VoxEU Column EU policies Global governance Institutions and economics

Public procurement markets: Where are we?

Public spending on large-scale projects is often a way of sneaking in protectionism through the back door and there are many cases of outright corruption. With the EU and US pushing hard for more open public procurement elsewhere in the world, this column asks just how open these markets are, particularly in the EU, which claims to have the most open market in the world.

Recently, the EU and US have pushed very hard for opening public procurement markets, as illustrated by the EU and US pressures on Japan and China, respectively. In particular, the EU claims that it is by far the most open market in the world. In March 2012, this belief has induced the European Commission to request from member states a mandate for closing EU public procurement markets to firms originating from countries using ‘restrictive practices’ in this domain – the so-called ‘reciprocity’ approach.

In such a conflicting context, the first thing to do is to test the claims. This requires a robust standard for measuring openness in public procurement markets. Existing claims rely on (quite opaque) information on commitments under the existing Government Procurement Agreement (GPA). This approach is flawed for two reasons.

  • During the negotiations on an improved Government Procurement Agreement (December 2011) all the countries – including the EU and the US – have agreed that existing data on Government Procurement Agreement commitments are so bad that they require a huge effort to improve them (Anderson 2012). Indeed, they are in such a pitiful state that it is impossible to reconcile them (Anderson et al. 2011).
  • More deeply, a measure of openness based on GPA commitments does not make sense. It is equivalent to measuring openness in goods by focusing exclusively on the tariff cuts which have been subjected to a negotiation while ignoring the rest of the tariff schedules left untouched by the deal. Rather, openness in trade in goods is routinely assessed by the whole range of products described exhaustively in the common tariff nomenclature (harmonised system). However, building a common nomenclature for public procurement is a task beyond reach because GPA commitments are defined not only in terms of economic activities (products and services) but also in terms of thresholds for bids and of an endlessly wide range of public entities included in the commitments.

There is thus a need to find an encompassing and robust definition of public demand that covers every cent spent by a public administration or an entity considered as a public agency – including utility sector bodies – on domestic and foreign goods and services. Foreign public procurement corresponds then to imports of goods and services absorbed by such public demand. National Accounts provide this framework to calculate public demand and imports (for details, see Messerlin and Miroudot 2012).

A new data source: The world as it is – not as it is claimed to be

The recent World Input-Output Database (Timmer et al. 2012) provides such data for the world's 14 largest economies (EU being one) and for the individual 27 EU Member States. It covers all the years since the 1995 Uruguay Round GPA.

The WIOD data provide a picture very different from the EU and US claims. Figure 1 plots the openness ratios (imports as a share of public demand) of the 41 countries against their GDPs in 2008.1 For the EU members, only imports coming from non-EU sources (‘extra-EU’) are taken into consideration – as it should be. The EU27 public demand and extra-EU imports are the sum of the individual EU member states’ corresponding figures. For illustration sake, Figure 1 labels the most important countries, and shows a correlation between GDP size and openness ratios.

Figure 1 provides two clear results:

  • The EU27 looks more open than the three largest EU member states, except the UK. As small economies are expected to have more recourse to foreign suppliers of public procurement than large economies (because of their narrow production base), openness ratios tend to be inversely related to economic size, as for goods.
  • As the EU has no significant ‘federal’ public procurement, the EU27 openness ratio mirrors the fact that decisions on public procurement have been made by 20 or so small economies and a handful of medium-sized economies.
  • The EU27 openness ratio is thus biased upward. (A similar bias may exist in the US, but it is probably weaker because of the huge US federal public procurement markets.)
  • The EU27 and the US are significantly less open than the East Asian economies – China, Japan, Korea, and Taiwan – targeted by EU and US pressures.

Observations at the EU member states level taking into account the size effect confirm this conclusion: Germany is less open than Japan or China, despite its significantly smaller GDP.

Figure 1. GDP and openness ratios, 2008

Source: Timmer (WIOD) 2012. Author’s calculations.

The world has changed profoundly

The evolution over 1995-2008 shows how fast the world is changing. Figure 2 focuses on the EU-Japan case. It is limited to the two largest EU member states behind EU pressures on Japan (Germany and, above all, France). Their smaller sizes (0.75 and 0.60 times Japan’s GDP) should make them be significantly more open than Japan. That has never been the case for France. Since 2004 Japan has caught up fast on Germany and it became more open than Germany in 2008.

Figure 2. Openness ratios, Japan, France, and Germany, 1995-2008

Source: WIOD 2012. Authors’ calculations.

Figure 3 focuses on the strong pressures on China from the US and EU. China has been significantly more open than the US and the EU27 since 2002, that is, precisely when its GDP was catching up the EU and US GDPs.

Figure 3. Openness ratios, China, EU and US, 1995-2008

Source: WIOD 2012. Authors’ calculations. Note: EU2 extra is the openness ratio of France and Germany combined.

The period 1995-2008 suggests two final observations:

  • The US openness ratio grows at a lower rate since 2003, long before the proposed changes of the Buy American Act tabled during the 2008 crisis.
  • French and German intra-EU openness ratios have slowed down since 2003, even decreasing in 2008. This relative evolution may reflect sluggish growth. But it may also reveal serious problems in the functioning of the internal market in public procurement.
Policy implications

The above evidence suggests three policy implications in public procurement:

  • Unsubstantiated claims should give way to balanced negotiations. No country is wide open and very few are closed (with the exception of Brazil and Russia where openness ratios are very low).
  • Negotiating on the basis of a global EU27 openness is meaningless in case of no ‘federal’ structure like in the EU where the openness ratios among the 27 EU member states vary by a factor of almost five. It is counterproductive from the EU’s own perspective because it hides the need to improve the functioning of the internal market in public procurement by opening the most closed EU members not only to non-EU firms, but also to the efficient firms originating from the most open EU members.
  • If the observed lower EU openness based on national accounts were to coincide with deeper GPA commitments, the EU situation would be similar to one of a country with very low tariffs but limited imports – a situation similar to Japan in trade in goods. Such a situation has two possible explanations. First, there may be good reasons – in the case of public procurements, granting bids to the firms best known because they are geographically close or more trusted. Second, there may be bad reasons – hidden 'non-tariff barriers' in EU member states' public procurement markets. Both explanations should strongly induce the EU to stop using claims on its alleged but unobserved openness for requesting down-payments, and rather to start to sit at the negotiating table for the benefits of both parties.

Even more crucially, the above evidence strongly suggests that the EU stance of ‘reciprocity’ is doomed to fail because it relies on two assumptions that are not fulfilled.

  • First is the higher openness of the EU. Above evidence refutes such claims, in particular when coming from the EU member states the most vocal about reciprocity.
  • Second, the threat of closing EU markets assumes a rapport de force. The credibility of such a threat is already gone: the combined French and German public demand, which was almost eight times larger than China’s public demand in 1995 and three times in 2000, was only 1.3 times larger in 2009.

Insisting on reciprocity in such a fast moving and deeply changing context is a sure recipe to induce powerful EU trading partners to strike back with similar and increasingly credible threats, engulfing EU firms in extremely costly retaliation wars for them.

The views expressed are only those of the authors, and do not reflect in any way those of the OECD Secretariat or the member countries of the OECD.


Anderson, RD, P Pelletier, K Osei-lah and AC Muller, 2011. "Assessing the value of future accessions to the WTO agreement on public procurement: Some new data sources, provisional estimates, and an evaluative framework for individual WTO Members considering accession", Staff Working Paper ERSD 2011-15. WTO Secretariat, Geneva.

Anderson, RD, (2012), "The conclusion of the renegotiation of the WTO Agreement on Government Procurement: what it means for the Agreement and for the world economy. Mimeo, WTO Secretariat. Forthcoming in the Public Procurement Law Review.

Messerlin, P and S Miroudot, (2012) "EU public procurement markets: How open are they?" mimeo, July, Groupe d’Economie Mondiale (GEM), http//gem.sciences-po.fr

Timmer, M. et al. 2012. The World Input-Output Database (WIOD): Contents, Sources and Methods. Available at www.wiod.org.

1 All the individual non-EU countries and most EUMS show a marked decline in the openness ratios in 2009. The fall is huge – higher than 10% except for a few cases. This evolution is, at least partly, related to the stimulus packages following the 2007-2008 crisis. Such packages have had a ‘domestic bias’ either because they have focused on public demand in sectors having relatively low foreign penetration (a mere composition effect) or because they have used procedures discriminating against foreign competitors (a protectionist effect).

1,785 Reads