Links I Liked [Public Procurement]

1.  Government of Western Australia shows Minimum Viable Product version of its new online procurement portal. Great to see a Government taking up on cheaper, more agile ways of developing services instead of building the whole widget without external feedback. Well done.

2. Is raising the micro-purchase threshold from $3,500 to $10,000 in the USA a good idea? In general, I am against raising thresholds for the reasons highlighted in the article, but allowing 18F to expand its micro-transaction platform would not be a bad outcome.

3.  Albert publishes an article about blacklisting.

4. Chicago has an hands on approach to innovation accelerators. Now, if only anyone in Europe would use the innovation partnership for the same purpose... 

5. Speaking of innovation accelerators, the Omidyar Network invested in CityMart. 

New paper: Public procurement financial thresholds in the EU and their relationship with the GPA

I have just uploaded to my SSRN repository a new paper which will be published later this year in the European Procurement and Public Private Partnership Law Review (EPPPL). The paper is focused on the relationship between the EU procurement thresholds and the GPA, arguing that the current threshold levels are arbitrary and effectively set by the EU's GPA commitments. Here's the abstract:

The regulation of procurement within the European Union is binary: above certain financial thresholds, contracts are subject to full EU regulation, whereas below they are only subject to national rules (in general). First introduced in the 1970s, the financial thresholds are arbitrary without a clear justification for their specific values. Thresholds remained fairly stable in nominal terms and over the years became solely dependent on the commitments assumed in the various revisions of multilateral procurement agreements, currently the Government Procurement Agreement (GPA) 2014. In consequence, the external market access commitments accepted by the EU in the GPA determine today the size of public procurement internal market.

While it is true that inflation and currency fluctuations have progressively reduced the real term value of thresholds, no proactive reductions have been undertaken by EU lawmakers, contrary to what was done with trade tariffs. In consequence, current threshold levels do not reflect any productivity improvements or transaction cost reductions achieved during the last 40 years. By remaining stable in nominal and changing only due to external pressures and inflation inertia, the thresholds have effectively functioned as a ceiling and a floor to the concept of internal market in public procurement within the EU.

This is the second paper of my "thresholds trilogy", with the first one focusing on cross-border interest for contracts below-thresholds (also available on SSRN). I am currently drafting the third one which will argue that we should make do with thresholds altogether or else we will never complete the internal market.

Links I Liked [Public Procurement]

1. Today is the last day for registrations for the Public Procurement Early Career Researcher Conference which will be held this Friday at the Center for Transnational Legal Studies in London. Attendance is free.

2. UK government’s data chief wants ‘data-as-a-service’ in Whitehall.

3. Californian hospitals suffer cyberattack. I think it's time for us to drop the word "cyber". but nonetheless this emphasis the importance of security/contingency plans to be taken seriously during public procurement procedures.

4. Open Data Impact case studies. Not strictly procurement, but still relevant for our field.

5. Chromebooks making inroads in education. As I have been saying for some time, in what concerns EU public procurement rules digital services are going to be so cheap sooner rather than later that significant business areas will be subject at most to Treaty principles due to the financial thresholds.

PS: This blog is currently on paternity leave.




New year, new EU thresholds

I forgot to cover this a few weeks ago, but we have new thresholds in place since January 1st:

Works: €5,225,000 (was €5,186,000)

Central services and supplies: €135,000 (was €134,000)

Sub-central services and supplies: €209,000 (was €207,000)

Utilities and defence: €418,00 (was €414,000)

Why do thresholds change every two years? Because that is the commitment the EU made in the context of the revised Government Procurement Agreement, where for third country access to the EU procurement market the thresholds are expressed in Special Drawing Rights.

As for why they exist and their current level, there is still no explanation other than these reflect the commitments the EU made for the GPA. I have a paper on this topic coming out in the very near future, but here's an older one about thresholds

In the UK they have fallen slightly from previous levels but only due to the changes in the the GBP-EU exchange rate.

Thoughts on the European Commission's Strategy to deepen the internal market

The Commission put out last week its strategy for a Deeper and Fairer Internal Market. Part of it is dedicated to public procurement and Albert commented on it here, pointing out the collusion risks of contract registers (which I do not subscribe entirely) and the perennial language problems in cross-border procurement (which I do agree with).

But there is more to drill out of the Commission's strategy. In my view, the focus on compliance is overstated. Most countries transpose the Directives on time and by and large compliance with EU requirements is ok, so that is not really the problem. The focus should have been instead on literally deepening the internal market for public procurement. It is time for the Commission to show some initiative in this area and solve the problem posed by the thresholds. I have a paper precisely about this coming out in the near future, explaining how they came about 40 years ago and how they are now effectively defined by the EU's external commitments with the GPA.

Today, due to the thresholds, only around 18% of public procurement spend is covered by the EU Directives, with the rest subject only to national rules. Well, not "only" national rules as if the contract has cross-border interest, then the Treaty principles of non-discrimination and transparency should be respected. As I wrote back in 2013, that is pretty much impossible to establish in advance and the Court has spent 15 years trying to solve the problem on a case-by-case approach.

There is another reason to deepen the internal market: digital services. By definition, digital services are cross-border (or at least they are not pry to the same transaction costs as other services) and tend to be cheaper than normal services. The vast, vast majority of digital services will come well below the current thresholds. This is for me a critical reason to review them in the near future, as if we do not do so, the digital services internal market will be as "complete" as the regular one.

As for the language issues, well Google translate is getting better and maybe Skype Translator will be the real deal.

EU procurement thresholds: it's time to bring them down

A version of the presentation I delivered yesterday about thresholds on Global Revolution VII is up on the presentation's tab.  For those of you in attendance yesterday, I used Perspective on the iPad to plot the graphics. Took me (and my wife) an eternity to format the spreadsheet tables correctly so that the graphs would turn up as they were supposed to.

If I have the time I may record a re-run of the presentation and put it up somewhere as a screencast as the slides follow my spartan style with limited text on each slide.

As for the content itself, the crux of my argument regarding thresholds and the internal market is that they are a product of compromises based on assumptions done in 1970s and 1980s which no longer hold true. We take them (and their values) as received wisdom and "that they are what they are" and have not taken into account any technological developments which have increased productivity, reduced transactions costs (IT, internet) or marginal costs in many industries. Thresholds remained nominally stable for the most part, although Inflation eroded the real value for goods/supplies. However, inflation has not done so for works (there was a big jump from 1M to 5M for the works threshold in 1989) and only to a limited extent for services (thresholds introduced in 1992).

The internal market thresholds have been glued to the EU's GPA commitments over the years and in my view conceding that the internal market is a common minimum denominator determined not by internal pressure but by those GPA commitments. Why is that?

Directive 2014/24/EU includes in its Article 92 a mandate for the European Commission to review the thresholds in the next three years in accordance with "inflation", "transaction costs" and to increase the GPA thresholds. Well, as our internal thresholds are identical to the GPA ones it is obvious that the real intention here is to increase procurement thresholds across the EU (as suggested by some scholars).

There is a fundamental misunderstanding on the equation that "higher prices = cross-border interest," as cross-border interest by economic suppliers depends on a number of different variables and factors not connected at all with a fixed value. Examples: time of the year; current workload; wider economy; resource availability to take part in tender or deliver contract within timeframes expected; etc. Not to talk about the two big ones: language barriers and legal differences (outside the strict procurement process). I have always been puzzled why private companies seem to be very keen and comfortable with private cross-border trade with suppliers on other Member States, but in public procurement only around 3% of money is spent directly with suppliers based on other Member States.

Raising thresholds would be bad for four reasons:

i) It would reduce what is already a small procurement "internal market" (only 18.5% of procurement spend covered), which may be the outcome some are looking for in the first place...

ii) It would misunderstand that as we introduce more technology into our lives certain classes of products/services will become cheaper, not more expensive. I am speaking about anything that involves computing power, economies of scale and/or bits&bytes.

iii) Connected with ii), it would ignore that the digital economy in the EU is getting close to 10% according to some guesstimates and that we are now finally building a digital internal market and there is nothing intrinsically more cross-border than bits&bytes.

iv) "But, Pedro what about cross-border interest for contracts below-thresholds?" If the cross-border interest is such a good idea why don't we apply it to all public procurement contracts then? When it takes someone as smart as Andrea Sundstrand 6 cases of the CJEU (and where 6 more could have been mentioned) to create a patchwork regime for a "super-procurer" to be able to comply with the cross-border requirement, I do not really need to make a case. If 15 years after Telaustria, we are still discussing what the hell is cross-border interest and how to reliably determine it, something is really wrong with the whole idea. But you can find some further thoughts of me here.

PS: Thanks for the pushback during the presentation and forcing me to think harder about this.

Public Contracts Regulations 2015 - Regulation 13

Regulation 13 - Contracts subsidised by contracting authorities

Regulation 13 takes us out of Sub-Section 3 (exclusions) and into Sub-Section 4 (specific situations) of Part 2. This Regulation covers contracts that are subsidised by contracting authorities, transposing Article 13 of Directive 2014/24/EU into England, Wales and Northern Ireland. As with most provisions we have covered so far, follows a pure copy & paste approach with only very limited differences.

The purpose of this Regulation and the underlying Article is to ensure that some contracts subsidised by contracting authorities are not taken out of public procurement just by the fact the public money is being spent via a subsidy instead of a direct procurement. This provision is only applicable to (some) works contracts and to ancillary service contracts when 50% of the funds come from a contracting authority.

The Regulation is not clear if the 50% funding threshold depends on a single contracting authority or can be cumulatively analysed when different contracting authorities chip in part of the subsidy, together bringing the total subsidy for the contract above the 50% mark. The way the provision is drafted, speaking of "contracting authorities" and its logic (to avoid a loophole) lead me to think that the second interpretation is correct.

Regulation 13 only applies to works contracts such as civil engineering works listed in Schedule 2 or building work for hospitals, facilities intended for sports, recreation and leisure, school and university buildings and buildings used for administrative purposes. I am puzzled about this distinction as pretty much any of the building work mentioned explicitly can be subsumed to the concept of civil engineering works under Schedule 2 (or Annex II of the Directive). However, the same distinction is present in the Directive and as such the responsibility for yet another tautology lies more with it than with the transposition.

Finally, the Regulation introduces a couple of small novelties. Whereas Article 13(1)(a) refers to a threshold of €5,186,000 and 13(1)(b) to €207,000 before the Directive is deemed applicable, the Regulation instead makes a direct cross-reference to the Article, effectively leaving in the hands of EU lawmakers any future update to these values. This is a safe option, however in alternative, the Regulation could have either provided its own values or - more likely - refer to the threshold rules of Regulation 5. As it stands, there is a slight risk that if the thresholds change in the future, this Regulation and the underlying Article may be out of sync with those new thresholds. Again, this is more a problem with the Directive than with the transposition itself.

Albert's entry for today is here.

Public Contracts Regulation 2015 - Regulation 6

Regulation 6 - Methods for calculating the estimated value of procurement

We ended last week talking about the threshold values under the Directive 2014/24/EU. Eagle-eyed Albert found something I missed out on my post, namely the fact that the sterling value is actually set by the Commission "for the time being considered" as indicated in Regulation 5(4). The most current one is as linked by Albert from late 2013 and that remains in force today. However, I would say that four lines of text without a single comma, or sub-paragraphs as common elsewhere in the Regulations, does not make for easy reading or interpretation. Try saying it out loud on one go. In fact, when I read it I assumed the Regulations implied the Commission's power was transitory (ie, "for the time being"). A comma or a couple of sentences/phrases would have been preferable.

We solve the legal issue, but leave the economic intact: currencies diverge over time and setting the exchange for 2 years is simply too long a period. I am not saying there is a easy way out of this conundrum, but as things stand, the forex risk is being borne by the internal market. What would happen if the euro depreciated 50% for example? Or vice-versa?

19 different ways of skinning an estimated value cat (Albert's entry is here)

Regulation 6 offers us a whole cornucopia of different ways to calculate the price of a procurement, which in consequence means they determine the applicable law. Although the first 5-6 paragraphs are straightforward and will be relevant for most cases, most of the time, they do not generate particular difficulties and the overall approach is to ensure that contracting authorities do not game the system by sub-dividing contracts or exclude VAT to avoid applying the Directive. I remember a talk back in 2010 by a scholar from another Member State who had access to procurement data and could pin point a "clustering" of contracts immediately below the thresholds which made no sense other than being an artificial cap to avoid applying more onerous rules.

Paragraph 7 tells us that the moment to define the procurement value is the start of the procurement procedure. Interestingly enough, this paragraph defines that moment not the moment the decision of procuring is taken, but any activity that indicates a procurement procedure will happen even if it indeed does not happen at all. The example of "contacting economic operators" indicates that the start of these preliminary activities will trigger the obligation of defining the procurement value and in consequence the application of Part 2 rules or not. I think this makes sense as it forecloses the possibility of the contracting authority going on fishing expeditions or blatantly lying to potential suppliers by indicating the procedure would be say a Request For Quotes (below-thresholds) when it knows that is not the fact.

Trainwrecks waiting to happen

My biggest concerns regarding Regulation 6 come from the following paragraphs 8 and 9, when the Regulation tries to determine the procurement value for procurement "things" which are not contracts and are hard to measure: framework agreements and innovation partnerships. Particularly the first is a case for deep consideration of how it can be gamed and abused by contracting authorities to avoid applying Part 2 or Part 3 of the Regulations. As mentioned by email by a friend of mine (who's attribution I will give if authorised), below thresholds contracts are now more regulated than in the past but framework agreements are not subject to the same transparency requirements. This creates a perverse incentive for contracting authorities to establish framework agreements with procurement values below thresholds to avoid any sort of transparency. Now let's marry this idea with a short initial term of the framework (say, one year) which coincidently justifies a below-threshold calculation value and that the framework then gets extended to the maximum 4 year period afterwards.

The methods to calculate the innovation partnership values are not much better though and in effect it does not make a lot of sense to do it. If the purpose is really to generate innovation, well then in advance is hard to estimate the determine the R&D and final procurement costs as there are simply too many variables. I suspect, however, that when contracting authorities really want to use the innovation partnership they will just assume a contract value above-thresholds and apply the appropriate rules. Having said that, I have devised at least a way to use the innovation partnership (or something like it) which would be great to run below-thresholds.

The ideas above expose the limitations of using procurement value as a proxy for determining the existence of internal market relevance and applying the Directive and Part 2 of the Regulations.

What about the rest?

The remaining paragraphs of Regulation 6, however detailed, do not offer so much scope for comment. Lots get some attention and detail as they should do to the Directive's preference for division into lots (which will be very welcome in cartelised markets...) and for the most part I agree with the provisions. I do have reservations identical to Albert's about Regulation 6(14)(a) and (b) which exclude the applications of Part 2 in case a supplies or services lot is smaller than €80,000 or €1,000,000 in the case of works. This paragraph does not have a specific cross-reference from this particular thresholds to the Commission's exchange rate calculation role mentioned in Regulation 5(4). Furthermore, Regulation 5(4) restricts the field of its own application to the values "mentioned in this regulation" and not "in these Regulations". Therefore, it seems we may have a gap in the law that will have to be filled in via interpretation. We can probably assume that the Commission will have identical power as this would be the harmonising interpretation and the overall sense of these threshold provisions, but this could have been easily avoided with some more care. As I said at the start, some more time to consider the transposition would have been preferable as this inconsistency is easy to solve. In the meanwhile, I can picture creative contracting authorities using small lots and/or exchange rates fluctuations to exclude procurements from Part 2.

Public Contracts Regulations 2015 - Regulation 5

Albert served a nice ammorti for me to talk about EU thresholds and that is the focus of today's post. Thresholds is a topic I have a lot to say about and will do so in June at the Global Revolution VII in Nottingham. You can find the gist of my ideas in this article from late 2013.

Public Contracts Regulations 2015 - Regulation 5 Threshold Amounts

Regulation 5 keeps it simple by cross-referencing to the application of Article 4(a) and (c) of Directive 2014/24/EU. The only major difference is the inclusion of threshold rules related to defence procurement  which naturally were not included directly in Directive 2014/24/EU.

Article 4 establishes the current thresholds applicable to EU public procurement for contracts/contracting authorities covered by Directive 2014/24/EU:

  • Works contracts - €5,186,000
  • Supplies and Services (central contracting authority) - €134,000
  • Supplies and Services (sub-central) - €207,000

This direct cross-reference to the values in the Directive strikes me as strange. In the past the UK (or at least England, Wales and Northern Ireland) would set in a Procurement Policy Note the conversion into sterling of the euro value for a certain period. It seems we moved from a system controlled by Central Government to a casuistic approach where for every procurement with a value close to the thresholds, contracting authorities will have to calculate the conversion in the days before launching the procedure to assess what if the value is above or below the thresholds. The GBP/EUR exchange rate as fluctuated a lot over the last few weeks due to the Greek crisis and in effect the real thresholds in England, Wales and Northern Ireland have gone down significantly from the Directive's 2014/24/EU original date (26.02.2014):


  • Works contracts - £4,273,160
  • Supplies and Services (central contracting authority) - £110,0413
  • Supplies and Services (sub-central) - £170,564


  • Works contracts - £3,832,959
  • Supplies and Services (central contracting authority) - £99,039
  • Supplies and Services (sub-central) - £152,993

The biggest problem I see with the current approach is that Central Government decided to trade certainty in practice for lawmaking simplicity. It can be argued that the new way is the correct one to ensure sterling values do not deviate from the original euro ones as it happened in the past, but the price to pay is to force every single public procedure close to those values to be manually checked by the contracting authority before tender. Furthermore, for really close call cases it is not clear what is the correct approach to determine the exchange rate applicable: is it the mid-market value? The end of day? Is it the value from the day before launching the procedure or when the decision of launching a procedure is being taken?

Lets remember that this provision as with most of the rules in the Regulations will come into force next week, indicating another rocky point that could have been avoided with a longer vacatio legis. Are contracting authorities aware of this change? I will be pinging a few over the next few days.

Unintended consequences

So far the effect of this new methodology has been one of effectively reducing the threshold values by close to 10% in England, Wales and Northern Ireland. For me this is good news. As per my 2013 article mentioned above, my view is that contracts currently below thresholds actually make part of the internal market and are artificially excluded from it by the arbitrary value of thresholds. But what about the cross-border interest test created by the CJEU? Well, every time that the CJEU looks into the test it tweaks (see C-388/12 Comune di Ancona and Albert's comment on it) but the reality is that this is a test impossible to undertake reliably in practice. It consists of an hypothetical question posed by the contracting authority before launching the procedure: is it likely this contract will generate cross-border interest? That is pretty much impossible to reliably do in practice other than running prior information notices quite often which then defeats the purpose of the test itself. It does not help that every time the CJEU has been asked to clarify matters since the original Telaustria decision the test gets changed ever so slightly. The CJEU appears to either have a difficulty in sorting this mess or enjoys the exercise of defining new ways to skin this proverbial cat.

There is no reason why the threshold values are what they are, other than being close to the Government Procurement Agreement (GPA) threshold values, effectively equating the public procurement internal market to the GPA, which does not make a lot of sense as the GPA is a multilateral agreement with parties outside the EU!

Current threshold values are an historical oddity. The current values made sense in a time where there was no Internet and transaction costs were high. European integration has grown leaps and bounds over the last 30 years and cross-border transaction costs have fallen down significantly. Furthermore, we are finally discussing an internal market for digital services which tend to be cheap, much cheaper than the current thresholds. Somehow the message in procurement is that we have first class and second class industries subject to different rules, not due to nature but solely on price of contract.

Imagine for a moment that the same twisted logic applied to EU freedoms. Looking at freedom of movement, imagine that this freedom would only be applicable if the employment contract had a value above €40,000/year. Below that employment contracts would not be deemed of cross-border interest in first instance but depend on a subjective and unreliable test. I hope I am not giving eurosceptics any ideas for "reforming" freedom of movement here.


PS: Apologies for the lateness of this entry as I spent yesterday dealing with Spanish bureaucracy, never an easy thing to do.