Econometric Analyses of Home Bias in Government Procurement

• DOI: http://doi.org/10.1111/roie.12164
• Published date: 01 February 2015
• Date: 01 February 2015
• Author: Anirudh Shingal

Econometric Analyses of Home Bias in

Government Procurement

Anirudh Shingal*

Abstract

The extent of discrimination in government procurement and its impact on economic efficiency has

attracted both theoretical and analytical work, but little econometric evidence.We bridge this gap by build-

ing a new sector-level dataset on domestic and foreign purchases by Japanese and Swiss governments over

1990–2003 to undertake “new” econometric analyses. Unlike previous work, we explain home-bias using

variables inspired by the political economy, trade-macroeconomic and procurement literatures. We also

provide “new”econometric evidence for previous theoretical predictions. Our results reveal the importance

of domestic-foreign productivity differences in governments’ cross-border purchases and also support pre-

vious theoretical predictions. However, Membership of the World Trade Organizations’s Agreement on

Government Procurement is not found to increase market access.

1. Introduction

Using United Nations–Organisation for Economic Co-operation and Development

(UN–OECD) data, Trionfetti (2000) estimates that contestable government procure-

ment markets account for 7–9% of gross domestic product (GDP) in developed coun-

tries. Thus the state has considerable influence over the allocation of resources in

market economies through procurement. A prominent aspect of such procurement is

the preference for domestic over foreign firms in the award of public contracts despite

cost and quality considerations. This “home-bias” in public purchase decisions has

non-trivial efficiency effects.A home-bias can reduce trade flows and influence inter-

national specialization especially in sectors where public demand is large relative to

domestic output and which are characterized by monopolistic competition and

increasing returns to scale (Trionfetti, 2000).

Factual evidence in support of this home bias has been provided in the literature

(Mastanduno, 1991; Hoekman and Mavroidis, 1997; Trionfetti, 2000; the European

Commission, 1997; Brülhart and Trionfetti, 2001; Evenett and Shingal, 2005; Shingal,

2011). Literature has also evolved to explain this home bias in public procurement

under alternative assumptions of domestic and foreign firm cost distributions and

attributes of the government’s objective function (McAfee and McMillan, 1989;

Laffont and Tirole, 1991; Rothenberg, 1993; Branco, 1994; Vagstad, 1995; Breton and

Salmon, 1996; Naegelen and Mougeot, 1998).

* Shingal: World Trade Institute, University of Bern, Hallerstrasse 6, 3012 Bern, Switzerland

(anirudh.shingal@wti.org). The paper has benefitted from comments made by the Editor on this manu-

script, AlanWoodland, two anonymous referees and by participants at the annual conferences of the Euro-

pean Trade Study Group, Copenhagen, the Canadian Economic Association, Calgary and the PEGGED

Research Workshop onTrade, Villars. I would also like to thank the Swiss NCCR in Trade Regulation for

financial support and Simon Evenett, Peter Holmes, Michael Barrow, Michael Gasiorek, Marius Brülhart,

Justina Fischer,Marcelo Olarreaga, James MacGee and Joseph Francois for helpful suggestions at different

stages.

Review of International Economics, 23(1), 188–219, 2015

DOI:10.1111/roie.12164

© 2014 JohnWiley & Sons Ltd

Another branch of the procurement literature has looked at corruption and bribery

as determinants of this home-bias (Rose-Ackerman, 1975, 1978; Burguet and Che,

2004; Compte et al., 2005; Burguet and Perry, 2007). Finally, Baldwin (1970, 1984),

Baldwin and Richardson (1972), Miyagiwa (1991), Chen (1995),Francois et al. (1997),

Trionfetti (1997, 2001), Weichenrieder (2001), Brülhart and Trionfetti (2001, 2004),

and Evenett and Hoekman (2005) have studied the consequences of a procurement

home bias on trade and specialization.

However, this literature is largely theoretical and analytical in nature and with the

exception of Brülhart and Trionfetti (2001, 2004) provides no econometric evidence.

In this paper, we bridge this gap by building a new dataset from WorldTrade Organi-

zation (WTO) notifications on domestic and foreign purchases by Japanese and Swiss

governments at the sector level over 1990–2003 to undertake “new” econometric

analyses. Our choice of countries is primarily determined by data availability. Both

countries have submitted detailed procurement data to the WTO sufficiently regularly

over 1990–20031and in a form amenable to econometric analyses.

Departing from previous studies that either provide factual evidence on the home

bias or model this theoretically, in this paper we explain the determinants of a home

bias econometrically using variables inspired by the political economy, trade–

macroeconomic and procurement literatures. Review of relevant literature suggests

the importance of these factors, but to the best of our knowledge, this area remains

unresearched. Concurrently, we also test for the effectiveness of the WTO’s Uruguay

Round Agreement on Government Procurement (URGPA) in increasing foreign

access to public markets.

Finally, we provide “new” econometric evidence for the theoretical predictions in

Baldwin (1970, 1984), and Baldwin and Richardson (1972) on the ineffectiveness of

discriminatory procurement and their examination in an oligopolistic set-up

(Miyagiwa, 1991) and in monopolistic competitive markets (Brülhart and Trionfetti,

2004). Other theoretical work on this subject includes that by Chen (1995), Trionfetti

(1997), Weichenrieder (2001), and Evenett and Hoekman (2005), but no empirical evi-

dence.The part exception to this is the study by Francois et al. (1997), which compares

public and private demand across 85 US industrial sectors and infers the ineffective-

ness of home bias from the “smallness” of public demand. Our database and econo-

metric approach enable a more complete assessment of these theoretical results.

Using EU data, Brülhart and Trionfetti (2001, 2004) also show that procurement

home bias matters for industrial location/specialization, but unlike us, their focus is

not the Baldwin–Richardson–Miyagiwa–Brülhart–Trionfetti (henceforth, BRMBT)

theoretical predictions.

Thus, to the best of our knowledge, we make several original empirical contribu-

tions to this literature,2the need for which has been long acknowledged (for instance,

see Mattoo, 1996). The propensity to source from foreign firms is generally found to

vary negatively with productivity differences in favor of domestic firms in our results,

which signifies the importance of efficiency considerations in these governments’

sourcing decisions. However, the expected impact of macroeconomic factors and

political budget cycles does not come through. Significantly, our empirical results

largely support the BRMBT theoretical predictions, but membership of the WTO’s

Agreement on Government Procurement (GPA) is not found to increase foreign

access to either country’s public markets.

In sum, this paper has two objectives: (i) explain the determinants of a procurement

home bias econometrically using variables inspired by the political economy, trade–

macroeconomic and procurement literatures; and (ii) provide “new” econometric

HOME BIAS IN GOVERNMENT PROCUREMENT 189

© 2014 JohnWiley & Sons Ltd

evidence for the BRMBT theoretical predictions in government procurement. Objec-

tive (i) is motivated by an analysis of the relevant political economy, trade–

macroeconomic and procurement literatures in section 2, which directly leads to five

testable hypotheses (section 2, last paragraph). These hypotheses are examined

empirically in section 3 using estimating equation (1) which itself is based on the

analysis in section 2 thus providing the theory-empirical connect. Objective (ii) is

embedded in the BRMBT theoretical propositions discussed in section 4, with esti-

mating equation (2) used to provide evidence on the theoretical predictions. Thereaf-

ter, section 5 looks at the data. Sections 6 and 7 discuss estimation issues and the

results from estimation. Section 8 concludes.

2. Determinants of a Home Bias in Public Procurement

The procurement literature has developed models showing that discrimination may

be necessary to minimize average procurement costs. McAfee and McMillan (1989)

consider one such case with asymmetric and lower foreign firm costs and a low

number of bidders, both domestic and foreign. They show that in such situations,

foreign firms may exploit their cost advantage by bidding just below the expected

domestic firm bid and hence, discriminating against them may be welfare enhancing.

Even when costs are assumed to be symmetric and the social costs of distortionary

taxation are accounted for, Branco (1994) has shown that in so far as foreign firm

profits do not enter the government’s objective function, it will be optimal for the gov-

ernment to prefer domestic suppliers, but unlike in McAfee and McMillan (1989)

where a home bias only results in foreign firms lowering their bids, in Branco (1994),a

home bias results in home purchases. A review of this theoretical literature thus sug-

gests that domestic–foreign cost differentials (cd−cf) and attributes of the govern-

ment’s objective function have a bearing on the choice of suppliers.

Other evidence discussed in some detail by Evenett and Hoekman (2004) suggests

that the preference for domestic suppliers may be driven by a range of procurement-

specific and domestic policy factors. These include the nature of the good or service3

being procured; the value of the procurement contract;4the extent of domestic com-

petition;5practical considerations of the tender;6compliance costs;7regulatory

burden;8and the domestic policy environment.9

Governments are also saddled with other objectives owing to which the award of

contracts is not always governed by efficiency considerations, further embedding dis-

crimination in procurement. For instance, governments may need to use procurement

as a tool of aggregate demand management to stimulate Keynesian multiplier effects

in an economy. Procurement home bias has been documented in major economies

during the recent economic crisis (Evenett, 2009a,b) which suggests that the propen-

sity to spend at home may be stronger during recessions. There are several variables,

used commonly in the trade and growth literatures, that can serve as reasonable

proxies for such determinants for, e.g. GDP growth, unemployment rates and real

exchange rates. In fact,changes in the latter may influence the choice of suppliers both

directly (a stronger currency also makes public imports less expensive) and indirectly

[by influencing foreign firms’ participation decisions (Baldwin, 1988; Baldwin and

Krugman, 1989; Campa, 2004) and via changes in firm mark-ups (Berman et al.,

2012)].

A change in procurement metrics can also be dictated by political economy com-

pulsions such as governments being in an election cycle. The impact of political insti-

tutions on economic policy has been studied extensively in the political budget cycles

190 Anirudh Shingal

© 2014 JohnWiley & Sons Ltd