In late December 2016, the European Commission published its Public Procurement Indicators 2015. The statistical information included in this report shows some interesting trends, such as the general increase of procurement expenditure in the EU in 2015 -- which was up by almost 7% from 2014, to reach a total of €450.21 billion -- as well as the continued trend of concentration of procurement expenditure that results from aggregation and/or centralisation of procurement at Member State level.
Regarding the trend towards greater concentration of procurement expenditure in large awards, it is interesting to note that 'at EU level more than one third of the value ... is awarded through contract award notices of 100 million euros or more. This relative concentration of procurement, in large awards, is extremely remarkable in the UK and to a lesser extent in Poland and France. On the opposite side Germany and France concentrate a large fraction of the value procured in the works sector in the smaller size awards'.
This seems surprising because projects of more than €100 million may be relatively common in works (ie infrastructure), as well as large framework agreements for common use equipment (notably, IT hardware), but services contracts of that size would have seemed much less common at first thought (although it is possible that IT expenditure is moving from goods to services as cloud computing and other services are 'virtualised'). In any case, the the fact that the trend is much stronger in the UK than in the rest of the EU (combined) strikes as odd.
Indeed, as the Commission's report stresses, the UK leads the statistics for the award of very large contracts (ie those of a value over €100 million), both for works (66%) and, possibly more remarkable, for services (70%) -- with a smaller but still very significant lead on goods (52%). What is worth emphasising is that the UK's figures are 10 times the magnitude of those for any other Member State (and around 50 to 100 times those of most Member States) both for works and services, and that they double the figures for any other Member State in goods (while still being around 50 to 100 times those of most Member States).
A recalculation of the figures concerning very large contracts excluding data for the UK shows that 22% of procurement expenditure at EU27 level is awarded through contracts of €100 million or more for works and services, and 25% for goods (and, anecdotally, it should be taken into account that a significant part of the latter is attributable to Italy's Consip centralised procurement activities). Thus, the fact that UK alone can move total figures up by 11% (ie a deviation of 50% of the EU27 statistics) seems quite striking.
Moreover, in general, the UK shows a disproportionately high share of contracts advertised in TED (and the estimated value of its contracts is larger throughout the value scale, with many less small contracts and many more large contracts than the EU average, which has an effect on the obligation to publish notices). This is particularly noticeable when compared to other EU countries with large procurement expenditure -- eg in 2015 the UK advertised an estimated 37% of its public expenditure, whereas Germany advertised less than 10% (see graphs below).
In view of these statistical divergences, a closer look at the UK numbers seems necessary in order to try to understand this trend towards concentrated expenditure through very large contracts. However, there is no detailed information in the report on the basis of which to carry out a qualitative analysis.
Many hypotheses are imaginable, such as the possibility that very large centralised contracts are tendered (for example, by the Crown Commercial Service) but they are not necessarily executed to a large percentage of their estimated value, or that the UK is actually significantly more centralised in terms of procurement than other Member States, particularly in services. Each of these possibilities opens itself up for speculation--for instance, about the reliability of statistical information that could include awarded but unexecuted procurement value (which may be very, extremely relevant in the inminent Brexit-related negotiations, as well as for any reevaluation of GPA coverage both for the EU and the UK) or, on the second scenario, about the drivers for such significant differences in centralisation volumes in different Member States and about the possibility of centralised procurement of services in a way that still allows for proper provision of (public) services to end users.
Either way, I find these issues most intriguing and, in case the issue of unexecuted (contracted) expenditure is included in the statistics, I think that more work should go into the collection of actual information and the publication of raw data that allows for refined analysis--ideally, in relation to the 2016 version of the public procurement indicators.