Engaging with the NHS on data driven health tech. Change is coming to the NHS (and right about time after WannaCry last year…)
Although the general deadline for transposition of the 2014 public procurement Directives was April 2016, Article 90 of Directive 2014/24/EU provided the Member States more time to get their act together in what concerns electronic procurement.
Well, the deadline for transposition of the remaining electronic procurement obligations contained in the Directive is today.
Now the million dollar question in the Member States that are yet to transpose (or implement) such provisions is: which have direct effect?
Last week, Bloomberg ran an article claiming the USA and two other countries were blocking the UKs accession to the GPA agreement. Yesterday, it doubled down on the story stating New Zealand and Moldova as the two other members blocking the UK. and provided more information about why Moldova is making life difficult for the UK. The Moldovan reasons are simply delicious and a prime example of asymmetric retaliation. In hindsight, they capture beautifully the zeitgeist of Brexit. All in all, what myself and Albert predicted about a year ago in our paper is panning out: UK going for a straightforward accession as possible but with the flank exposed to demands from current members.
So far it seems that the current members are willing to run down the clock to November 27th when the WTO government procurement meeting occurs. To be fair, there is no reason or incentive to do otherwise for a number of reasons. First, because the UK is not leaving the European Union until March 29th, 2019, so there may be time for an agreement until then. If ratifications are required, then agreeing now or in March does not make a significant difference.
Second, the longer the uncertainty lasts the weaker the UK bargaining position and the more willing it will be to make concessions. And herein lies the rub: those demands for concessions can come from anywhere in the spectrum of interests of the other members, effectively meaning they may be completely unconnected with procurement. Procurement is simply being used as leverage to obtain concessions elsewhere (again, read between the lines of the Moldovan reasons…).
Finally, contrary to popular perception, the UK procurement market is not that open to foreign bidders. Only large contracts are subject to the GPA rules and those tend to be of interest to large companies. And which countries have large companies operating in foreign public procurement markets? Above all, two: USA and the UK. So, the USA is effectively reducing competition for procurement contracts inside its market and also - probably more crucially - taking key players out of competition abroad. So for the USA it makes sense to make life as difficult as possible to the UK unless really good sweeteners are thrown in (NHS privatisation anyone?). So for the price of losing access to the UK market the USA is blocking competition in all other markets (exception may be EU of course) as the UK also has no Free Trade Agreements in place. As for Moldova, it sits on the other side of the spectrum. It knows its companies stand no chance in hell of winning contracts in the UK so why open its home procurement market for free? Better to try and win a concession elsewhere like, say, visas.
Overall, I suspect the overarching interest of all parties will lead to a deal sooner or later, but so far we’re still in the multidimensional chess part of the game.
PS: The irony of New Zealand being the third blocker is not lost on me. Eat your hat, brexiteers.
This is a slide I will be using on Monday’s EU law lecture. I have not drilled down into the details, namely if there is any evolution or if there has been an increase on Commission’s delegated Regulations but I was suprised at the findings.
The EU legal secondary law framework is mostly one of Regulations and not Directives, with all the implications that brings.
The UK Government published recently a guidance note on the potential impact for public contracts access in case there is no deal with the EU before March 29 2019. There is not really much actionable information and perhaps calling it “guidance” is a slight misnomer as the document is more of a “heads up, this may happen” type of document.
Post March 29 2019, the Government implies that UK contracting authorities will be using a new UK-based e-notification service instead of OJEU/TED. However, there is no information whatsoever about this new service, who will set it up, by what date and how it will operate. In short, it adds no legal certainty to the implications of the UK departure. It might have been preferable to simply refer to the need to use Contracts Finder and similar regional portals for the time being instead of re-inventing the wheel once more.
Looking into the part about procedures ongoing at that date also yields reasons for concern. Here’s what the guidance contains:
“There will be more engagement on about how to deal with ongoing procurement procedures in the handover period between the two systems nearer the time. This will be described via appropriate communication channels and in guidance, which will be made available on GOV.UK.”
Again, not exactly reassuring. What will happen to those situations whereby the contracting authorities (and suppliers) are reliant on the European Single Procurement Document or e-Certis and associated databases to get data about economic operators taking part in ongoing procedures? And would the EU economic operators (and GPA ones) lose their status halfway through the procedure?
Finally, a word about the GPA. The guidance confirms that the UK is seeking individual accession to the agreement (as forecast by myself and Albert Sanchez-Graells). As the accession request was submitted in June 2018, the process will take time and it is simply impossible that the accession would be wrapped up by the end of March 2019. In consequence, UK economic operators would not only lose access to the EU procurement market but also to those of GPA members. As for economic operators from GPA countries, it would be up to the UK to decide how to treat them, but even if they were admitted to tendering doubts will remain about their eligibility for remedies.
Liz David-Barret and Mihaly Fazekas have a new research project called Curbing Corruption in Government Contracting, funded by the Department for International Development Anti-Corruption Evidence Programme. The project aims to look at how corruption can manipulate procurement and strategies to identify variables, patterns and trends that may indicate a corruption risk.
As the project evolves, it will be possible to find on the website working papers, policy briefs, datasets and a stream of blog posts on their work. You can find them as well on Twitter.
The House of Commons Library has just published an interesting new briefing paper on public procurement and contracts. It is an introductory text, but one that sets up the scene well to explain how public procurement works in the UK.
The briefing paper can be found here.
Arecent decision by the Technology and Construction Court (Lancashire Care NHS Foundation Trust v Lancashire County Council  EWHC 1589) is a tour de force on how contracting authorities are often still unable to deal properly with the need to have good decision making models *and* keeping track of their decision making processes. Discretion does not mean arbitrariness...
Even bearing in mind that the contract at hand was for care services and as such subject to the "light touch" regime of Articles 74-76 of the Public Contracts Regulations 2015, that does not mean contracting authorities can simply run the procedure as they see fit. The award decision must follow the award criteria disclosed (it did in this case) and any element of the decision that is material needs to be logged and justified.
It is also worth noting the importance that should be given to moderation when multiple individuals are involved in the assessment process and on this point paras 30-40 of the judgment. are scathing "[a]s this summary shows, there was no consistency either in identifying what were said to be key points or in highlighting points to show that they had been influential. The approach differed even within the record of the same question(...)."
Here's an article (in Spanish) that caught my eye this morning. A couple of interesting tidbits:
"The Spanish cleaning companies Association and the Unions trust that the new Public Contracts Law covers labour costs and to guarantee the social rights of workers in all tenders from the various contracting authorities."
Here's what is meant by it, in the words of the Association president:
"...to include in the award criteria various elements related to the service quality, such as the working conditions offered by the companies..."
I find this article interesting for a couple of reasons. First, it is very uncommon to see both business and unions agreeing on labour costs. Second, it is possible to explain it by looking at the incentives and how they are both aligned in this instance. Let's start with the companies.
Cleaning services are incredible price sensitive and (as it is claimed) 90% of the cost incurred with each contract is simply labour costs. Companies hate competition and honestly price is the most liquid of comparators wether we like it or not. Since those contracts tend to be awarded based on price, if the costs is essentially fixed (the minimum wage) then they are effectively competing in the narrow sliver of their margin (those 10%) and that is where it hurts. No wonder they want to either take price out of the equation or dilute with "service quality" criteria. More about this in a second.
As for the unions, they simply want a better deal for their members and there is nothing wrong with that, so they also want prices to rise assuming they translate into higher wages (they won't) or at least better working conditions for cleaning staff.
So, both parties interests are aligned in reducing price importance in the equation. In other words, both want the contracts to get more expensive. One side wants better margins, the other either more pay or better working conditions.
About the "service quality" then. The second citation above is a direct citation. The example of service quality provided by the President of the cleaning companies association has nothing to do with service quality (well, at least not directly) but with working conditions instead. To conflate the two is disingenuous to say the least. Working conditions are a problem for companies like cleaning services companies due to attrition and costs of training/recruiting new staff.
It is not surprising for me that he did not pick up other award criteria for quality. No mention of efficiency (although price is a proxy for it), availability/turn around of staff in case of spike in cleaning needs, technology to manage the contract/communications, etc. It may well be, however, that he did mention them but the reporter chose that tidbit instead.
In any case this is a roundabout way of solving the fundamental problem: wages are probably too low. It will lead to worse outcomes than solving the fundamental problem. And even then, let's be honest and assume that solving that problem implies higher taxes. There are no free lunches in public procurement.
I came across this paper on Hacker News yesterday and decided to have a look at it since it merges two of my interests: distributed ledgers and procurement. The paper by Hardwick, Akram and Markantonakis proposes a theoretical framework for a tendering system based on smart contracts, running on the Ethereum blockchain.
I cannot comment on some of the substance, particularly the security arguments presented, but keeping to the theoretical framework for now I have two comments on (i)mutability and the idea behind smart contracts to provide.
(I)mutability of a tender is not a feature, but a bug.
The authors approach the problem of public contract tendering from a vector broadly in the citizen participation/open governance/system integrity/trust scope. Nothing wrong with that, and it is refreshing to see new takes on existing ideas without the constraints imposed by the current mental models.
On p.4 the authors put forward as a security feature of their proposal:
"R1) The tendering Organisation cannot change the tender once it is placed on the blockchain. If due to some unforeseeable reasons they have to change it, then they have to create a new tender (smart contract) on the blockchain."
Keeping legal considerations out of the way, I can understand the theoretical underpinnings of this proposal for immutability. However, it is very problematic on a practical basis: time and opportunity cost for everyone involved. Going back to square one every single time a minor error with the tender is detected or due to further information having been requested by the bidders is simply inefficient and disproportionate. If the objective is to ensure the security of the process and enable auditing ex post facto, this could be achieved by allowing changes to be made to the smart contract by the contracting organisation as long as all those changed states are tracked/timestamped as well. This would allow in my view to achieve the same objective but with lower transaction costs for the parties.
Smart contracts, not contracts
Next semester I will be teaching law and blockchain (also smart contracts) and this paper is a timely reminder how prose language can be tricky to interpret. What is described here as a smart contract (ie, the set of operations between launching a tender and awarding the contract) is perceived in legal terms not as a contract at all, but pre-contractual actions which may or not impose obligations on the parties (it varies from jurisdiction to jurisdiction) but not those of a contractual nature. The contract is the end result and usually it only involves two parties (the contracting body and the private party), not all participants in the procedure.
In essence as far as I can define smart contracts they are a set of scripts that run automatically in the same way they already exist in other areas that can lead to the formation of a contract (stock buying/selling orders at a pre-defined value come to mind). They are not (necessarily) by themselves a contract. Having said that, depending on the jurisdiction it is possible that smart contracts really are contracts assuming a) all the elements for contract formation are present (again, jurisdiction specific), or b) definition of what constitutes a contract changes in the future.
Where could this be used?
Coming back to the idea behind the paper - running public tenders on Ethereum blockchain - and where it could be used. My view about the use cases for blockchain in general is that it will be useful and used in scenarios where no other current technological alternative has been deployed or where trust issues are so profound that a centralised database would be more efficient but cannot be used due to lack of trust in the operators.
In both cases I cannot see countries which have already adopted centralised electronic procurement systems to jettison those in favour of a blockchain based system any time soon. Let's not forget the public sector tends to be a laggard in terms of technology and that even today electronic procurement is not mandatory in many countries (EU, I'm looking at you...) even though buying over the internet has been commonplace for more than 15 years. Small elements of what is proposed in the paper are being trialled in Aragon (Spain) but only as a mechanism to timestamp the delivery of bids and I have in the past called for a similar system to be created to gather contract feedback data.
I suspect the potential bulk of use cases will thus be in the developing world, especially those that do not currently have electronic procurement systems and where trust is a paramount issue. It is no surprise that mobile banking (on feature phones) took off in countries without strong bank branches penetration or even internet. And if banking can be done on low bandwidth phones, there is no reason why tendering could not follow a similar development path. Plus, on the long run piggy backing on a major blockchain like Ethereum may lead to a cheaper alternative than the set up and management of a centralised system.
Another potential scenario for use is for regime transitions, ie countries that have changed regimes and where the new one wants more transparency, quickly and where trust is - once more - an issue. Ukraine takes great pride (and rightly so) on developing and using ProZorro after the Maidan Revolution but the next Ukraine may find it easier and cheaper to achieve a similar result but without the need to deploy a centralised system.
The Public Accounts Select Committee has published a report into the risk assessments of Carillion in the run up to its failure. Here's a snippet:
"The Carillion assessments show that:
- Although Carillion had been rated Amber owing to performance against contracts with the Ministry of Defence and Ministry of Justice, it was not until after Carillion issued a profit warning in July 2017 that Government downgraded Carillion to Red. It appears the Government was not aware of Carillion’s financial distress until this point.
- In November 2017, officials recommended a provisional Black rating for Carillion. However, following representations from the company, the Cabinet Office did not confirm the designation. Carillion collapsed less than two months later."
On a very strange (but welcome) transparency note, even the assessments themselves were made public.
What it appears to me is that by November 2017, plans should have been privately put in motion to forestall a potential (though not certain yet) entrance into administration, ie how to handle ongoing contracts, securing access to sites, etc. That for me is much more important than knowing exactly when the rating moved from amber to red and then finally to black.
On a final note: who and when had access to these risk assessment reports, ie only the Cabinet Office or the wider public sector which could have been tendering contracts with Carillion as a bidder.
Somehow I missed this excellent blogpost by Peter Smith last week about the Public Administration and Constitutional Affairs Committee hearing on public sector outsourcing. The whole thing is worth reading, but here's an interesting bit:
"Asked about contracts that are in difficulties, “we have re-priced in some case” says Manzoni. He then backs off somewhat and says “we have to be careful with regulations”. Really? Tell us more, do explain where you have broken the law! “Several I can think of where a re-pricing has taken place, where we have got it wrong”."
I do not fully agree with Peter's immediate conclusion of illegality. It may be that what is meant in this context is that the service delivery changed and as such the prices changed as well. This would be legal (within the limits provided for in the Regulations). Even a pure price change might be legal if it met the three criteria mentioned by Peter in his post and present in Article 72 of Directive 2014/24/EU:
(i) the need for modification has been brought about by circumstances which a diligent contracting authority could not foresee;
(ii) the modification does not alter the overall nature of the contract;
(iii) any increase in price is not higher than 50 % of the value of the original contract or framework agreement. Where several successive modifications are made, that limitation shall apply to the value of each modification. Such consecutive modifications shall not be aimed at circumventing this Directive;
The burden of proof for the conditions to be met remains with the contracting authority, so in this case evidence would have to be provided that a diligent contracting authority could not foresee the need to increase the price. An of course, the more times the repricing happens to any given contracting authority the less likely it is the test will be met. But who would put forward a complaint even it that was the case?
In the bigger picture, this is yet another example of what I have been harping for years: the price of regulatory focus in the procedure has moved the pressure points (corruption/illegalities/violation of competition etc) to other areas of the system, namely contract performance and pre-tender launch.
It is also evidence of the winner's curse at play, but we have known that for ages.
Very interesting piece on Wired about how Barcelona is dealing with smart city surveillance, even if I do not buy the whole political worldview.* On the procurement side, this bit at the start caught my eye:
“Now we have a big contract with Vodafone, and every month Vodafone has to give machine readable data to city hall. Before, that didn’t happen. They just took all the data and used it for their own benefit”
I will take this at face value, but even so it shows the importance of understanding where value (and risk) lies. By giving Vodafone free reign on using the data generated in that contract the City Council was effectively paying them twice for the same service: first, in cash. Then, in data Vodafone could use as well for its own purposes. That the current City Council understands that the value generated by its contract is valuable (and also a key reason why incumbents usually have a built in advantage IMHO) is a welcome development.
In general I am in favour of more, not less transparency even though it is not exactly risk free in some markets due to the collusion opportunity it offers. But my experience in public procurement tells me that more detailed data provided to tenderers helps them reduce uncertainty and provide more detailed bids based on that data (it just so happens it might as well help collusion).
There is another important point to think about here as well and that is the potential State aid implications. If usage data has value for the incumbent and it is already being paid to deliver the contract then it is arguable the intrinsic data value goes beyond the market rate and that might constitute a case of implicit (?) State Aid.
The fact the data is controlled by the City Council and (hopefully) made available to tenderers in the following tender allows to level the ground between the incumbents and challengers, negates part of their inbuilt advantage and the value the data has for the first.
*The idea behind Barcelona as a smart city predates 2015 and the current preoccupation with data ownership as well.
In the wake of Carillion's implosion, we are coming to see the internal workings of the UK Government in a light that might be surprising or unexpected for those looking from the afar. The most recent piece of information comes in today's Telegraph and is connected to what award criteria are used. Mostly, lowest price even in large outsourcing contracts.
The Public Contracts Regulation 67(1) establishes - correctly - the principle that contracts need to be tendered in accordance with the most economic advantageous tender. 3 years after the Regulations came into force, lowest price or price only contracts should be a thing of the past. But they are not, why?
First, as mentioned in the Telegraph's piece, there is a huge pressure on budgets and that means the pressure is passed on to the private sector via price only contracts. If a large percentage of a pool of contracting authorities operating in a market all have the same approach, then margins of economic operators are indeed squeezed. By itself that is not a problem and is part and parcel on the economic world. No company likes competition, so take with a grain of salt the tears claiming contract prices in the public sector are too low. Having said that...
The winner's curse exists really is a thing in public procurement. If only one contractor can win and the lowest price is the award criteria, there is a huuuuuuge pressure to win the contract at any cost - and try and make the difference up during the contract. In effect, quite often the prices presented to win the contract are not sustainable, ie, will not cover costs and allow for a profit to be made. Have too many of those, a little bit of a head wind and the positive cashflow won't be enough to keep the ship from keeling over. In effect, this is roughly what happened to Carillion (and may be happening to Capita as well).
In consequence, I would argue there are a lot more tenders that are abnormally low than those that are formally checked for that condition. In other words, there are plenty of apparently normally priced tenders which are effectively abnormally low depending on the financial circumstances of the tenderer and eventually other contracts the economic operator has already won or will be winning in the contract. In a world where the majority of winning tenders are abnormally low, when looked in context they will all look, well, normal.
That is only part of the story though.
Second, a longer term problem looms: lack of capacity in the public sector to deal with tender complexity. Lowest price tenders are easier to assess/compare and the corresponding decision to justify. It is no surprise they remain very popular in the UK (and elsewhere too), making irrelevant the default MEAT rule in both the Directive 2014/24/EU and the Public Contracts Regulations 2015.
This brings me to the final point: the UK pushed for significant changes to be introduced in Directive 2014/24/EU, which then led to the creation of the competitive procedure with negotiation and the innovation partnership amid other measures that I classify as being "for the 1%" of contracting authorities.
What Carillion's demise is showing is that not even the "1% of contracting authorities" have the chops to do the 'advanced' basics well (moving from lower price to MEAT) let alone deal with all those new toys included in the Directive. For years I have clamoured that what we need is to support contracting authorities in doing the bread and butter well. That means doing the actual work of training, supporting, upskilling and rewarding those at the coal face. To date I have seen zero investment from the UK Government on this (publishing guidance doesn't count...)
But that makes for less compelling marketing case than the new toys. Plus, it costs lots of money. Up front and for uncertain or diffuse benefits (ie, improving tenders, reducing risk and avoiding the next Carillion). For now, it is important to cut those budgets a little more though.
In the meanwhile, lowest price contracts remain the rule and not the exception.
The European Commission has just published earlier today a lengthy guidance document on procurement of innovation. It is welcome since procurement of innovation is the kind of approach (policy? idea?) that is useful mostly for the top 1% of procurers and even then, just some of the time. As such, guidance is required to make it easier for it to be deployed.
Here's the blurb:
"The 2014 modernised public procurement directives 2 adjusted the public procurement framework to the needs of public buyers and economic operators arising from technological developments, economic trends and increased societal focus on sustainable public spending.
Public procurement rules are no longer only concerned with “how to buy” – they provide scope for incentives on “what to buy”, without prescribing them. The objective of spending taxpayers’ money well is gaining new dimensions, beyond merely satisfying the primary needs of public entities. With each public purchase, the public opinion is rightly interested to know whether the procured solution is not only formally compliant, but also whether it brings the best added value in terms of quality, cost-efficiency, environmental and social impact and whether it brings opportunities for the suppliers’ market."
The more we try to achieve with procurement the more complex it becomes. Remember when the next cycle of complaints about procurement complexity comes about.
1. UK East Coast rail franchise expected to be scrapped 'in days'. For the third time in 10 years one should add...
2. Carillion used its suppliers to "prop up falling business model". 120 payment days are not really new or unexpected in the field of public procurement (and others - just check Amazon's payment terms and how it survives on free cashflow...)
1. Microsoft deal helps NHS ramp up cyber security- oh, the irony.
3. When to challenge public contract awards - lessons learned from the UK passport saga - easy, when you have reasonable grounds for it.
4. Did Your Company Make that Warplane? Don’t Count on the Upgrade Work Anymore - on the importance of taking control of IP for maintenance/upgrades.
Myself and Albert have just made available on SSRN a new working paper on Brexit and public procurement, focused mostly on the transition agreement. As usual, we welcome comments and criticism.
Here's the abstract:
Can't really say I am surprised by De La Rue's decision, but the motives are a joy to read:
"We've done our homework, we've taken legal advice, we've looked at, frankly, the likelihood of overturning the decision and the sensible thing for us to do is to refocus our efforts elsewhere and to move forward."
"Following four weeks of intense consideration and clear legal advice, we have taken the decision not to challenge the award of the UK passport contract."
Seriously, was that no evident already two weeks ago? Or was it all an outburst to be seen doing something?
Bonus points for burying the news of the profit warning with the announcement of dropping the challenge. It will be interesting to see if the "profit warning should lead to lost contracts" brigade which sprung after Carillion's demise will be calling for their pound of flesh of De La Rue now.