation.

The availability of long spans of data on goods market openness lends itself to conventional panel

data methods. In contrast, given the discrete nature of many capital account liberalizations, event studies

market openness and ﬁnancial openness are not independent. The empirical w ork of

Aizenman and Noy (2003), Aizenman (2004) and Chinn and Ito (2005) suggests that

capital account liberalization is often preceded by goods market openness, perhaps be-

cause trade integration makes restrictions on capital ﬂows harder to sustain.

As background to our work, we now consider other determinants of ﬁnancial depth.

Se veral contrib utions have examined the effects of the legal, regulatory and macroeco-

nomic en vironment on the functioning of the banking sector and equity markets. Most

prominently, La Porta et al. (1998) argue that the origins of the legal code are im-

portant for ﬁnancial development, because legal s ystems differ in their treatment of

creditors and shareholders, and in contract enforcement. La Porta et al. argue that the

English common law tradition protects the rights of minority shareholders and credi-

tors, while the French civil code is associated with less efﬁcient contract enforcement,

weaker investor protection and possibly higher corruption. Countries with German or

Scandinavian legal origins are said to have intermediate levels of investor protection

and contract enforcement.

Whether or not the origins of the legal code m atter, gov ernment regulation clearly

plays a strong role. The starting point of the arguments in Rajan and Zingales (2003)

is that gov ernment regulation is needed to ensure ef fective contract enforcement, and

transparency in accounting and disclosure. Regulations concerning information dis-

closure, accountings standards, permissible practice of banks and deposit insurance do

appear to have material effects on ﬁnancial development (Mayer and Sussman 2001).

Less central to the current paper , the macroeconomic environment may also be

relevant. Huybens and Smith (1999) examine theoretically, and Boyd, Levine and

Smith (2001) empirically, the effects of inﬂation on ﬁnancial depth. They conclude

that economies with higher inﬂation rates are likely to have smaller, less activ e, and

less efﬁcient banks and equity markets.

Finally, it is worth pointing out that ﬁnancial development can, in turn, inﬂuence

the structure and extent of trade. Two papers by Beck (2002, 2003) have examined this

issue in detail. Drawing on the arguments of Kletzer and Bardhan (1987), Beck (2002)

develops a model in which countries with better-de v eloped ﬁnancial markets will tend

to have a comparative advantage in manufacturing.

Using a 30-year panel with 65

countries, he shows that ﬁnancial depth is associated with a higher level of manufactur-

ing exports and a higher trade balance in manufactured goods. In a companion paper,

may be the best approach in that context, as in Henry (2003).

As with any innovative line of research, this is not without its critics. Looking at historical data,

Rajan and Zingales (2003) ﬁnd that common law legal codes have not always been associated with greater

ﬁnancial dev elopment. They argue that law could play a stronger role in ﬁ ltering the eff ects of interest

groups and incentiv es than in inﬂuencing the overall level of ﬁnancial development.

Recent work by Ju and Wei (2005) and Wynne (2005) also develops connections between ﬁnancial

development and trade specialization.

Beck (2003) shows that economies with greater ﬁnancial depth ha v e higher manufac-

turing export shares and higher trade balances in industries that rely more on external

ﬁnance. These results help to motiv ate our Granger-causality approach: since ﬁnance

may inﬂuence trade, as well as vice versa, evidence on timing becomes especially im-

portant.

3Thedata

This section describes the data and variables that we will use in our analysis. Section

3.1 describes the data on openness and GDP. In section 3.2, we will outline some of the

widely used measures of ﬁnancial development. Our empirical work will combine the

standard measures into aggregate indices of ﬁnancial de velopment, an approach that we

describe in section 3.3. Deﬁnitions of the main variables we use, and information on

data sources, can be found in Table 1.

3.1 Basic data

Our cross-section sample relates to the period 1990-2001. We have excluded countries

with populations of less than 500,000 in 1990, using population data from the Penn

World Table (PWT).

We also exclude transition economies from the sample.

To measure initial GDP in our regressions, we use real chain-weighted GDP per

capita (denoted RGDPCH in PWT 6.1) a veraged ov er 1988-90 to lessen the effect of

temporary measurement errors and departures from trend. The data on openness are

also from PWT 6.1. Openness is measured as the sum of exports and imports divided

by GDP, in either current prices (OPENC) or constant international prices (OPENK),

averaged ov er 1988-1990. Following Svaleryd and Vlachos (2002), we always exclude

Hong Kong and Singapore, two city-states for which openness is unusually high, re-

ﬂecting transit trade and production that involves a high import content (for e xample,

electronics). We also exclude other countries for which measured openness exceeds

150%.

In order to capture the geographic predisposition to trade, we use the measure of

natural openness due to Frankel and Romer (1999) and call this CTRADE, for con-

structed trade share. Information on the country of legal origin, and classiﬁcations by

income levels and export specialization, are obtained from the Global Development

Network (GDN) Growth Database. To keep sample sizes reasonably large, our empir-

ical work aggregates the GDN income classes into two larger groups: lo w er -income

countries made up of low-income and lower-middle income countries, and higher-

income countries made up of high-income and upper-middle income countries.

The exact source is version 6.1 due to Heston et al. (2002).

3.2 Standard measures of ﬁnancial development

A number of measures of ﬁnancial development have been used in recent work. Our

basicdataonﬁnancial development are from the Financial Structure Database intro-

duced by Beck et al. (2000). We average standard indicators over 1990-2001, omitting

any observations for which fewer than three years of data available. We no w sum-

marize in more detail the available indicators, drawing on Demirguc-Kunt and Le vine

(1999) in particular. We start with indicators of bank-based ﬁnancial depth, which play

the dominant role in our empirical work, and then turn to indicators of stock market

dev elopment.

For measuring overall ﬁnancial development, an especially popular measure is the

ratio of liquid liabilities to GDP, based on the liquid liabilities of the ﬁnancial system

(currency plus demand and interest-bearing liabilities of bank and nonbank ﬁnancial

intermediaries). This measure has been used by McKinnon (1973) and King and Levine

(1993) among others; we denote this measure LLY, as is standard in the literature. Other

widely used measures include the ratio to GDP of credit issued to the private sector by

banks and other ﬁnancial intermediaries (denoted PRIV O) and the ratio of commercial

bank assets to the sum of commercial bank assets and central bank assets (denoted

BTOT). LLY, PRIVO and BTOT are highly correlated. They will be used to construct the

main indicator of ﬁnancial depth in our later panel data analysis, given the availability

of long spans of data for these measures.

The Beck et al. (2000) database also includes two measures of the efﬁciency of

ﬁnancial intermediation. The variable OVC is the ratio of ov erhead costs to total bank

assets. In the short run, high overhead costs may be related to in vestments by compet-

itive banks in improving ﬁnancial services, b ut over longer time spans, high overhead

costs are likely to reﬂect inefﬁciency and a lack of competition. A second measure, the

NetInterestMarginorNIM, equals the difference between bank interest income and

interest expenses, di vided by total assets. Again, high values for this variable tend to

suggest a lack of competition among banks.

Recently, some studies have taken a wider view of ﬁnancial de v elopment. Le vine

and Zervos (1998) discuss the independent effects of banks and stock markets on eco-

nomic growth, rather than focusing on simply the extent of intermediation. Some mea-

sures of stock market development, such as stock market turnover, can be seen as in-

dices of ﬁnancial sector efﬁciency or sophistication rather than simply ﬁnancial depth.

As well as a standard measure, market capitalization relativ e to GDP (MCAP), Levine

and Zervos (1998) use Total Value Traded (TVT) as an indicator of stock market activ-

ity. This is the ratio of trades in domestic shares (on domestic exchanges) to GDP, and

can be used to gauge market liquidity relative to the size of the economy. Finally, the

T urnover Ratio (TOR) is the ratio of trades in domestic shares to market capitalization.

High values for TOR indicate a more activ e equity market, which may be associated

with a relativ ely ef ﬁcient allocation of capital.

3.3 New measures of ﬁnancial development

The indicators described in the abov e section are standard. Since they are all intended

as proxies for an underlying, latent variable - ﬁnancial development - there may be sig-

niﬁcant advantages in combining them. This could help to alleviate measurement errors

and outlier problems that might be associated with the use of a single indicator. We take

one of the simplest possible approaches, namely to use principal components analysis.

The method takes N speciﬁc indicators and yields new indices (the principal compo-

nents) P

, ...P

that are mutually uncorrelated and so capture different dimensions

of the data. In our w ork we use solely the ﬁrst principal component. Formally, this

is deﬁned by a vector of weights a =(a

, ..., a

)

on the (standardized) indica-

tors X =(X

,...X

)

such that a

X has the maximum variance for any possible

weights, subject to the constraint that a

a =1.

We use this method to aggregate dif ferent sets of components into ﬁve new mea-

sures of ﬁnancial depth. Their structure can be seen from Table 2, together with the

weights on each component.

Our ﬁrst aggregate measure is designed to capture over -

all ﬁnancial dev elopment, and is denoted FD. This is based on the complete set of eight

components, namely LLY, PRIVO , BTOT, OVC, NIM, MCAP, TVT and TOR.Theﬁrst

principal component accounts for 63% of the variation in these eight indicators. Here

and subsequently, all the weights have the expected signs: positiv e for all variables

except OVC and NIM, given that high values for these latter two variables indicate in-

efﬁciency in the ﬁnancial sector.

Our second measure, FDSIZE, is effectiv ely the av erage of LLY and MCAP,and

provides a summary of the combined importance of bank-based and equity-based ﬁ-

nance, relative to GDP. In contrast, FDEFF isdesignedtocaptureﬁnancial efﬁciency,

and is based on OVC, NIM, TVT and TOR.Afourthmeasure,FDBANK,capturesthe

extent of bank-based intermediation. This uses LLY, PRIVO, BTOT, OVC and NIM,and

accounts for 71% of the variation in these ﬁve indicators. FDSTOCK captures equity

market development, and accounts for 86% of the variation in MCAP, TVT and TOR.

Finally, a measure of ﬁnancial depth, FDEPTH,usesonlyLLY, PRIVO and BTOT,and

accounts for 74% of the variation in these indicators.

By construction, all the new indices have a mean of zero. Panel A of Table 3 pro-

vides some other descriptive statistics for the new indicators and their components. In

panel C of Table 3, we present the correlations among the ne w measures. As expected,

All components are measured in natural logarithms, which helps to reduce outlier problems in this

particular application, and are then standardized to have mean zero and unit standard deviation.

they are highly correlated with one another. It is interesting to note that by far the

lo west correlation is between FDBANK and FDSTOCK (0.63), reﬂecting the tendency

for intermediation to be either bank-based or equity-based. Another interesting aspect

of Table 3 is that the correlations between openness and the new indicators, shown in

Panel C, are noticeably higher than when the original proxies are used (Panel B). This is

consistent with the idea that aggregation of the original measures has reduced measure-

ment error . The correlations with openness are still low, but these simple associations

do not control for the level of development and other inﬂuences on ﬁnancial depth.

We no w take a look at scatter plots of the trade-ﬁnance relationship. Figure 2 plots

two measures of ﬁnancial development, LLY and FD, against openness. Openness and

LLY are measured in natural logarithms, which helps to reduce outlier problems in this

case. In ﬁgure 2, there is some evidence that ﬁnancial development and openness are

positively related, although outliers such as Japan (coded JPN in the ﬁgure) and El

Salvador (SLV) may obscure the relationship to some extent. In ﬁgure 3, we present

partial scatter plots between the same sets of v ariables, after conditioning on initial GDP

per capita, and legal origin dummies (for English, French and German legal origin).

Again, there is some evidence of a positive association between ﬁnancial development

and openness, whether using LLY or the additional information contained in the new

indicator FD.

4 Cross-section: methods and results

4.1 Cross-section methods

Our basic framework will be regressions of the form:

Finance = α + β log(OP ENC)+f(controls)+ε (1)

Here Finance is the lev el of ﬁnancial dev elopment, based on either individual

ﬁnancial indicators, or our new aggregate indices, measured over 1990-2001. The ex-

planatory variables are the natural logarithm of openness, and some controls. As in

Rajan and Zingales (2003) we condition on the log of real GDP per capita, to control

for the demand for ﬁnance. Our other explanatory v ariables are dummies for country of

legal origin, as in La Porta et al. (1998), Beck et al. (2003) and Berko witz et al. (2003).

Figures 2 and 3 suggested that inﬂuential outliers could play an important role in

any cross-section analysis. To increase robustness in this dimension, we estimate our

models in two stages. We ﬁrst apply median or least absolute deviation regression,

in which the parameters are chosen to minimise the sum of the absolute values of the

residuals, rather than the sum of their squares as in OLS. We then exclude any ob-

servations for which the LAD residual is more than two standard deviations from the

mean residual, before estimating the model by OLS or GMM. This procedure is not

perfect, but should help to exclude the worst outliers, including some that would not be

identiﬁed by more conventional OLS diagnostics.

Once outliers have been excluded, we run straightforward regressions using OLS.

Since there are man y reasons that openness could be correlated with the error term in

such a regression - including reverse causality, measurement error and omitted vari-

ables - we supplement OLS with IV procedures. We treat openness as an endogenous

explanatory variable, and use the Frankel-Romer (1999) trade share as an instrument.

Their work models openness using the size of the domestic population (giv en that large

countries trade a lower fraction of GDP internationally) and the proximity of large ex-

ternal markets. Our maintained assumption will be that this measure of “natural open-

ness” affects ﬁnancial development through external trade, but is uncorrelated with the

error term in the structural equation. This rules out, among other things, any inﬂuences

of geography on ﬁnancial development that act through routes other than openness.

For the Frankel-Romer measure of natural openness to be a good instrument, it must

be correlated with openness, after conditioning on other (exogenous) v ariables. We

investigate instrument relevance using a F-test on the excluded instrument in the ﬁrst-

stage regression of 2SLS (the reduced form regression of openness on all the exogenous

v ariables, including natural openness). These tests suggest that natural openness is not

a weak instrument, consistent with the ﬁndings in Frankel and Romer (1999).

We use 2SLS to construct diagnostic tests, including a test for heteroskedasticity

based on Pagan and Hall (1983). Compared to OLS, the use of IV procedures is as-

sociated with a loss of efﬁciency if openness is uncorrelated with the error term. We

therefore implement two tests for endogeneity. The ﬁrst is a standard Wu-Hausman test

based on auxiliary regressions in the 2SLS case.

The second is a difference-Sargan

type test, sometimes called the C statistic in the GMM context. Not surprisingly, the

results are close to those of the auxiliary regression approach. Both tests indicate that

openness is endogenous in many of the models we consider.

4.2 Cross-section results

This section presents cross-country regression results. To anticipate some of our main

ﬁndings, partitioning the sample into a higher-income group and a lower-income group

will be an important step. Openness is associated with ﬁnancial development in higher-

This follows similar analyses in Rajan and Zingales (2003) and Stulz and Williamson (2003), among

others.

See Davidson and MacKinnon (1993, p. 237-242) for details of this approach.

income countries, but the evidence for this is much weaker in the low er -income sample.

We ﬁnd this when using the individual ﬁnance measures and also when using the new

aggregate indicators.

We present our ﬁrst set of results in Table 4. We report the point estimates and

heteroskedasticity-robust p-values only for openness; the estimates for initial GDP and

the legal origin dummies are not reported, to save space. The relationship between

openness and ﬁnance is weak for the sample as a whole. In the higher-income group of

countries, however, Panel A of Table 5 shows that openness and bank-based ﬁnance are

strongly associated. This result emerges whether using OLS or GMM, and also when

using individual ﬁnance measures like LLY and PRIVO (results not reported).

In Panel B of Table 5, we consider the lower-income group. Here, the GMM re-

sults suggest that, if anything, the more open countries are less ﬁnancially de v eloped,

although these results are fragile. The R

fortheOLSregressionsismuchlowerin

the lower -income sample than in the higher-income sample, suggesting that these re-

gressions omit important determinants of ﬁnancial depth. One possibility is that, even

conditional on the level of income, countries that specialize in primary commodities are

likely to have weak ﬁnancial de velopment, relative to countries where manufacturing

activity has a greater role. This effect may be especially pronounced for economies

dominated by point-source resources like oil or diamonds, since oil and mineral ex-

traction is often carried out by the state or multinationals, implying less demand for

external ﬁnance at any given level of income. We have explored this effect using export

classiﬁcations from the GDN database, and a dummy variable for exporters of point-

source resources, based on the work of Isham e t al. (2002). Financial development

is signiﬁcantly lower for primary commodity exporters (results available on request)

but controlling for this effect, or splitting the sample by export classiﬁcation, does not

strengthen the estimated effect of openness.

We hav e considered the robustness of these ﬁndings in a number of other dimen-

sions. One simple test is to replace the trade share in current prices (OPENC) with the

trade share evaluated at prices that are constant across countries (OPENK). This makes

little difference to the results, tending to weaken the ef fect of openness only slightly.

We have also experimented with aggregate measures of ﬁnancial development in which

the individual components are not in logarithms. Again, the results are very similar to

those shown in Tables 4 and 5.

In summary, these regressions indicate that external trade is associated with ﬁnan-

cial dev elopment in richer countries. These regressions have relatively high explanatory

power. The trade effect is robust to treating openness as endogenous, and emerges for

many different indicators of ﬁnancial development. Perhaps contrary to some of the

arguments in Rajan and Zingales (2003), the trade effect is stronger for bank-based ﬁ-

nance than it is for equity-based ﬁnance. For poorer countries, the picture is a great

deal more ambiguous. We ﬁnd scant evidence that trade and ﬁnance are positively as-

sociated. The GMM estimates indicate that the association may even be ne gati ve, but

there is also evidence that important determinants of ﬁnancial development have been

omitted.

5 Panel data methods

Cross-section methods are simple and easy to interpret, but have some important weak-

nesses. Relationships may be artiﬁcially created or obscured by unobserved hetero-

geneity and outliers. The use of panel data can ov ercome these problems, and has other

advantages. We can look at relationships over time, to see whether increases in open-

ness are followed by increases in ﬁnancial depth, and to distinguish between short-run

and long-run effects. In principle, we can also use the time series v ariation to obtain

more precise estimates of the parameters of interest.

For our panel data analysis, we will use 88 countries. The data are ﬁve-year av-

erages over the period 1960-99, giving a maximum of 8 cross-sections per country,

although the panel is unbalanced. As in the empirical gro wth literature, averaging the

data over ﬁve-year intervals means that the results are less likely to be driven by co-

mov ements at very short horizons. Averaging is also likely to lessen the impact of

measurement error, and simpliﬁes the speciﬁcation of the dynamics of the model. We

will see that, even using ﬁve-year averages, we need an AR(2) rather than an AR(1)

model to capture the dynamics adequately. This implies that a model based on annual

datawouldbelikelytorequiremanyparameters.

To measure ﬁnancial development for as many countries as possible, we drop the

indicators for which data are incomplete, and construct a measure FDEPTH. As before,

this is the ﬁrst principal component of LLY, PRIVO,andBTOT, but now the loadings

are based on the pooled cross-section time series data for the individual measures. As

before, our openness data are from version 6.1 of the Penn World Table.

We will carry out a panel data version of a Granger-causality test. Given that causal-

ity may run in either direction, we cannot treat openness as strictly exogenous. Instead,

we estimate partial adjustment models that allow feedback, using sequential moment

conditions to identify the model. We provide a brief introduction to this approach,

which is covered in more detail in Arellano (2003, chapter 8).

We start with the sim-

ple AR(1) case, and then consider the AR(2) model that seems more appropriate for

For a complementary approach that does make use of annual data, but for a smaller number of coun-

tries, see Law and Demetriades (2005).

We draw on some aspects of Arellano’s presentation, including his notation. Another useful introduc-

tion can be found in Bond (2002), and a briefer overview in Durlauf et al. (2005).

this application.

The most common approach in the empirical literature would be to specify an

AR(1) model of the form:

= α

it−1

+ β

it−1

+ η

+ φ

+ v

| α

|< 1 (2)

i =1, 2, ...88 and t =2...8

where in our application y

willbeameasureofﬁnancial development and x

will be openness. The model allows for unobserved heterogeneity through the indi-

vidual ef fect η

capturing the combined ef fect of time-invariant omitted variables. φ

is a common time effect, while v

is the disturbance term. We assume that x

is po-

tentially correlated with the individual effect η

and may be correlated with v

,but

is uncorrelated with future shocks v

it+1,

it+2,

... Under these assumptions, x

it−1

predetermined with respect to v

and the errors can be assumed to satisfy sequential

moment conditions of the form

E(v

| y

t−1

,η

,φ

)=0 (3)

where y

t−1

=(y

...., y

i,t−1

)

and x

t−1

=(x

...., x

i,t−1

)

When these moment conditions are satisﬁed, the transient errors v

are condition-

ally serially uncorrelated, for any j>0

E(v

it−j

| y

t−1

,η

,φ

)=0 (4)

and this implies (by the law of iterated expectations) that

E(v

it−j

)=0 (5)

Under these assumptions, the model can be estimated by ﬁrst-differencing equation

(2) to eliminate the individual effects, and then using lagged lev els of y

and x

dated

t − 2 as instruments. But a more efﬁcient GMM estimator can be obtained by using

more of the available moment conditions, as proposed by Arellano and Bond (1991).

They suggest using all available lagged levels of x

and y

dated t − 2 and earlier, so

that the relevant moment conditions take the form

E(y

it−s

∆v

)=0 s ≥ 2; t =3,..,8

E(x

it−s

∆v

)=0 s ≥ 2; t =3,..,8 (6)

(There are also six moment conditions associated with the period-speciﬁcconstants,

which we omit for simplicity.)

We call this estimator DIF-GMM. Giv en the strict assumptions needed for identiﬁ-

cation, it is important that speciﬁcation tests are applied to the estimated models. First

of all, we use the standard Arellano and Bond (1991) tests for serial correlation in the

ﬁrst-differenced errors. We expect to ﬁnd ﬁrst-order serial correlation in the differenced

errors, but second-order serial correlation would imply that (4), and therefore some of

the moment conditions, are in valid. We can also use a Sargan-type test to assess the

model speciﬁcation and o veridentifying restrictions (also known in the GMM context

as Hansen’s J test).

A number of limitations of DIF-GMM should be noted. A well-known property of

two-step GMM estimators is that the standard errors may be severely biased downwards

in small samples. To address this problem, we adopt the W indmeijer (2005) ﬁnite

sample correction to the standard errors. All reported standard errors and test statistics

are heteroskedasticity-robust.

Importantly, when using DIF-GMM, we also experiment with restricted instrument

sets. This can help to avoid the overﬁtting biases that are sometimes associated with

using all the available (linear) moment conditions. Throughout the paper, our ‘reduced’

instrument sets use no lags dated further back than t − 4. Using fewer moment condi-

tions can also help to improv e the power of Sargan-type tests; see for example Bowsher

(2002).

A perhaps more fundamental weakness of DIF-GMM is that lagged levels of the se-

ries may be weak instruments for ﬁrst differences, especially when the series are highly

persistent, or the v ariance of the individual effects (η

) is high relative to the variance

of the transient shocks (v

). In this case, identiﬁcation of the model can sometimes

be improved by making additional assumptions on the initial conditions of the process.

Under assumptions developed in Arellano and Bov er (1995) and Blundell and Bond

(1998), the “system GMM” estimator can be used to alleviate the weak instruments

problem.

This estimator (which we call SYS-GMM) adds a system of equations in levels

to that in ﬁrst differences. To achieve identiﬁcation, the lagged ﬁrst-differences of

the series (y

) dated t − 1 are used as instruments in the untransformed (lev els)

equations. The additional moment conditions are

E[∆y

it−1

(η

+ v

)] = 0 t =3, .., 8 (7)

E[∆x

it−1

(η

+ v

)] = 0 t =3, .., 8 (8)

Note that, given these moment conditions, differences lagged two periods or more

are then redundant as instruments for the levels equations, because the corresponding

moment conditions are linear combinations of those already in use.

The simulation results in Blundell and Bond (1998) suggest that the combined or

system GMM estimator is more robust than difference GMM to weak instrument bi-

ases, and this method has become increasingly popular in the cross-country empirical

literature. Note that the v alidity of the additional moment conditions used by this esti-

mator (or a subset of them) can be tested using an incremental Sargan statistic. Also, in

implementing SYS-GMM, we will again restrict the instrument set, to avoid overﬁtting

biases. As before, in the transformed equations, we use only lagged levels at dates t−2,

t − 3, and t − 4 as instruments.

Our later empirical work will suggest that the AR(1) speciﬁcation is invalid for this

particular application. We therefore mov e to an AR(2) model, with two lags of the

dependent variable, and two lags of openness. This model is given by:

= α

it−1

+ α

it−2

+ β

it−1

+ β

it−2

+ η

+ φ

+ v

i =1, 2,...88 and t =3...8 (9)

and so the ﬁrst-differenced equation is:

∆y

= α

∆y

it−1

+ α

∆y

it−2

+ β

∆x

it−1

+ β

∆x

it−2

+ φ

− φ

t−1

+ ∆v

i =1, 2,...88 and t =4...8 (10)

No w the relevant moment conditions for DIF-GMM are:

E(y

it−s

∆v

)=0 s ≥ 2; t =4,..,8 (11)

E(x

it−s

∆v

)=0 s ≥ 2; t =4,..,8

Note that, perhaps surprisingly, we can continue to use moment conditions based on

the lagged levels y

it−2

and x

it−2

for the equations in ﬁrst differences, even when y

it−2

and x

it−2

are included in the untransformed model. Arellano (2003, section 6.7) is an

example of this approach. To see how this works, it can alternatively be interpreted

as exploiting the exogeneity of ∆y

it−2

and ∆x

it−2

in the ﬁrst-differenced equations,

together with the use of lagged lev els dated t − 3 and earlier:

E(∆y

it−2

∆v

)=0 t =4, .., 8

E(∆x

it−2

∆v

)=0 t =4, .., 8

E(y

it−s

∆v

)=0 s ≥ 3; t =4,..,8

E(x

it−s

∆v

)=0 s ≥ 3; t =4,..,8

But clearly these moment conditions are linear combinations of the set (11). This

alternative, equivalent way of writing down the moment conditions helps to clarify that,

in the AR(2) model, there is a sense in which identiﬁcation relies on lagged levels dated

t−3 and earlier. (This can be seen more explicitly by considering how a 2SLS approach

would be implemented.) Since these longer lags may be weak instruments, this again

suggests the potential usefulness of the system GMM estimator in this context.

In the AR(2) model, one hypothesis of economic interest is the joint null β

= β

0, which can be interpreted as a panel data test for Granger-causality. Although a Wald-

type test of this restriction could be implemented, we use an alternative approach. This

is to estimate both the unrestricted and the restricted models using the same moment

conditions, and compare their (two-step) Sargan statistics using an incremental Sargan

test of the form:

= n(J(˜γ) − J(ˆγ))

where J(˜γ) is the minimized GMM criterion for the restricted model, J(ˆγ) for

the unrestricted model, and n is the number of observations. Under the null, D

asymptotically distributed as χ

where r is the number of restrictions. The intuition

for the test is that, if the parameter restrictions are valid, the moment conditions should

remain valid even in the restricted model. F or more details, see Bond et al. (2001a) and

Bond and Windmeijer (2005).

We no w turn to some additional issues of interpretation raised by the use of an

AR(2) model. First of all, we may be interested in the stability of the estimated model.

For stability we require the roots of the relevant lag polynomial, namely the bracketed

term on the left-hand side of:

1 − α

L − α

= β

it−1

+ β

it−2

+ η

+ φ

+ v

to lie outside the unit circle. This can easily be checked by either dynamic simula-

tion or direct calculation; see for example Hamilton (1994, p. 17-18) on the latter . The

majority of our estimated models are stable, the main exceptions arising when pooled

OLS is used rather than our preferred ﬁxed-effects methods.

If the model is stable, we can calculate a point estimate for the long-run ef fect of

openness on ﬁnancial development:

+ β

(1 − α

− α

)

We can estimate an approximate standard error for this long-run effect using the

delta method. Ho wever, we should note that the long-run effect, as a nonlinear function

of the model parameters, may be imprecisely estimated. In particular , if the sum α

+α

is close to one and imprecisely estimated, the data can appear consistent with very high

v alues for the long-run effect, because the ratio quickly blows up as α

+α

nears unity.

Since the conﬁdence interval generated by the delta method is, perhaps mistakenly,

assumed to be symmetric around the point estimate, it may well overlap zero too often.

For this reason, we complement the delta method with a test of the restriction

+ β

=0. If this restriction is rejected, it suggests that there is evidence for a

long-run effect of openness on ﬁnancial development. If the restriction holds and the

parameters are non-zero (β

= −β

6=0)thenﬁnancial development depends on the

change in openness, and not on its level. This would be consistent with a story in which

increases in openness, through restructuring and in v estment, lead to a short-term boom

in lending that does not persist into the long run. Again, we will test this restriction on

the coefﬁcients using a criterion-based approach.

In some of our estimated models, the lag polynomial associated with the depen-

dent variable has one root close to unity, indicating a high degree of persistence. This

makes it especially relevant to ask how the estimation methods will perform when the

data-generating process is characterized by a high degree of persistence. The GMM es-

timators will remain consistent, because the relevant asymptotics here are for large N,

ﬁxed T .Aqualiﬁcation is that in the case of an exact unit root, the moment conditions

used in DIF-GMM are no longer sufﬁcient to identify the parameters. Identiﬁcation

may require the use of mean stationarity assumptions, as in the system GMM estima-

tor; see Arellano (2003, p. 116) for the AR(1) case.

Finally, we can test for unobserved heterogeneity using a procedure originally sug-

gested by Holtz-Eakin (1988). In the absence of individual ef fects, the following ad-

ditional moment conditions become valid, corresponding to the use of lagged levels as

instruments in the levels equations:

E[y

it−1

− α

it−1

− α

it−2

− β

it−1

− β

it−2

− φ

)] = 0 (12)

E[x

it−1

− α

it−1

− α

it−2

− β

it−1

− β

it−2

− φ

)] = 0

t =3, .., 8

The validity of these additional moment conditions, which can be tested using an

incremental Sargan test relative to difference or system GMM, then becomes a simple

test for the presence of unobserved heterogeneity (where the null is no heterogeneity).

One motivation for using this test is that, if individual effects are not present, pooled

OLS will be a consistent estimator, although not fully efﬁcient given the likely presence

of heteroskedasticity.

6Paneldataresults

This section presents the results of our panel data analysis. The main ﬁndings can

be summarized as follows. In the short run (here, 5-10 years) increases in openness

are followed by increases in ﬁnancial depth. For the whole sample, and a sample of

lo wer -income countries, this result is robust across different estimation methods, and

to v ariation in the choice of moment conditions. Not surprisingly, the long-run effect is

estimated less precisely, but in most cases we can reject the restriction that β

+β

=0.

This suggests that the effect of openness on ﬁnancial development persists into the long

run, although the extent of support for this hypothesis varies with the precise choice of

moment conditions.

We will consider OLS estimates, Within Groups (WG) estimates, and two versions

of difference GMM. All models include a full set of time dummies. The ﬁrst version

of DIF-GMM uses all a vailable linear moment conditions, while the second does not

use any lags dated further back than t − 4. We will also consider three versions of sys-

tem GMM, the ﬁrst using both sets of moment conditions (7) and (8). The second two

versions use separately either (7), which we call SYS-GMM-1, or (8), which we call

SYS-GMM-2. This approach helps to avoid overﬁtting and also reﬂects the possibil-

ity that the system GMM assumptions may be incorrect. For example, if the countries

with relatively small individual effects (and hence tending to be less developed ﬁnan-

cially) are also the countries in which openness has increased the fastest, the moment

conditions in (8) would be in valid.

First of all, we consider an AR(1) model of the form often studied in the cross-

country empirical literature. These results are shown in Table 6.

The table provides

se veral reasons to believ e this model is badly mis-speciﬁed. One source for concern

is the evidence for second-order serial correlation. In the case of system GMM, in the

ﬁnal column, the Sargan statistic also suggests that either the model speciﬁcation or the

moment conditions are in valid.

With these problems in mind, we do not give further consideration to AR(1) models.

We turn instead to AR(2) models of the form:

All our panel data estimates are obtained using the xtabond2 package for Stata, written by David

Roodman. We were able to obtain almost identical results and test statistics using the DPD for PcGive

software described in Doornik et al. (2002).

More subtly, it is worth noting that the autoregressive parameter estimated by DIF-GMM or SYS-

GMM does not lie in the interval deﬁned by the OLS and WG estimates. Since the OLS and WG estimates

of the autoregressive parameter should be biased in opposite directions, a consistent estimate of this pa-

rameter might be expected to lie between these two extremes (see Bond et al. 2001b for more discussion).

Again, this tends to call into question the validity of the moment conditions, and may also hint at weak

instrument biases.

= α

it−1

+ α

it−2

+ β

it−1

+ β

it−2

+ η

+ φ

+ v

i =1, 2,...88 and t =3...8

We will be especially interested in whether the lev el of ﬁnancial development de-

pends primarily on the change in openness, or the le vel. To examine this, we will test

the restriction that β

+ β

=0,inwhichcaseβ

it−1

+ β

it−2

= β

∆x

it−1

.We

will see that the degree of support for this interpretation is limited. It is always strongest

when pooled OLS is used, typically less strong under system GMM, and weakest under

difference GMM. In general, when we allow for ﬁxed effects, we ﬁnd evidence that the

effects of openness persist into the long run.

Table 7 presents results for the whole sample of 88 countries.

Recall that, when-

ever we apply the GMM estimators, we report tests for serial correlation (ﬁrst- and

second-order), the Sargan statistic, a test for Granger causality (β

= β

=0), a test

for a long-run effect (β

+ β

=0) and the implied point estimate and approximate

standard error of the long-run effect. We report the tests of restrictions on parameters,

and long-run effects, for OLS and Within Groups also. In the OLS and WG cases, the

tests of restrictions are based on con ventional Wald tests.

Looking at Table 7, the ﬁrst point to note is that, across all the estimation methods,

we ﬁnd strong evidence that increases in openness are associated with increases in

ﬁnancial depth in the short-run. The coef ﬁcient on the ﬁrst lag of openness is positively

signed and signiﬁcantly different from zero across all se v en models reported in Table

7. The effect of the second lag of openness varies more across the models. In the

system GMM estimates, especially, we ﬁnd evidence that the coefﬁcient on the second

lag of openness is negatively signed, which points to the importance of specifying the

dynamics carefully. The Granger-causality test rejects the null of non-causality at the

5% lev el for all seven models.

The long-run effect is less clear. In the ﬁnal two rows of the Table, we report the

point estimate and approximate standard error associated with the long-run effect. In

the case of the WG estimates, and the two varieties of DIF-GMM estimates, we ﬁnd a

stable long-run effect, and one that is signiﬁcantly different from zero at the 5% level

in columns 2 and 4. When we turn t o the different versions of SYS-GMM (columns

5-7) the long-run effect is less precisely estimated.

Note, howev er, that whenev er we

Note that, following our earlier discussion, it may again b e tempting to compare the ﬁrst autoregres-

sive parameter in the GMM estimates with that obtained under WG and OLS; but this would not have the

same justiﬁcation in the AR(2) model as in the AR(1) model.

This arises partly because of the signiﬁcantly negative coefﬁcient on the second lag of openness

(corresponding to β

), which makes a zero long-run effect harder to reject, and partly because α

+α

≈ 1

in these models, which tends to blow up the ratio of coefﬁcients that forms the long-run effect.

allow for ﬁxed effects, the restriction that β

+ β

=0is rejected at the 20% level, and

typically at the 5% level. This suggests that the effect of greater openness persists into

the long run.

Our simple heterogeneity tests reject the null of no individual effects in column

7, and are close to doing so in the other two cases. This suggests that, as expected,

unobserved heterogeneity is present, and the pooled OLS estimates are likely to be

inconsistent. An alternative way to address this problem is to look for samples that

are more likely to be homogenous. For this reason, we follo w our earlier cross-section

analysis, and split the sample into a higher-income group (Table 8) and a lower-income

group (Table 9).

We consider the higher-income group ﬁrst, in Table 8. In this group of countries,

theevidenceforasigniﬁcant short-run ef fect is much weaker than before. We typically

cannot reject the null hypothesis that the two lags of openness are jointly insigniﬁcant,

or the alternative null that β

+ β

=0. Consistent with this pattern, the long-run

effects are imprecisely estimated for all estimation methods other than Within Groups.

At ﬁrst glance, these ﬁndings might seem to go against our earlier OLS and GMM

cross-section results, in which openness and ﬁnancial development were positively as-

sociated in the higher-income sample. One reason for weaker results in the higher-

income panel may be that its cross-section dimension (N =35) is unusually small for

an application of GMM. It is interesting to note that our test for individual effects does

not reject the null of a common intercept at conventional levels. If taken at face value,

the lack of v ariation in the individual effects in the higher-income group implies that

panel data methods are not required. Simple cross-section or pooled OLS estimates

may provide a reasonable estimate of the long-run effect, and have the advantage that

they retain the information in the “between” variation.

The results for the lower-income group are shown in Table 9. In this sample, we

ﬁnd much stronger evidence that increased openness leads to greater ﬁnancial depth in

the short run. This effect is signiﬁcant at the 10% level in all sev en models. Again we

ﬁnd evidence that the coefﬁcient on the second lag of openness is negativ ely signed,

and can see that it is hard to estimate the long-run effect precisely in a sample of this

size, with the exception of the SYS-GMM-2 estimates in column 7. But the restriction

that β

+ β

=0is usually either rejected, or is close to being rejected.

The diagnostic tests support the identifying assumptions, and overall we conclude

that there is strong evidence that changes in openness are followed by changes in ﬁnan-

cial depth, which persist at least over the medium term. In the whole sample, or the

subset of poorer countries, there is fairly strong e vidence that the ef fect persists into the

long run. In the group of higher-income countries, the estimates are less precise, which

may reﬂect the small cross-section dimension of the relev ant panel. An alternative ex-

planation is that the stronger cross-section results for the higher-income sample are an

artifact of individual effects, but we ﬁnd relatively little statistical evidence for these

individual effects in the higher-income panel.

7 Conclusions

The determinants of ﬁnancial development have become a focus of recent research.

In this paper, we have examined whether ﬁnance is inﬂuenced by external trade, per-

haps for the political economy reasons identiﬁed by Rajan and Zingales (2003), or

the risk diversiﬁcation considerations emphasized by Svaleryd and Vlachos (2002).

The ﬁrst sections of the paper emphasize cross-section results, in which the Frankel-

Romer measure of “natural openness” is used to isolate exogenous variation in open-

ness. Whether using OLS or instrumental variable procedures, we ﬁnd strong evidence

that trade promotes bank-based ﬁnancial development in higher-income countries, but

not in the lower-income group.

These ﬁndings may be contaminated by omitted variables that are correlated with

openness. The main contribution of the paper is to use panel data to examine whether

increases in openness are followed by increases in ﬁnancial de v elopment, and whether

this ef fect persists into the long run. We ﬁnd strong support for this hypothesis in the

sample as a whole, and in the lower-income group. The results are robust to several

different estimation methods, and the short-run effects in particular are not sensitive to

the precise choice of moment conditions. For the higher-income group, the panel data

results are weaker than in the cross-section.

Our results suggest that the long-run effects of trade go beyond those en visaged in

traditional models. But as with much of the empirical literature on ﬁnancial dev elop-

ment, it is not clear whether the ef fects work primarily through the demand for external

ﬁnance, or through improving the supply side, or some interaction of the two. For ex-

ample, it may be that greater openness is associated with changes in sectoral structure

that increase the demand for external ﬁnance. Alternatively, it may be that increased

exposure to foreign competition inﬂuences the supply-side more directly, as envisaged

in Rajan and Zingales (2003). Discriminating between the various explanations is a

difﬁcult task, but may be an interesting area for further research, given the ﬁndings we

present here.

Refere nces

[1] Aizenman, J. (2004). “Financial opening and development: evidence and policy

controversies, ” American Economic Review, 94(2), 65-70.

[2] Aizenman, J. and Noy, I. (2003). “Endogenous ﬁnancial openness: efﬁciency and

political economy considerations,” NBER working paper no. 10144.

[3] Alessandria, G. and J. Qian (2005), “Endogenous ﬁnancial intermediation and

real effects of capital account liberalization,” Journal of International Economics,

forthcoming.

[4] Arellano, M. and S. R. Bond (1991), “Some tests of speciﬁcation for panel data:

Monte carlo evidence and an application to employment equations”, Review of

Economic Studies, 58: 277-297.

[5] Arellano, M. (2003). Panel data econometrics. Oxford University Press, Oxford.

[6] Beck, T. (2002), “Financial dev elopment and international trade. Is there a link?”,

Journal of International Economics, 57: 107-131

[7] Beck, T. (2003), “Financial dependence and international trade”, Review of Inter-

national Economics, 11: 296-316.

[8] Beck, T., A. Demirguc-Kunt, and R. Lev ine (2000). “A new database on the

structure and dev elopment of the ﬁnancial sector ,” World Bank Economi c Revi ew,

14(3), 597-605.

[9] Beck, T., Demirguc-Kunt, A. and Levine, R. (2003). “Law, endo wments, and

ﬁnance,” Journal of Financial Economics, 70, 137-181.

[10] Berkowitz, D., K. Pistor and J.-F. Richard (2003), ”Economic de velopment, legal-

ity, and the transplant effect”, Eur opean Economic Review, 47(1), 165-95.

[11] Blundell, R. and S. Bond (1998), ”Initial conditions and moment restrictions in

dynamic panel data models”, JournalofEconometrics, 87: 115-43.

[12] Bond, S. R. (2002), ”Dynamic panel data models: a guide to micro data methods

and practice”, Portuguese Economic Journal,1:141-162.

[13] Bond, S. R., C. G. Bowsher , and F. Windmeijer (2001a), ”Criterion-based in-

ference for GMM in autoregressive panel data models”, Economics Letters,73:

379-88.

[14] Bond, S. R., A. HoefﬂerandJ.R.W.Temple(2001b),”GMMestimationofem-

pirical growth models”, CEPR discussion paper no. 3048.

[15] Bond, S. R. and Windmeijer, F. (2005). “Reliable inference for GMM estimators?

Finite sample properties of alternative test procedures in linear panel data models,”

Econometric Reviews, 24(1), 1-37.

[16] Bowsher , C. G. (2002). “On testing overidentifying restrictions in dynamic panel

data models”, Economics Letters 77, 211-220.

[17] Boyd, J. H., R. Le vine and Bruce D. Smith (2001), “The impact of inﬂation on

ﬁnancial sector performance”, Journal of Monetary Economics, 47, 221-48.

[18] Chinn, M. D. and Ito, H. (2005). “What matters for ﬁnancial development? Capi-

tal controls, institutions, and interactions,” NBER working paper no. 11370.

[19] Davidson, R. and J. G. MacKinnon (1993), Estimation and infer ence in econo-

metrics. Oxford University Press.

[20] Dell’Ariccia, G. and Marquez, R. (2005). “Lending booms and lending stan-

dards,” CEPR discussion paper no. 5095.

[21] Demirguc-Kunt, A. and R. Lev ine (1999), “Bank-based and market-based ﬁnan-

cial systems: Cross-country comparisons”, World Bank Policy Research Working

Paper No. 2143.

[22] Dollar, D. and Kraay, A. (2004). “Trade, growth, and poverty,” Economic Journal,

114, F22-F49.

[23] Doornik, J., Arellano, M. and Bond, S. (2002). “Panel data estimation using DPD

for O X ,” Manuscript, Nufﬁeld College, Oxford.

[24] Durlauf, S. N., P. A. Johnson and J. R. W. Temple (2005) “Growth econometrics”,

in P. Aghion and S. N. Durlauf (eds) Handbook of Economic Growth,North-

Holland, forthcoming.

[25] Frankel, J. A. and D. Romer (1999), “Does trade cause growth?”, American Eco-

nomic Review, 89: 379-98.

[26] Giavazzi, F. and Tabellini, G. (2004), “Economic and political liberalizations,”

CEPR discussion paper no. 4579.

[27] Hamilton, J. D. (1994). Time series analysis. Princeton University Press, Prince-

ton, Ne w Jersey.

[28] Henry, P. B. (2003). “Capital-account liberalization, the cost of capital, and eco-

nomic growth,” American Economic Review, 93(2), 91-96.

[29] Heston, A., Summers, R. and Aten, B. (2002). Penn World Table Version 6.1,

Center for International Comparisons at the University of Pennsylvania (CICUP),

October.

[30] Holtz-Eakin, D. (1988). “Testing for individual effects in autoregressive models,”

Journal of Econometrics, 39(3), 297-307.

[31] Huybens, E. and B.D. Smith (1999), “Inﬂation, ﬁnancial markets and long-run

real activity”, Journal of Monetary Economics, 43: 283-315.

[32] Isham, J., Woolcock, M., Pritchett, L. and Busby, G. (2002). “The varieties of ren-

tier experience: how natural resource export structures af fect the political econ-

omy of economic growth,” Manuscript, March.

[33] Ju, J. and Wei, S.-J. (2005). “Endow ment versus ﬁnance: a wooden barrel theory

of international trade,” CEPR discussion paper no. 5109.

[34] King, R. G. and R. Levine (1993). “Finance and growth: Schumpeter might be

right”, Quarterly Journal of Economics, 108: 717-737.

[35] Kletzer, K. and Bardhan, P. K. (1987). “Credit markets and patterns of interna-

tional trade,” Journal of Development Economics,October.

[36] La Porta, R., Lopez-de-Silanes, F., Shleifer, A., and Vishny, R.W. (1998), “Law

and ﬁnance”, Journal of Political Economy, 106: 1113-1155.

[37] Law, S. H. and Demetriades, P. (2005). “Openness, institutions and ﬁnancial de-

velopment,” University of Leicester, working paper no. 05/08.

[38] Levine, R. (1997). “Financial development and economic growth: views and

agenda”. Journal of Economic Literatur e , 35: 688-726.

[39] Levine, R. (2005). “Finance and growth: theory and evidence,” in P. Aghion and

S. N. Durlauf (eds) Handbook of Economic Growth, North-Holland, forthcoming.

[40] Levine, R., N. Loayza and T. Beck (2000). “Financial intermediation and gro wth:

causality and causes,” Journal of Monetary Economics, 46(1): 31-77.

[41] Levine, R. and Renelt, D. (1992). “A sensitivity analysis of cross-country growth

regressions,” American Economic Review, 82, 942-963.

[42] Levine, R. and S. Zervos (1998), “Stock markets, banks and economic growth”,

American Economic Review, 88: 537-58.

[43] Li, K., R. Morck, F. Yang and B. Yeung (2004). “Firm-speciﬁc variation and

openness in emerging markets,” Review of Economics and Statistics, 86(3), 658-

669.

[44] Mayer, C. and O. Sussman (2001), ”The assessment: ﬁnance, law and growth”,

Oxfor d Review of Economic Policy, 17( 4), 457-66.

[45] McKinnon, R. I. (1973). Money and capital in economc development.Washington

D.C.: Brookings Institution.

[46] Pagano, M. and Volpin, P. (2001a). “The political economy of ﬁnance,” Oxford

Review of Economic Policy, 17(4), 502-519.

[47] Pagano, M. and Volpin, P. (2001b). “The political economy of corporate gov er -

nance,” CEPR discussion paper no. 2682.

[48] Rajan, R. G. and L. Zingales (1998). “Financial dependence and growth,” Ameri-

can Economic Revie w, 88: 559-86.

[49] Rajan, R. G. and Zingales, L. (2003). “The great rev ersals: the politics of ﬁnancial

development in the twentieth century,” Journal of Financial Economies, 69, 5-50.

[50] Stulz, R. M. and Williamson, R. (2003). “Culture, openness, and ﬁnance,” Journal

of Financial Economics, 70, 313-349.

[51] Svaleryd, H. and Vlachos, J. (2002), “Markets for risk and openness to trade: how

are they related?” Journal of International Economics, 57: 369-395.

[52] Windmeijer, F. (2005). “A ﬁnite sample correction for the variance of linear efﬁ-

cient two-step GMM estimators,” Journal of Econometrics, 126(1), 25-51.

[53] Wynne, J. (2005). “Wealth as a determinant of comparativ e advantage,” American

Economic Review, 95(1), 226-254.

Density

0 .5 1 1.5 2

Openness (trade share/GDP)

1960−64

1995−99

Figure 1: The increase in openness

This figure shows kernel density plots of the distributions of average openness in

1960-64 (left) and 1995-99 (right). The figure shows the trend towards increased

openness over this period.

AGO

ARG

AUS

BDI

BEN

BFA

BGD

BOL

BRA

BWA

CAF

CAN

CHE

CHL

CHN

CIV

CMR

COG

COL

CRI

CYP

DEU

DNK

DOM

DZA

ECU

EGY

ETH

FIN

FJI

GAB

GBR

GHA

GMB

GNB

GTM

GUY

HND

HTI

IDN

IND

IRL

IRN

ISR

ITA

JAM

JOR

JPN

KEN

KOR

LKA

LSO

MAR

MDG

MEX

MLI

MOZ

MRT

MUS

MWI

MYS

NER

NGA

NIC

NLD

NOR

NPL

NZL

PAK

PAN

PER

PHL

PNG

PRY

RWA

SEN

SLE

SLV

SWE

SYR

TCD

TGO

THA

TTO

TUN

TUR

UGA

URY

USA

VEN

ZAF

ZMB

ZWE

−3 −2 −1 0 1

Log of LLY

2.5 3 3.5 4 4.5 5

Log of openness

Openness and LLY

ARG

AUS

BGD

BOL

BRA

CAN

CHE

CHL

CHN

CIV

COL

CRI

CYP

DEU

DNK

ECU

EGY

FIN

GBR

GHA

GTM

IDN

IND

IRL

IRN

ISR

ITA

JAM

JOR

JPN

KEN

KOR

LKA

MAR

MEX

MUS

MYS

NGA

NLD

NOR

NPL

NZL

PAK

PAN

PER

PHL

PRY

SLV

SWE

THA

TTO

TUN

TUR

URY

USA

VEN

ZAF

ZMB

ZWE

−6 −4 −2 0 2 4

FD index

2.5 3 3.5 4 4.5 5

Log of openness

Openness and our FD index

Figure 2: Trade and finance

These figures show scatter plots of the logarithm of LLY, and the aggregate index

FD, against the logarithm of openness.

BRA

USA

IND

ARG

BGD

JPN

RWA

AUS

PER

UGA

ETH

IRN

SLE

CHN

NPL

COL

TUR

ITA

PAK

MEX

DZA

CMR

ZAF

GBR

CAN

NZL

ZWE

GHA

BDI

GTM

SLV

URY

BFA

NER

MAR

CAF

HTI

MDG

BOL

FIN

IDN

KEN

TCD

SYR

MOZ

VEN

EGY

NGA

MLI

MWI

LKA

SWE

AGO

TTO

DEU

PHL

ECU

SEN

THA

ZMB

PRY

CIV

CHL

DNK

ISR

BEN

GNB

PAN

HND

NOR

CRI

KOR

GAB

CHE

DOM

PNG

CYP

NIC

IRL

TUN

COG

JAM

BWA

NLD

TGO

FJI

MRT

GMB

MYS

GUY

MUS

LSO

JOR

−2 −1 0 1

Log of LLY

−1.5 −1 −.5 0 .5 1

Log of openness

coef = .18732092, (robust) se = .10315694, t = 1.82

Openness and LLY (conditional)

BRA

USA

ARG

IND

JPN

BGD

AUS

PER

IRN

CHN

ITA

COL

TUR

NPL

MEX

PAK

GBR

CAN

ZAF

NZL

URY

GTM

SLV

FIN

ZWE

GHA

MAR

BOL

IDN

VEN

EGY

SWE

DEU

KEN

DNK

ECU

PHL

CHL

PRY

TTO

NOR

CIV

NGA

LKA

PAN

ISR

THA

KOR

ZMB

CHE

CRI

NLD

CYP

TUN

IRL

JAM

MYS

MUS

JOR

−4 −2 0 2 4

FD index

−1 −.5 0 .5 1

Log of openness

coef = .39212878, (robust) se = .38869956, t = 1.01

Openness and FD (conditional)

Figure 3: Partial scatter plots

These figures show partial scatter plots of the logarithm of LLY, and the aggregate

index FD, against the logarithm of openness. These are the partial associations

between trade and finance, conditional on the logarithm of GDP and legal origin

dummies.

Table 1. The Variables

Variable Description Sources

OPENC The sum of exports and imports over GDP (at current prices) Penn World Table 6.1

OPENK The sum of exports and imports over GDP (at international prices) Penn World Table 6.1

RGDPCH Average real GDP per capita over 1988-90 (cross-section study only) Penn World Table 6.1

CTRADE Natural propensity to trade, as derived by Frankel and Romer. Based on

aggregated fitted values of bilateral trade equation.

Frankel and Romer (1999)

LLY Ratio of liquid liabilities of financial system (currency plus demand and interest-

bearing liabilities of banks and nonbanks) to GDP

FSD

PRIVO Ratio of credit issued to private sector by banks and other financial intermediaries,

to GDP

FSD

BTOT

Ratio of commercial bank assets to sum of commercial and central bank assets FSD

OVC Ratio of overhead costs to total assets of the banks FSD

NIM Bank interest income minus interest expenses over total assets FSD

MCAP Ratio of the value of listed shares to GDP FSD

TVT Ratio of the value of shares traded on domestic exchanges to GDP FSD

TOR Ratio of the value of shares traded to market capitalization FSD

FD First principal component of LLY, PRIVO, BTOT, NIM, MCAP, TVT, TOR (all in

logs)

See text

FDSIZE First principal component of LLY and MCAP (all in logs) See text

FDEFF First principal component of OVC, NIM, TVT and TOR (all in logs) See text

FDBANK First principal component of LLY, PRIVO, BTOT, OVC and NIM (all in logs) See text

FDSTOCK First principal component of MCAP, TVT and TOR (all in logs) See text

FDEPTH First principal component of LLY, PRIVO and BTOT. This index is used in the

panel data analysis.

See text

LEGOR_UK British legal origin GDN

LEGOR_FR French legal origin GDN

LEGOR_GE German legal origin GDN

LEGOR_SC Scandinavian legal origin GDN

Key to Table 1: FSD – Financial Structure Database introduced by Beck, Demirguc-Kunt and Levine (2000). GDN – Global Development Network Database, 2002

Table 2. The indices of financial development

Measure Proportion LLY PRIVO BTOT OVC NIM MCAP TVT TOR

0.63 0.38 0.40 0.30 -0.32 -0.36 0.36 0.38 0.32

FDSIZE

0.80 0.71 0.71

FDEFF

0.68 -0.48 -0.51 0.52 0.49

FDBANK

0.71 0.48 0.48 0.38 -0.43 -0.46

FDSTOCK

0.86 0.55 0.62 0.56

FDEPTH

0.74 0.60 0.63 0.49

Notes: This table shows how our various indicators of financial development (FD, FDSIZE, FDEFF, FDBANK, FDSTOCK, FDEPTH) are constructed, from

the raw data on different measures of financial development. We construct indicators from the raw data using the first principal component of a number of

variables, namely the linear combination of the variables that has the highest sample variance, subject to the constraint that the sum-of-squares of the

coefficients equals unity. The table shows the weights that each index places on each of the (standardized) variables, and the proportion of the variance in

the original data that is explained by the first principal component.

The raw measures used are the natural logarithms of LLY = the ratio of liquid liabilities of the financial system (currency plus demand and interest-bearing

liabilities of banks and nonbanks) to GDP; PRIVO = the ratio of credit issued to the private sector by banks and other financial intermediaries to GDP;

BTOT= the ratio of commercial bank assets to the sum of commercial bank and central bank assets; OVC = the ratio of overhead costs to total assets of the

banks; NIM = the bank interest income minus interest expenses over total assets; MCAP = the ratio of the value of shares listed on domestic exchanges

(market capitalization) to GDP; TVT = the ratio of the value of shares traded on domestic exchanges to GDP; TOR = the ratio of the value of shares traded

on domestic exchanges to total market capitalization.

Table 3. Descriptive Statistics: 1990-2001

===============================================================

A. Summary Statistics for Openness and financial development measures

Variable Observation Mean Std. Dev. Min Max

OPENC

101 59.73 30.71 14.99 140.90

LLY

93 0.44 0.31 0.05 1.87

PRIVO

99 0.42 0.38 0.02 1.65

BTOT

98 0.77 0.21 0.17 1.00

OVC

92 0.05 0.02 0.01 0.11

NIM

91 0.06 0.03 0.01 0.17

MCAP

67 0.36 0.37 0.01 1.76

TVT

68 0.21 0.29 0.00 1.34

TOR

67 0.42 0.42 0.01 1.92

59 0.00 2.25 -5.22 3.45

FDSIZE

61 0.00 1.26 -3.06 2.58

FDEFF

66 0.00 1.64 -3.44 3.16

FDBANK

82 0.00 1.88 -4.80 3.51

FDSTOCK

67 0.00 1.60 -3.99 2.41

B. Correlations between openness and the existing financial development measures

OPENC LLY PRIVO BTOT OVC NIM MCAP TVT TOR

OPENC

1.00

LLY

0.22 1.00

PRIVO

0.18 0.86 1.00

BTOT

0.15 0.50 0.64 1.00

OVC

-0.21 -0.56 -0.54 -0.46 1.00

NIM

-0.15 -0.57 -0.59 -0.51 0.86 1.00

MCAP

0.24 0.58 0.71 0.41 -0.35 -0.40 1.00

TVT

0.04 0.54 0.77 0.44 -0.32 -0.37 0.79 1.00

TOR

-0.23 0.32 0.47 0.38 -0.29 -0.25 0.23 0.60 1.00

C. Correlations among openness and the new financial development measures

OPENC FD FDSIZE FDEFF FDBANK FDSTOCK

OPENC

1.00

0.27 1.00

FDSIZE

0.30 0.92 1.00

FDEFF

0.15 0.95 0.80 1.00

FDBANK

0.27 0.94 0.84 0.87 1.00

FDSTOCK

0.03 0.86 0.82 0.86 0.63 1.00

Table 4. External trade and financial development (whole sample), 1990-2001

FD FDSIZE FDEFF FDBANK FDSTOCK

OPENC (OLS) 0.64 0.244 0.516 0.56 -0.497

(0.08)* -0.25 (0.04)** (0.07)* (0.06)*

OPENC (GMM) 0.138 0.139 0.238 0.532 -0.883

(0.78) (0.48) (0.49) (0.15) (0.01)***

(a) R-squared 0.59 0.60 0.57 0.52 0.52

(b) First-stage F

69.07 75.75 68.55 68.60 87.56

0.32 0.13 0.14 0.27 0.10

(d) Wu-Hausman

0.16 0.49 0.29 0.92 0.06

(e) C statistic

0.17 0.47 0.28 0.91 0.08

Observations 57 56 62 79 64

Notes: This Table shows the point estimates and p-values for openness in OLS and GMM estimates, using

five alternative measures of financial development as the dependent variable. The coefficients and

heteroskedasticity-robust p values correspond to the natural logarithm of openness. Table 1 describes all

variables in detail. Other explanatory variables included in each of the regressions are the natural logarithm

of initial real GDP, and one or more dummy variables for legal origin. In the GMM estimates, the instrument

is the Frankel-Romer measure of the propensity to trade (CTRADE). All regressions exclude outliers, as

identified by a preliminary median regression (see text). Most of the diagnostics are based on 2SLS rather

than GMM.

significant at 10%; ** significant at 5%; *** significant at 1%

¹ This tests the significance of the excluded instrument (CTRADE) in the first-stage regression of 2SLS.

² This tests the homoskedasticity of the system of equations when 2SLS is used. Based on Pagan and Hall

(1983).

³ This is a Wu-Hausman test that the difference between the OLS and 2SLS coefficients is not systematic.

⁴ This is a GMM-based test of the null hypothesis that OPENC is orthogonal to the disturbances.

Table 5. External trade and financial development (subsamples), 1990-2001

A. Higher-income group

FD FDSIZE FDEFF FDBANK FDSTOCK

OPENC (OLS) 0.88 0.472 0.442 1.563 -0.371

(0.00)*** (0.01)*** (0.08)* (0.00)*** (0.09)*

OPENC (GMM) 0.831 0.514 0.264 1.254 -0.431

(0.00)*** (0.00)*** (0.27) (0.00)*** (0.03)**

(a) R-squared 0.77 0.72 0.55 0.84 0.58

(b) First stage F 46.99 44.95 46.46 30.98 55.16

(d) Wu-Hausman 0.78 0.78 0.33 0.05 0.63

(e) C statistic 0.74 0.73 0.27 0.05 0.59

Observations 27 28 34 26 34

B. Lower-income group

FD FDSIZE FDEFF FDBANK FDSTOCK

OPENC (OLS) -1.023 -0.435 -1.565 -0.221 -0.753

(0.13) (0.23) (0.01)*** (0.63) (0.19)

OPENC (GMM) -4.027 -1.357 -3.994 -0.723 -1.301

(0.05)** (0.02)** (0.01)*** (0.36) (0.25)

(a) R-squared 0.26 0.20 0.18 0.21 0.22

(b) First stage F 6.80 22.04 10.26 25.66 19.34

(d) Wu-Hausman 0.00 0.01 0.01 0.41 0.48

(e) C statistic 0.06 0.07 0.06 0.39 0.46

Observations 28 28 29 51 29

Notes: The upper panel is based on high-income and upper-middle countries while

lower panel is for lower-middle and low-income countries, as classified in the GDN

Growth Database. For other notes, please see Table 4.

Table 6. External trade and financial development (whole sample)

Dependent variable: FDEPTH OLS levels Within groups DIF-GMM DIF-GMM

SYS-GMM

Instrument set None None Full Reduced Reduced

Observations 520 520 432 432 520

Lag 1 FDEPTH 1.032 0.775 0.689 0.453 1.100

(61.620)*** (19.24) *** (9.23) *** (2.15) ** (24.71) ***

Lag 1 OPENC 0.071 0.621 0.973 1.114 0.421

(1.14) (3.52) *** (3.80) *** (1.92) * (1.77) *

Serial correlation (m1) p-value 0.00 0.45 0.00

Serial correlation (m2) p-value 0.02 0.12 0.01

Sargan p-value 0.44 0.38 0.03

Granger causality p-value 0.25 0.00 0.00 0.08 0.08

LR effect point estimate -2.229 2.765 3.126 2.038 -4.206

(Standard error) (2.51) (0.85) *** (1.26) *** (1.30) (3.46)

Notes: 88 countries, 1960-1999. Year dummies are included in all models (coefficients not reported). Figures in parentheses below point estimates are t-ratios. * significant at

10%; ** significant at 5%; *** significant at 1%

The GMM results reported here are two-step estimates with heteroskedasticity-consistent standard errors and test statistics; the standard errors are based on the finite sample

adjustment of Windmeijer (2005). m1 and m2 are tests for first-order and second-order serial correlation. First-order serial correlation is expected due to first-differencing, but

identification of the models relies on the absence of second-order serial correlation.

The Sargan test is used to assess the overidentifying restrictions and is asymptotically distributed as χ². The test uses the minimised value of the corresponding two-step GMM

estimator. The difference Sargan test is used to test the additional moment conditions used by the system GMM (SYS-GMM) estimator. The Granger causality test examines the

null hypothesis that financial development is not Granger-caused by openness; the test statistic is criterion based, using restricted and unrestricted models (see text).

The LR effect is the point estimate of the long-run effect of openness on financial development. Its standard error is approximated using the delta method.

Table 7. External trade and financial development (whole sample)

Dependent variable: FDEPTH OLS levels WG DIF-GMM DIF-GMM SYS-GMM SYS-GMM-1 SYS-GMM-2

Instrument set None None Full Reduced Reduced Reduced Reduced

Observations 432 432 346 344 432 432 432

Lag 1 FDEPTH 1.251 0.824 0.653 0.388 1.250 1.264 1.147

(17.10) *** (14.14) *** (3.01) *** (1.43) (14.99) *** (14.93) *** (6.52) ***

Lag 2 FDEPTH -0.250 -0.253 -0.205 -0.155 -0.294 -0.344 -0.258

(-3.11) *** (-4.01) *** (-2.47) ** (-1.76) * (-3.61) *** (-4.20) *** (-2.48) **

Lag 1 OPENC 0.731 0.803 1.240 1.235 0.968 1.174 0.914

(3.10) *** (3.65) *** (2.48) ** (2.78) *** (3.72) *** (2.87) *** (2.69) ***

Lag 2 OPENC -0.652 0.027 0.037 0.222 -0.518 -0.547 -0.434

(-2.96)*** (0.11) (0.18) (0.84) (-2.05) ** (-2.55) ** (-1.94) *

Serial correlation (m1) p-value 0.06 0.55 0.00 0.00 0.00

Serial correlation (m2) p-value 0.86 0.95 0.36 0.53 0.36

Sargan p-value 0.50 0.29 0.20 0.26 0.11

Diff-Sargan p-value 0.21 0.31 0.05

Heterogeneity test p-value 0.16 0.23 0.05

Granger causality p-value 0.01 0.00 0.01 0.02 0.01 0.00 0.02

Test of β

+β

=0 p-value 0.25 0.00 0.00 0.01 0.16 0.04 0.14

LR effect point estimate -47.369 1.935 2.314 1.900 10.328 7.856 4.330

(Standard error) (581.61) (0.59) *** (1.49) (0.86) ** (14.50) (7.84) (3.25)

Notes: 88 countries, 1960-1999. Year dummies are included in all models (coefficients not reported). Figures in parentheses below point estimates are t-ratios. * significant at 10%; ** significant at 5%; ***

significant at 1%. The GMM results reported here are two-step estimates with heteroskedasticity-consistent standard errors and test statistics; the standard errors are based on the finite sample

adjustment of Windmeijer (2005). m1 and m2 are tests for first-order and second-order serial correlation. First-order serial correlation is expected due to first-differencing, but identification of the models

relies on the absence of second-order serial correlation. The Sargan test is used to assess the overidentifying restrictions and is asymptotically distributed as χ². The test uses the minimised value of the

corresponding two-step GMM estimator. The difference Sargan test is used to test the additional moment conditions used by the system GMM estimators in which SYS GMM uses the standard moment

conditions, while SYS GMM (Modified 1) only uses the lagged first-differences of FDEPTH dated t-1 as instruments in levels and SYS GMM (Modified 2) only uses lagged first-differences of OPENC

dated t-1 as instruments in levels. The heterogeneity test is used to test the null that there are no individual effects (see text). The Granger causality test examines the null hypothesis that financial

development is not Granger-caused by openness; the test statistic is criterion based, using restricted and unrestricted models (see text). The LR effect is the point estimate of the long-run effect of

openness on financial development. Its standard error is approximated using the delta method.

Table 8. External trade and financial development (higher-income group)

Dependent variable: FDEPTH OLS levels WG DIF-GMM DIF-GMM SYS-GMM SYS-GMM-1 SYS-GMM-2

Instrument set None None Full Reduced Reduced Reduced Reduced

Observations 185 185 150 150 185 185 185

Lag 1 FDEPTH 1.166 0.753 0.368 0.123 1.173 1.218 0.957

(10.01) *** (8.65) *** (1.98) (0.64) (9.66) *** (7.01) *** (2.75) ***

Lag 2 FDEPTH -0.183 -0.298 -0.273 -0.214 -0.149 -0.393 -0.154

(-1.38) (-3.25) *** (-2.12) ** (-1.91) * (-1.19) (-3.02)*** (-1.10)

Lag 1 OPENC 0.686 1.237 0.842 0.734 0.814 0.734 0.586

(1.58) (3.10) *** (1.11) (1.31) (1.47) (1.60) (0.75)

Lag 2 OPENC -0.610 -0.013 -0.135 0.063 -0.374 -0.238 -0.573

(-1.50) (-0.03) (-0.42) (0.29) (-1.04) (-0.44) (-1.16)

Serial correlation (m1) p-value 0.22 0.87 0.02 0.01 0.07

Serial correlation (m2) p-value 0.33 0.47 0.41 0.97 0.68

Sargan p-value 0.92 0.71 0.79 0.53 0.48

Diff-Sargan p-value 0.68 0.17 0.12

Heterogeneity test P-value 0.78 0.28 0.41

Granger causality p-value 0.29 0.00 0.54 0.66 0.56 0.55 1.00

Test of β

+β

=0 p-value 0.47 0.00 0.21 0.08 0.39 0.27 1.00

LR effect point estimate 4.574 2.245 0.780 0.731 -18.448 8.582 0.066

(Standard error) (13.18) (0.747) *** (0.93) (0.58) (81.76) (6.72) (4.56)

Notes: 35 countries, 1960-1999. For other notes please see Table 7.

Table 9. External trade and financial development (lower-income group)

Dependent variable: FDEPTH OLS levels WG DIF-GMM DIF-GMM SYS-GMM SYS-GMM-1 SYS-GMM-2

Instrument set None None Full Reduced Reduced Reduced Reduced

Observations 247 247 194 194 247 247 247

Lag 1 FDEPTH 1.279 0.849 1.054 0.930 1.202 1.250 0.961

(15.46) *** (10.16) *** (4.40) *** (2.33) ** (13.43) *** (9.70) *** (5.51) ***

Lag 2 FDEPTH -0.356 -0.252 -0.306 -0.309 -0.366 -0.380 -0.258

(-4.32) *** (-2.77) *** (-3.23) *** (-2.61) *** (-5.12) *** (-4.64) *** (-2.45) **

Lag 1 OPENC 0.756 0.640 2.258 2.281 0.933 1.262 1.307

(2.99) *** (2.52) *** (2.65) *** (1.72) * (2.76) *** (1.87) * (3.39) ***

Lag 2 OPENC -0.666 -0.207 -0.201 -0.204 -0.665 -0.703 -0.453

(-2.73) *** (-0.67) (-0.55) (-0.45) (-2.99) *** (-2.36) ** (-1.42)

Serial correlation (m1) p-value 0.02 0.15 0.00 0.00 0.01

Serial correlation (m2) p-value 0.20 0.34 0.30 0.31 0.35

Sargan p-value 0.74 0.40 0.53 0.47 0.42

Diff-Sargan p-value 0.65 0.57 0.45

Heterogeneity test P-value 0.51 0.32 0.13

Granger causality p-value 0.01 0.04 0.00 0.06 0.06 0.05 0.02

Test of β

+β

=0 p-value 0.40 0.19 0.01 0.01 0.31 0.04 0.17

LR effect point estimate 1.158 1.075 8.140 5.484 1.640 4.332 2.877

(Standard error) (1.13) (0.80) (7.37) (7.26) (1.63) (4.74) (0.94)***

Notes: 53 countries, 1960-1999.

For other notes please see Table 7.

Appendix Tables

A1. External trade and financial development (whole sample), 1990-2001

LLY PRIVO BTOT OVC NIM MCAP TVT TOR

OPENC (OLS)

0.267 0.221 0.024 -0.101 -0.228 -0.033 -0.702 -0.709

(0.01)*** (0.04)** (0.68) (0.44) (0.02)** (0.86) (0.05)** (0.00)***

OPENC (GMM)

0.272 0.077 -0.046 -0.263 -0.193 -0.148 -0.984 -0.879

(0.04)** (0.64) (0.52) (0.14) (0.24) (0.59) (0.06)* (0.00)***

(a) R-squared

0.62 0.70 0.42 0.21 0.49 0.50 0.53 0.47

(b) First stage F ¹

70.30 70.30 68.73 65.59 65.30 82.88 82.82 82.88

0.45 0.90 0.43 0.58 0.42 0.29 0.47 0.67

(d) Wu-Hausman³

0.96 0.10 0.22 0.57 0.49 0.12 0.15 0.26

(e) C statistic⁴

0.96 0.11 0.22 0.55 0.46 0.13 0.15 0.24

Observations

89 89 88 81 80 59 60 59

Notes: This table reports the coefficients and heteroskedasticity-robust p values for the natural logarithm of openness, for eight alternative dependent

variables (in logarithms) using OLS and GMM. Table 1 describes all variables in detail. Other explanatory variables included in each of the regressions are

the natural logarithm of initial real GDP, and dummy variables for British, French and German legal origin. In GMM estimates, the instrument is the

Frankel-Romer measure of the propensity to trade (CTRADE). All regressions exclude outliers, as identified by a preliminary median regression (see text).

Most of the diagnostics are based on 2SLS rather than GMM.

* significant at 10%; ** significant at 5%; *** significant at 1%

¹ This tests the significance of the excluded instrument (CTRADE) in the first-stage regression of 2SLS.

² This tests the homoskedasticity of the system of equations when 2SLS is used. Based on Pagan and Hall (1983).

³ This is a Wu-Hausman test that the difference between the OLS and 2SLS coefficients is not systematic.

⁴ This is a GMM-based test of the null hypothesis that OPENC is orthogonal to the disturbances.

A2. External trade and financial development (subsamples), 1990-2001

A. Higher-income group

LLY PRIVO BTOT OVC NIM MCAP TVT TOR

OPENC (OLS) 0.468 0.177 0.044 -0.602 -0.505 0.057 -0.334 -0.448

(0.00)*** (0.14) (0.04)** (0.00)*** (0.00)*** (0.67) (0.35) (0.08)*

OPENC (GMM) 0.525 0.313 0.044 -0.63 -0.475 -0.093 -0.841 -0.735

(0.00)*** (0.04)** (0.09)* (0.00)*** (0.01)*** (0.72) (0.09)* (0.02)**

(a) R-squared 0.72 0.71 0.58 0.53 0.64 0.48 0.40 0.32

(b) First stage F 31.68 31.68 31.91 31.68 31.68 31.68 31.68 31.68

(d) Wu-Hausman 0.48 0.56 0.89 0.20 0.72 0.07 0.18 0.50

(e) C statistic 0.41 0.51 0.87 0.19 0.69 0.07 0.14 0.42

Observations 29 29 28 29 29 29 29 29

B. Lower-income group

LLY PRIVO BTOT OVC NIM MCAP TVT TOR

OPENC (OLS) 0.146 0.182 -0.027 0.365 0.183 -0.49 -0.947 -1.105

(0.28) (0.23) (0.79) (0.03)** (0.29) (0.21) (0.21) (0.01)***

OPENC (GMM) 0.004 -0.266 -0.192 0.275 0.276 -0.327 -1.663 -1.584

(0.99) (0.41) (0.23) (0.35) (0.33) (0.68) (0.26) (0.07)*

(a) R-squared 0.34 0.45 0.25 0.16 0.10 0.25 0.20 0.18

(b) First stage F 31.28 31.28 31.28 22.22 22.17 25.20 26.25 25.20

(d) Wu-Hausman 0.41 0.05 0.16 0.88 0.74 0.58 0.46 0.34

(e) C statistic 0.40 0.08 0.19 0.87 0.73 0.55 0.44 0.32

Observations 60 60 60 52 51 29 30 29

Notes: The upper panel is based on high-income and upper-middle countries while lower panel is for lower-middle and low-income countries, as classified in the GDN Growth Database in World

Bank. For other notes, please see Table 4.

A3. External trade and

financial development (whole sample), 1990-2001

LLY PRIVO BTOT OVC NIM MCAP TVT TOR

OPENC (OLS) 0.187 0.192 0.035 -0.129 -0.077 0.104 -0.613 -0.713

(0.07)* (0.11) (0.58) (0.34) (0.53) (0.61) (0.10)* (0.00)***

OPENC (GMM) -0.016 -0.228 -0.093 -0.115 -0.134 -0.235 -1.27 -1.094

(0.93) (0.26) (0.20) (0.52) (0.39) (0.34) (0.01)*** (0.00)***

(a) R-squared 0.46 0.60 0.39 0.19 0.32 0.42 0.48 0.40

(b) F test¹ 72.54 79.72 78.34 68.56 68.46 84.26 85.63 84.26

(d) Wu-Hausman³ 0.13 0.01 0.10 0.90 0.61 0.08 0.05 0.08

(e) C statistic

⁴

0.14 0.02 0.11 0.90 0.58 0.08 0.07 0.09

Observations 93 99 98 92 91 67 68 67

Notes: As previously, but these regressions do not exclude outliers.

A4. External trade and financial development (subsamples), 1990-2001

A. Higher-income group

LLY PRIVO BTOT OVC NIM MCAP TVT TOR

OPENC (OLS) 0.290 0.210 0.049 -0.547 -0.457 0.224 -0.254 -0.443

(0.04)** (0.13) (0.03)** (0.00)*** (0.00)*** (0.24) (0.49) (0.08)*

OPENC (GMM) 0.318 0.112 0.041 -0.458 -0.445 -0.235 -0.645 -0.547

(0.03)** (0.44) (0.04)** (0.01)*** (0.00)*** (0.34) (0.04)** (0.01)***

(a) R-squared 0.62 0.55 0.59 0.40 0.55 0.36 0.36 0.34

(b) First stage F ¹ 41.33 53.26 52.54 51.11 51.11 51.11 51.11 51.11

(d) Wu-Hausman³ 0.74 0.36 0.66 0.23 0.90 0.08 0.21 0.57

(e) C statistics⁴ 0.71 0.31 0.62 0.89 0.08 0.16 0.44

Observations 31 37 36 36 36 67 36 36

B. Lower-income group

LLY PRIVO BTOT OVC NIM MCAP TVT TOR

OPENC (OLS) 0.115 0.174 0.001 0.331 0.256 -0.07 -1.085 -1.129

(0.47) (0.37) (1.00) (0.10)* (0.15) (0.88) (0.18) (0.01)***

OPENC (GMM) -0.376 -0.687 -0.243 0.529 0.437 -0.626 -2.827 -2.53

(0.29) (0.13) (0.11) (0.14) (0.14) (0.45) (0.10)* (0.04)**

(a) R-squared 0.20 0.29 0.24 0.12 0.09 0.14 0.13 0.17

(b) First stage F¹ 33.82 33.82 33.82 24.41 24.40 20.03 22.74 20.03

(d) Wu-Hausman³ 0.06 0.00 0.08 0.43 0.41 0.35 0.09 0.03

(e) C statistics⁴ 0.10 0.03 0.12 0.45 0.42 0.33 0.14 0.08

Observations 62 62 62 56 55 31 32 31

Notes: As previously, but these regressions do not exclude outliers.

A5. External trade and financial development (higher-income group, OPENK), 1990-

2001

FD FDSIZE FDEFF FDBANK FDSTOCK

OPENK (OLS) 0.763 0.460 0.307 1.223 -0.456

(0.01)*** (0.01)*** (0.18) (0.00)*** (0.03)**

OPENK (GMM) 0.708 0.414 0.245 1.046 -0.400

(0.00)*** (0.02)** (0.28) (0.00)*** (0.02)**

(a) R-squared 0.78 0.71 0.53 0.71 0.61

(b) First stage F¹ 37.86 34.58 39.04 32.24 44.69

(d) Wu-Hausman³ 0.61 0.78 0.68 0.20 0.51

(e) C statistics⁴ 0.54 0.74 0.63 0.14 0.47

Observations 26 29 34 28 34

Notes: These regressions are based on higher-income group for OPENK instead of

OPENC. For other notes, please see earlier tables.

External trade and financial development (higher-income group), 1990-2001

FD FDSIZE FDEFF FDBANK FDSTOCK

OPENC (OLS) 0.019 0.008 -0.011 0.028 -0.009

(0.00)*** (0.05)** (0.11) (0.00)*** (0.17)

OPENC (GMM) 0.017 0.011 -0.003 0.021 -0.013

(0.01)*** (0.05)** (0.65) (0.00)*** (0.04)**

(a) R-squared 0.82 0.66 0.55 0.88 0.36

(b) First stage F¹ 22.44 18.29 22.22 11.61 30.04

(d) Wu-Hausman³ 0.63 0.43 0.08 0.02 0.16

(e) C statistic⁴ 0.58 0.34 0.09 0.03 0.09

Observations 27 29 34 26 34

Notes: These regressions are based on higher-income group in which the first principal component for

any financial indicator is the linear combination of measures that are not in logarithms. OPENC is not in

logarithms, but GDP is. For other notes, please see earlier tables.

A6. External trade and financial development (whole sample), 1990-2001

FD FDSIZE FDEFF FDBANK FDSTOCK

OPENC (OLS) 0.392 0.346 -0.036 0.487 -0.438

(0.32) (0.15) (0.91) (0.14) (0.11)

OPENC (GMM) -0.349 -0.081 -0.336 -0.008 -0.964

(0.56) (0.82) (0.420 (0.99) (0.01)***

(a) R-squared 0.54 0.49 0.41 0.48 0.47

(b) First stage F¹ 72.39 75.47 83.21 61.91 84.26

(d) Wu-Hausman³ 0.05 0.06 0.29 0.20 0.04

(e) C statistic⁴ 0.07 0.08 0.28 0.22 0.05

Observations 59 61 66 82 67

Notes: as previously, but these regressions do not exclude outliers.

A7. External trade and financial development (subsamples), 1990-2001

a. Higher-income group

FD FDSIZE FDEFF FDBANK FDSTOCK

OPENC (OLS) 1.048 0.602 0.592 1.303 -0.156

(0.00)*** (0.01)*** (0.03)** (0.00)*** (0.55)

OPENC (GMM) 0.707 0.421 0.377 1.204 -0.47

(0.06)* (0.03)** (0.18) (0.00)*** (0.03)**

(a) R-squared 0.71 0.63 0.52 0.69 0.38

(b) First stage F¹ 38.17 38.03 51.11 38.17 51.11

(d)W-Hausman³ 0.17 0.35 0.28 0.61 0.16

(e) C statistic⁴ 0.16 0.31 0.23 0.57 0.12

Observations 29 30 36 29 36

b. Lower-income group

FD FDSIZE FDEFF FDBANK FDSTOCK

OPENC (OLS) -0.775 0.111 -1.004 -0.309 -0.854

(0.40) (0.82) (0.17) (0.55) (0.16)

OPENC (GMM) -2.78 -1.029 -2.147 -1.879 -2.29

(0.20) (0.39) (0.16) (0.10)* (0.11)

(a) R-squared 0.12 0.10 0.10 0.19 0.14

(b) F test¹ 20.06 20.03 20.06 25.02 20.03

(d) Wu-Hausman³ 0.10 0.11 0.22 0.04 0.08

(e) C statistic⁴ 0.14 0.15 0.24 0.09 0.12

Observations 30 31 30 53 31

Notes: As previously, but these regressions do not exclude outliers.

A8. Panel data results using only information on LLY to measure financial development

Dependent variable: LLY Whole

Sample

Higher-income

Group

Lower-income

Group

Instrument set Reduced Reduced Reduced

Lag 1 LLY 1.293 1.139 1.163

(12.21)*** (6.58)*** (9.48)***

Lag 2 LLY -0.403 -0.086 -0.475

(-2.59)*** (-0.29) (-4.34)***

Lag 1 OPENC 0.120 0.066 0.224

(1.84)* (-0.83) (4.29)***

Lag 2 OPENC -0.056 -0.083 -0.054

(-0.79) (-0.71) (-0.92)

Serial correlation (m1) p-value 0.03 0.18 0.01

Serial correlation (m2) p-value 0.93 0.94 0.43

Sargan p-value 0.17 0.92 0.67

Diff-Sargan p-value 0.15 0.99 0.97

Heterogeneity test P-value 0.27 1.00 0.69

Granger causality p-value 0.05 0.13 0.09

LR effect point estimate 0.528 2.794 0.546

(Standard error) (0.58) (6.08) (0.21) ***

Notes: 88 countries in total, 35 higher-income countries and 53 lower-income countries; 1960-1999. For other notes

please see Table 7.

Financial Development and Central Bank Bilateral Currency Swaps: Is there Trade Effect?

Preprint

Full-text available

Feb 2024

Abdullahi Ahmed Mohammed

The proper assessment and understanding of the financial system are at the core of a robust analysis of macroeconomic fundamentals (Svirydzenka, 2016). Bilateral currency swap enables countries to boost their liquidity access in the financial system for trade and financial transaction. Significantly, we examine the financial development of both China and its currency swap partners. We test our empirical model using data on financial development for a sample of 27 countries. We provide empirical evidence that currency swap is important for trade especially for countries with relatively low level of financial development. It is well documented that the differences in development amongst countries are substantial, and such differences are important in the determination of trade pattern. The level of financial development was proxied by the interaction term of disaggregated measure of financial development such access, depth, and efficiency each interacted with swaps. We provide empirical evidence that differential level of financial development can be a key determinant of whether a country can use currency swap lines for international trade. In rich countries, strong financial system promote trade, the opposite is the case in poorer ones. Perhaps, empirical tests on the influence of financial system and on trade remain on the research agenda especially looking at industry-level import and export data.

Financial Development and Central Bank Bilateral Currency Swaps: Is there Trade Effect?

Article

Full-text available

Dec 2023

Abdullahi Ahmed Mohammed

The Nexus Between Banking Development and Economic Growth in the Presence of Trade Openness and Inflation: Case of Vietnam

Article

Full-text available

Jul 2023

The purpose of this research is to examine studies on the relationship between banking development and economic growth in conditions of trade openness and inflation in Vietnam by using ARDL regression estimation method on data. time series from the first quarter of 2000 to the fourth quarter of 2019. Research results have found: There is a relationship between banking development and economic growth in terms of trade openness and inflation in Vietnam.; There was an external economic shock affecting the relationship between banking development and economic growth in Vietnam in 2008. A threshold value of 32.86% was found and a threshold value of 32.86% was found. of inflation is 9.19%. With a trade openness of less than 32.86%, it shows that banking development does not contribute to economic growth in Vietnam. However, with a trade openness greater than 32.86%, it shows that banking development has a positive impact on economic growth in Vietnam. Meanwhile, with an inflation rate below 9.19%, banking development has a positive impact on economic growth through domestic credit to the private sector. And vice versa, with an inflation rate above the threshold of 9.19%, we find a positive impact of banking development on economic growth through the interest rate spread. JEL classification: G10, G21, G28, G38.

The heterogeneous effects of macroeconomic and financial factors on financial deepening in Africa: evidence from a method of moments quantile regression analysis

Article

Jun 2024

Purpose This paper aims to investigate the heterogeneous effects of macroeconomic and financial factors across various distributions of financial deepening in 22 African countries over the past two decades (2000–2019). Design/methodology/approach The paper uses a recent method of moments quantile regression, which accounts for the often overlooked heterogeneity effects. The analysis focuses on the banking sector, which is predominant in Africa, using a broad range of macroeconomic and financial indicators. Findings The findings show that gross domestic product per capita positively and significantly impacts financing deepening with an increasing marginal benefit as depth increases. Trade openness positively and substantially affects only high financial deepening. Real interest rate, real exchange rate and inflations negatively and significantly affect financial deepening, especially at higher than lower levels. Financial stability positively and substantially influences financial deepening with an increasing marginal benefit as the depth increases. Bank lending interest rate, bank lending–deposit rate spread, bank concentration and return on equity negatively and substantially impact higher levels of financial deepening than lower levels. Practical implications These findings are crucial to policymakers and development partners, as promoting a favourable financial environment and stable macroeconomic policies based on the heterogeneity of financial depths can increase debt financing in Africa. Originality/value To the best of the authors’ knowledge, this paper is one of the first attempts to analyse the heterogeneous effects of macroeconomic and financial determinants on varying levels of financial depth in Africa.

Are Trade Openness and Institutional Quality Important For Financial Development?

Article

Jun 2019

This study examines the importance of trade openness and institutional quality for financial development in sub-Saharan Africa. The study covers37 Sub-Saharan African countries over the period 1986-2016. Using the system Generalised Method of Moments technique, it finds that trade openness has a positive impact on financial development while only democracy accountability and government stability contribute significantly to financial development. The study finds that institutional quality does not complement trade openness to exact positive impacts on financial development. There is a need for policies that will foster the development of institutional quality in Sub-Saharan Africa.

Asian Economic Sustainability: A Focus on Financial Development, Education, and Governance

Article

Full-text available

Sep 2023

In order to achieve economic sustainability, Asian nations must coordinate their efforts with the Sustainable Development Goals (SDGs) of the United Nations. This paper explores this vital issue. Asia plays a crucial role in the global economy, which emphasizes how urgent it is to improve economic sustainability in order to promote justice and resilience on a global scale. In order to examine the factors that influence economic sustainability in the area between 2000 and 2021, this study looks into the effects of financial development, education, governance, and labor force dynamics. The study applies a rigorous econometric technique and makes use of panel regression and panel two-stage least squares (2SLS) models to illuminate the various aspects that impact the sustainability of economies in Asia. The empirical results highlight the critical roles that government expenditures on financial development, workforce expansion, and education play in promoting economic sustainability. Moreover, governance metrics positively influence GDP, suggesting that governance plays a critical role in determining long-term economic results. For steady and fair economic growth, this paper recommends giving priority to policies that assist financial development, wise educational investments, and the promotion of good governance. Policymakers can build focused initiatives that support economic sustainability and are in line with the larger global goal for resilient and inclusive development by using the identified determinants and their interaction.

Do Financial Markets in Developed Countries Continue Benefiting from Economic Openness - An Evidence from G10 Group.docx

Article

Full-text available

Jan 2024

The interactions between economic openness with other variables including inflation and GDP growth help to capture the different channels through which economic liberalization can benefit financial market development. This paper examines the influences of economic openness measured by trade openness and financial openness, on the development of financial markets in developed economies proxied by the Financial Market Index, a composite index representing the sum of financial market depth, access, and efficiency measurements. Yearly panel data from World Bank and International Monetary Fund for G10 countries spanning the period 2001-2019 and Autoregressive Distributed Lag model (panel-ARDL) are employed for analysis, and main results follow. Trade openness, financial openness, and economic growth are crucial determinants of financial market development in the long-run. However, in the short run, only financial openness has a positive and a significant impact on financial market development. The empirical results also reveal that Inflation has a detrimental effect on the evolution of the financial markets in the short and long term. The error correction term is negative in sign and statistically significant suggesting a long run equilibrium relationship amongst our variables, where, 28% of disequilibria in financial market development is corrected annually; a fairly high speed of adjustment that takes around 3.5 years to restore equilibrium. Finally, decision-makers should be attentive regarding the economic openness, and able to assess and manage trade and financial openness, keeping a growing rate of economic growth and low inflation rate, in order to promote the stability of the financial markets, and allow developed economies to exploit the benefits of economic openness for financial markets development.

Openness Policies and Financial Development, an Ardl Approach: The Case of Ghana

Preprint

Full-text available

Jan 2023

China's Bilateral Currency Swap Agreement: Strategic move to Foster Political and Financial Hegemony

Article

Full-text available

Dec 2023

Abdullahi Ahmed Mohammed

The global financial crisis of 2008 nearly put a halt to China's export-led and current account surpluses trajectory, in 2007 China's current account surplus fell from 10% of GDP to about 2% in 2013. This necessitates the internationalization of the Chinese Renminbi to boost trade, investment and hedge against foreign currency risk through bilateral currency swap. In bilateral currency swap, on the trade date, counter parties exchange notional amounts in two different currencies. For instance, one party receives 30 million British pounds while the other receives 3.3 million Chinese Renminbi. This implies a GBP/RMB exchange rate of 1.1, and at the end of the deal they swap again using the same exchange rate. Evidently, the currency bilateral swap agreements signed by the People's Bank of China and some Central Banks in advanced, emerging markets and developing economies is reinforcing the trend of Renminbi internationalization in global trade. Our empirical results show that Furthermore, the relative trade shares and exports intensity depicts a large positive swing especially for the swap provider, further suggesting that swap line's primal motive perhaps resolves around the provider country's self-interest, even though the benefits are substantially symbiotic for the recipient and provider country.

The role of international tourism, trade openness and renewable energy in the financial development of selected African countries

Article

Oct 2023

Estimation and Inference in Econometrics

Book

Full-text available

Jan 1993

Offering a unifying theoretical perspective not readily available in any other text, this innovative guide to econometrics uses simple geometrical arguments to develop students' intuitive understanding of basic and advanced topics, emphasizing throughout the practical applications of modern theory and nonlinear techniques of estimation. One theme of the text is the use of artificial regressions for estimation, reference, and specification testing of nonlinear models, including diagnostic tests for parameter constancy, serial correlation, heteroscedasticity, and other types of mis-specification. Explaining how estimates can be obtained and tests can be carried out, the authors go beyond a mere algebraic description to one that can be easily translated into the commands of a standard econometric software package. Covering an unprecedented range of problems with a consistent emphasis on those that arise in applied work, this accessible and coherent guide to the most vital topics in econometrics today is indispensable for advanced students of econometrics and students of statistics interested in regression and related topics. It will also suit practising econometricians who want to update their skills. Flexibly designed to accommodate a variety of course levels, it offers both complete coverage of the basic material and separate chapters on areas of specialized interest.

Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations

Article

Jan 1991

Panel data econometrics: advanced texts in econometrics

Article

Jan 2003

M. Arellano

Bank-based and market-based financial systems: Cross-country comparisons

Article

Jan 2001

Financial Dependence and International Trade

Article

May 2003

Thorsten Beck

Does financial development translate into a comparative advantage in industries that use more external finance? The author uses industry-level data on firms’ dependence on external finance for 36 industries and 56 countries to examine this question. It is shown that countries with better-developed financial systems have higher export shares and trade balances in industries that use more external finance. These results are robust to the use of alternative measures of external dependence and financial development and are not due to reverse causality or simultaneity bias.

The Political Economy of War Finance

Article

A Finite Sample Correction for the Variance of Linear Two-Step GMM Estimators

Article

Jan 2000

Frank Windmeijer

Monte Carlo studies have shown that estimated asymptotic standard errors of the efficient two-step generalised method of moments (GMM) estimator can be severely downward biased in small samples. The weight matrix used in the calculation of the efficient two-step GMM estimator is based on initial consistent parameter estimates. In this paper it is shown that the extra variation due to the presence of these estimated parameters in the weight matrix accounts for much of the difference between the finite sample and the asymptotic variance of the two-step GMM estimator that utilises moment conditions that are linear in the parameters. This difference can be estimated, resuling in a finite sample corrected estimate of the variance. In a Monte Carlo study of a panel data model it is shown that the corrected variance estimate approximates the final sample variance well, leading to more accurate inference.

Dynamic Panel Data Estimation Using DPD-a Guide for Users

Article

Jan 1988

Reliable Inference for GMM Estimators? Finite Sample Properties of Alternative Test Procedures in Linear Panel Data Models

Article

Mar 2005
ECONOMET REV

We compare the finite sample performance of a range of tests of linear restrictions for linear panel data models estimated using the generalized method of moments (GMM). These include standard asymptotic Wald tests based on one-step and two-step GMM estimators; two bootstrapped versions of these Wald tests; a version of the two-step Wald test that uses a finite sample corrected estimate of the variance of the two-step GMM estimator; the LM test; and three criterion-based tests that have recently been proposed. We consider both the AR(1) panel model and a design with predetermined regressors. The corrected two-step Wald test performs similarly to the standard one-step Wald test, whilst the bootstrapped one-step Wald test, the LM test, and a simple criterion-difference test can provide more reliable finite sample inference in some cases.

Financial intermediation and growth: Causality and causes without outliers

Article

Jan 2009

Corrado Andini

In a seminal paper, Levine et al. (J Monet Econ 46:31–77, 2000) provide cross-sectional evidence showing that financial development has positive average impact on long-run growth, using a sample of 71 countries. We argue that the evidence is sensitive to the presence of outliers.

Do External Trade Promote Financial Development?

Abstract and Figures

Recommended publications

Understanding the Dynamic Effects of Government Spending on Foreign Trade

Liquidity Premia and Interest Rate Parity

Is inflation an international phenomenon?

The Effects of Foreign Shocks When Interest Rates Are at Zero