Financial intermediation, trade agreements and international trade*

• Published date: 01 March 2021
• Author: Anne‐Gaël Vaubourg,Duc Bao Nguyen
• Date: 01 March 2021
• DOI: http://doi.org/10.1111/twec.13010

788

|

wileyonlinelibrary.com/journal/twec World Econ. 2021;44:788–817.

© 2020 John Wiley & Sons Ltd

Received: 6 May 2019

|

Revised: 5 July 2020

|

Accepted: 13 July 2020

DOI: 10.1111/twec.13010

ORIGINAL ARTICLE

Financial intermediation, trade agreements and

international trade*

Duc BaoNguyen1,2

|

Anne-GaëlVaubourg3

1VNU University of Economics and Business, Vietnam National University, Hanoi, Vietnam

2Univ. Bordeaux, CNRS, GRETHA, UMR 5113, Pessac, France

3CRIEF, EA 2249, University of Poitiers, Poitiers, France

KEYWORDS

exports, financial intermediation, international trade, regional trade agreements

1

|

INTRODUCTION

After several decades of trade openness and financial liberalisation, the great trade collapse observed

subsequent to the 2008 financial crisis appeared as a major opportunity to consider the links between

finance and trade (Auboin,2009, 2011; Auboin & Engemann,2014).

The key theoretical rationale for the impact of finance on trade is the existence of upfront export

costs that firms face when they sell abroad. These costs, which are related to advertising, gathering

information on foreign customers, translation and organising foreign distribution networks as well as

administrative procedures and compliance with the regulatory environment, must be externally fi-

nanced. As emphasised by Amiti and Weinstein (2011), Berman, De Sousa, Martin, and Mayer (2012)

and Schmidt-Eisenlohr (2013), export costs also result from the time gap between production and sale.

Hence, working capital needs to be financed while goods are in transit.1 Of course, time to ship is

particularly long in international trade and costs for exporters can be exacerbated by customs and ad-

ministrative entry procedures (Djankov, Freund, & Pham, 2010; Hummels & Schaur,2013).2 For

1Time costs also encompass depreciation costs such as literal spoilage and technological obsolescence (Hummels & Schaur,

2013).

2It is noteworthy that, in contrast to Anderson and van Wincoop (2004, p. 691), who provide a broad definition of trade costs

(‘costs incurred in getting a good to a final user other than the marginal cost of producing that good itself: transportation costs

(freight cost and time cost), policy barriers (tariffs and non-tariffs measures), information costs, contract enforcement costs,

costs associated with the use of different currencies, legal and regulatory costs, and local distribution costs’), the literature on

trade and finance focuses on the costs incurred by the exporting firm.

*We thank the editor, two anonymous referees, Jérôme Héricourt and Vincent Vicard for many fruitful discussions on the

paper, and Thomas Zylkin for useful remarks on the econometric model and the database. We are also grateful to seminar

participants at the 20th Warsaw European Trade Study Group (ETSG) Conference, the Global Political Economy Network

(GPEN) Conference (London), the seminar at GREThA (University of Bordeaux), the seminar at CRIEF (University of

Poitiers), the International Network for Economic Research (INFER) Workshop on New Challenges of Economic and

Financial Integration (Bordeaux) and the 36th International Symposium on Money, Banking and Finance (Besançon) for their

excellent feedback. All errors are our own.

|

789

NGUYEN aNd VaUBOURG

these reasons, exports crucially depend on the strength of firms' financial constraint (Chaney,2016;

Manova,2013) and the level of financial development (Beck,2002, 2003). The favourable effect of

financial development on exports is also shown to be particularly strong in more financially vulnera-

ble sectors (Manova,2008, 2013), during financial crisis (Berman etal.,2012; Chor & Manova,2012;

Iacovone, Ferro, Pereira-López, & Zavacka,2019) and when export fixed costs are high, that is when

the exporting country is weakly opened to trade (Manova,2008). Finally, the literature also empha-

sises the key role of banks and other financial institutions (e.g. insurance companies) in the provision

of trade finance tools (Amiti & Weinstein,2011; Auboin & Engemann,2014; Egger & Url, 2006;

Felbermayr & Yalcin,2013; Moser, Nestmann, & Wedow,2008; Niepmann & Schmidt-Eisenlohr,2017;

Schmidt-Eisenlohr,2013; van der Veer,2015), thus suggesting that financial intermediation has a

particularly favourable effect on international trade.

However, the literature on trade and finance does not consider an element of trade policies that cru-

cially determines the level of export costs, that is the signature of regional trade agreements (RTAs).

Since the early 1990s, the number of RTAs concluded among countries has steadily increased. These

agreements have increasingly gone beyond regional boundaries over time and turned into more and

more cross-regional ones.

Regional trade agreements help to substantially reduce traditional tariff and non-tariff measures

(e.g. technical standards, sanitary and phytosanitary conditions, quotas) among member countries

as well as other ‘behind-the-border’ barriers relating to environment regulation or employment law

(Chauffour & Maur,2010; Pomfret & Sourdin,2009). But RTAs also contain trade facilitation provi-

sions that alleviate ‘cross-border’ barriers by simplifying customs paperwork and procedures (i.e. clear-

ance, inspections or technical controls), thus reducing time to export (Djankov etal.,2010; Hillberry

& Zhang, 2018; World Trade Organization,2015). Because reduced enforcement frictions and a

shorter time gap between production and sale imply lower export costs (Schmidt-Eisenlohr,2013),

RTAs contribute to mitigating exporting firms' need for external funds. This argument suggests that

the positive impact of external finance, especially financial intermediation, on exports should be lower

when the exporting and the importing countries are involved in an RTA. The first contribution of this

paper is to check for the existence of such an interaction between RTAs and financial intermediation.

In line with Rajan and Zingales (1998), Manova (2008) and Manova, Wei, and Zhang (2015), the

second contribution of this paper is to show that the intensity of interactions between RTAs and fi-

nancial intermediation also has a sectoral dimension. Our results indicate that the existence of an RTA

between two trading partners mitigates the export-promoting effect of financial intermediation in the

exporting country but to a weaker extent for the most financially constrained sectors.

Finally, our paper also indicates that RTAs interact with financial intermediation not only in the ex-

porting country but also in the importing one. Indeed, it is noteworthy that our measure of trade open-

ness, that is the existence of an RTA between two countries, is a country-pair-specific variable that

captures the bilateral dimension of trade liberalisation. Hence, following the idea that the cost of ex-

ternal finance in the importing country also matters for trade (Niepmann & Schmidt-Eisenlohr,2017;

Schmidt-Eisenlohr,2013), we show that RTAs mitigate the favourable effect of financial intermedia-

tion not only in the source but also in the destination country.

To conduct our empirical investigation, we estimate a gravity model on a data set of 69 devel-

oped and developing countries over the period 1986–2006. To our knowledge, this study is one of

the first to identify the impacts of the interaction term between a financial intermediation indica-

tor (a time-varying country-specific variable) with regional trade agreements (a bilateral determi-

nant of trade) on international trade within a panel data gravity model with fixed effects. Because

country-time fixed effects control for all time-varying country-specific factors, they do not allow

one to identify the impact of financial indicators, which are perfectly collinear with fixed effects. To