Services trade and credit frictions: Evidence with matched bank–firm data

• Date: 01 May 2020
• Author: Francesco Bripi,David Loschiavo,Davide Revelli
• Published date: 01 May 2020
• DOI: http://doi.org/10.1111/twec.12930

1216

|

wileyonlinelibrary.com/journal/twec World Econ. 2020;43:1216–1252.

© 2020 John Wiley & Sons Ltd

Received: 19 December 2019

|

Accepted: 2 January 2020

DOI: 10.1111/twec.12930

SPECIAL ISSUE ARTICLE

Services trade and credit frictions: Evidence with

matched bank–firm data

FrancescoBripi1

|

DavidLoschiavo2

|

DavideRevelli3

1Bank of Italy, Milan, Italy

2Bank of Italy, Roma, Italy

3Bank of Italy, Genoa, Italy

KEYWORDS

bank–firm data, credit frictions, trade in services

1

|

INTRODUCTION

Due to the technology improvements, services trade has grown considerably in the recent decades.1 In

this context, competitiveness is a key factor for firms exporting services, but it requires dedicated in-

vestments in intangible capital and product customisation, that may be subject to credit frictions, es-

pecially during a financial shock.

After the Great Recession, a large body of empirical literature has established that financial fric-

tions have important effects on several firm outcomes, such as investments, employment and exports

of goods (e.g. Amiti & Weinstein, 2018; Chodorow-Reich, 2014; Foley & Manova, 2015). However,

we still have a limited understanding of the influence of financial factors on firms' abilities to export

services.

A clear prediction of the role of credit frictions for services exports is not obvious a priori. On one

hand, services typically require lower investment in physical initial capital than manufacturing; hence,

it is reasonable to expect lower credit needs for starting up a business or the export activity.

On the other hand, exporting services requires high upfront sunk costs that must be borne to enter

into the export market. In particular, investments to acquire knowledge of the characteristics of the

destination country, such as the legal framework and clients network, are often necessary (Eickelpasch

& Vogel, 2011; Lejpras, 2009; Silva & Carreira, 2016). The prevailing intangible nature of these

assets does not allow to be used as collateral in case of default (Haskel & Westlake, 2017), thus

enhancing the risk of credit constraints (Almeida & Campello, 2007). The greater reliance of firms

on the traditional system of bank lending in many countries does not help overcome the tyranny of

collateral that is inherent in financing intangible investments (Dell'Ariccia, Kadyrzhanova, Minoiu,

& Ratnovski, 2017).

1 The share of services in world trade has increased from 15% at the beginning of the 80s to 20% in 2010 (Jensen, 2016).

|

1217

BRIPI et al.

Variable costs may also affect the international competitiveness of services. Even though ship-

ments are typically rapid in the case of service exports (the delivery is immediate2), production lags

due to product customisation can justify external financing. In the case of some complex services such

as software production, they, could plausibly be longer than that of many manufactured goods.

Moreover, several services tend to be less standardised to fit the specific needs of customers, so that

dedicated costs in product customisation are required (Hoekman & Mattoo, 2012; Vogel, 2011; Wong,

Wu, & Zhang, 2006).

Overall, credit constraints may occur if a relevant part of the costs of providing services is financed

externally due to the existence of high entry costs, the tyranny of collateral in financing intangible

assets, production lags and product customisation. It turns out that this is a likely situation, as various

papers show by calculating the external financial dependence for manufacturing and services.3

It follows that the role of credit frictions on the exports of services is a likely result and has to be

tested empirically. To this aim, we use a high-quality data set obtained by merging matched bank–firm

relationships data, supervisory data on banks' balance sheets and survey data on services exports of

Italian companies (mainly business and personal services4). We focus on a specific event—the sover-

eign debt crisis between 2011 and 2012—to disentangle credit supply from demand factors.5 Our

identification strategy exploits the fact that during the crisis, banks reliant on retail funding were less

exposed to the shock and thus were able to continue to expand credit supply, while the more exposed

banks reduced credit supply (see Section 4 for further details).

The empirical exercise consists of cross section IV regressions of the growth rate of export ser-

vices in 2012 on the change in the credit granted in the same period. The growth rate of credit is

instrumented with the “retail” components of domestic bank funding (the ratio of deposits and bank

bonds held by domestic households over bank assets) lagged one year. The validity of our instrument

is tested by using matched bank–firm credit data, to show that this credit change was higher for firms

financed more intensively by banks with a greater relevance of retail funding (see Section 5).

We find that credit supply had a significant impact on services exports of Italian firms during the

period considered. A bank credit change of 1% induced a rise in exports flows between 0.28% and

0.41%. This result holds true even after the inclusion of various firm controls and several robustness

checks.

Our contribution not only provides the first evidence of the importance of credit frictions for ex-

porting services, but it also extends to other dimensions. To unpack the adjustment mechanisms, we

focus on two features closely related to services. First, we show that the significant role of credit

2 Services are non-storable; that is, the delivery is likely to be immediate (especially if they are traded over the internet); and

this implies the absence of shipment delays (as in the case of goods) that justify exporters’ working capital needs during the

transfer: for a discussion on shipment delays in services trade, see Ariu (2016a).

3 In the Appendix (Table A1), we report as an example the measure computed by Catão et al. (2009) on manufacturing and

services US firms, which clearly shows that external financial dependence is higher in various service sectors than in other

manufacturing ones. For example, “sale, maintenance and repair of motor vehicles and motorcycles, sector code 50,” has a

higher level of external financial dependence than “manufacture of other transport equipment, code 35”: 1.12 Vs 0.69,

respectively. Other works in this vein have similar findings (Balta and Nikolov, 2013; De Serres, Kobayakawa, Sløk, &

Vartia, 2007; Duygan-Bump et al., 2015).

4 In detail, these are computer and information services; other business services; personal, cultural and recreational services;

and communication services.

5 The crisis derived from increased uncertainty over sovereign risk, and subsequently, it affected banks with a different

intensity. Since during that period financial tensions derived from the sovereign debt market (and not by the imports of Italian

services from the rest of the world), we consider this crisis as an exogenous event with respect to the services traded by

Italian firms.

1218

|

BRIPI et al.

supply arises when services are exported to countries with weaker institutional background, where

contractual risks are greater.6 This result is consistent with the “complementarity of inputs” assump-

tion. This is a typical feature of services provision such that the production process ends in the country

where it is consumed (Lennon, Mirza, & Nicoletti, 2009). Moreover, credit supply shocks affect ex-

ports of services that require more complex tasks in production and these effects are only significant

when exporting to countries with weaker institutional backgrounds.

Second, we find that credit constraints affect the exports of services when these are exported as

secondary products by a multi-product firm, typically a manufacturer; the case is known as servitisa-

tion (Breinlich, Soderbery, & Wright, 2018; Crozet & Milet, 2017). While there are various reasons

for manufacturers to provide services, credit needs may arise if marginal costs increase as one moves

away from the product of core competence (Eckel, Javorcik, Iacovone, & Neary, 2015; Eckel & Neary,

2010).

The paper is structured as follows. In Section 2, we briefly review the literature related to trade

services and to trade and finance. The data sets and the sample definition are described in Section

3. In Section 4, we shortly discuss the sovereign crisis, and the empirical methodology is outlined in

Section 5. The main results are presented in Section 6; the robustness checks are in Section 7. Section

8 concludes the paper. An Appendix gives a short description of the survey, discusses sample restric-

tions and reports some additional statistics. Other robustness checks are in Appendix S1.

2

|

RELATED LITERATURE

The literature on international trade in services has grown remarkably in the last decade. Several

papers have analysed firm-level data on exporters and importers of services: Breinlich and Criscuolo

(2011) in the UK, Kelle, Kleinert, Raff, and Toubal (2013) in Germany, Ariu (2016a, 2016b) in

Belgium, and Federico and Tosti (2007) in Italy, among others. Most of these works highlight the

relevant heterogeneity among firms exporting services and, consistently with the consensus in the

international trade in goods literature, that exporting firms are more productive and bigger in size.

Despite the growing interest for firms' exports of services, the literature on trade services and fi-

nance is still very scant, and to the best of our knowledge, only few papers have considered this topic

so far. Biewen, Harsch, and Spies (2012) show that the level of financial development of the exporting

countries did not have a significant effect on services imports by German multinational firms during

the years 2002–08. Borchert and Mattoo (2012) provide anecdotal evidence of Indian firms to show

that the crisis resilience of services trade (relative to the collapse of trade in goods) in 2008–09 was

due to the lower dependence on external financing of services with respect to the production of goods.

Using balance sheet data, Ariu (2016a) shows that services exports of Belgian firms have been quite

resilient to the financial crisis and that services exports were not affected by external finance depen-

dence, nor by long-term financial debt. Our results on the effect of credit for export services depart

from Ariu (2016a), most likely because we study a different country and the sovereign crisis which

did not affect world trade in services, aside from 2008 to 2009.

Differently from these contributions, our work explores the role of external financing on service

trade using very detailed data on the financial exposure of firms, that is, the outstanding credit by

banks to Italian firms exporting services. We match these data with information from the banks'

6 For example, this is the case when the importer's decision to buy foreign services critically depends on local financing

intermediaries (Niepmann & Schmidt-Eisenlohr, 2017).