Why Is So Much ICT Spend Missing From Contracts Finder?

hot topic jpeg

Kelvin Prescott Jan 15And so we are back with our Hot Topics after a short lull. To kick off April's topic - Procurement Technology - Kelvin Prescott, Information Director at Innopsis, the public sector tech trade association that is championing better procurement practices between the public sector and its suppliers, brings us this thought-provoking discussion on why ICT spending is missing from Contracts Finder.

As a result of the updated procurement regulations in the UK, all public bodies are required to publish their opportunities with a value greater than £10K on Contracts Finder. In theory, this should make it much easier for suppliers to identify potential business in the public sector, particularly for contracts that fall below the current EU thresholds.

However, research conducted by Newbury Management Consultants, on behalf of Innopsis, found that the majority of ICT spend was missing from Contracts Finder. Total government spend on ICT is estimated to be about  £17bn per annum, but the annualised value of ICT contracts advertised or awarded in the 2015 to 2016 financial year was only £5.98bn. The majority of government spend, approximately two-thirds, does not appear to be advertised on Contracts Finder at all.

We analysed data published on Contracts Finder for the period April 2015 to December 2016 to identify how much ICT spend was advertised. The graph below summarises the results:

Innopsis article 2

In total, Contracts Finder has seen £4.5bn of ICT contracts advertised in the period April to December 2015, across 1,635 opportunities, with an average value of £2.7m. Annualised, that implies a total spend of about £5.98bn on ICT.

Where is the missing spend? One explanation may be that the spend is hidden in one of a number of framework agreements.

Framework Spend Exemption

The government’s policy has an important exemption: individual contracts that are awarded under a framework agreement don’t have to be published. This has the effect of drawing a veil over the whole procurement process, at least for any companies that aren’t on the framework. A public body can choose to issue tenders only to suppliers on a framework, thus limiting the number of potential bids they have to evaluate.

This exemption was highlighted in an article by Peter Smith of Spend Matters here. His concern was that public bodies would tend to increase their use of frameworks in order to avoid the cost and compliance overhead associated with publishing lower-value contract opportunities.

This can create problems within the marketplace, as there are multiple frameworks within the public sector covering ICT. Suppliers cannot anticipate in advance which framework a customer will use, and so are forced to bid on all of them, increasing their bid costs and prices.

There are other explanations for the missing spend information. Some public bodies may not be following the rules. There are likely to be variations from year to year in the total value of contracts that are awarded. Finally, some ICT spend may be bundled into contracts for other types of goods and service. However, it seems likely that a significant part of the missing spend is flowing through frameworks.

Innopsis believes that spend under framework agreements should be subject to the same rules on transparency that apply to other types of contract. In order to maintain a vibrant marketplace, potential suppliers must have visibility of the opportunity for them in the public sector.

Share on Procurious

Voices (13)

  1. life:

    All of the above and some. I’ve previously undertaken a reconciliation of actual individual company revenue (of which I had an accurate and definitive view!) against spend as published with that company and it was completely wrong and, overall, under reported.

    It’s not just innacurate it’s just wrong and not useful.

    Other factors that (I am only guessing) might make for material differences is the reissue of large contracts through different portals/frameworks etc. over time, large extensions of work rather than “proper” re-procurements at contract expiry (of which there are many!), shifts of contracts and hence spend into, out of, and then back into, arms length companies (ALMOs etc.) particularly in Local Government, changing ratio of Local Government/Health etc. within the totals due to sporadic drives from CCS and predecessors for sectors other than central Gov. to use frameworks (and thus “hide” spend as explained above), etc. etc. etc.

    The accountant in me also ponders “You cannot control what you do not measure” vs “Gathering information with which you will not make a decision is wasted effort”. Obviously it would in a generic sense be good to have this info, and comorting to know we could (although to do it properly would be extremely costly) but actually over and above the estimates we can already get in different ways what would we actually do with an exaustive record? I’m not even sure it’s that useful for suppliers in reality.

  2. Agrajag:

    They haven’t taken their data over a period of time like that. They have selected 9 months worth of data and then extrapolated to make a whole year. Why not take a full 12 month’s worth of Contracts Finder data? The cynic in me might wonder if the other three months didn’t help them to make their point.
    But that isn’t really the point. You cannot compare two sets of data that report different information. In this case historic expenditure is being compared to estimated future contract value. I would also challenge their assertion that £17bn p.a. is spent by government on ICT. I suspect that they mean public sector expenditure, and not all public sector bodies are required to publish contract opportunities on Contracts Finder. Incidentally, and probably coincidentally, central government ICT expenditure is estimated at about £7bn p.a.

    1. Kelvin Prescott:

      Hi Agrajag

      Good questions. To answer:
      1. We took 9 months data because contract information prior to March 2015 was archived by Contracts Finder (and is in a different format to the data now available). Extending the analysis to prior years is on the to-do list, albeit that both the format of the data, and the amount of validation needed, means that it isn’t a straightforward task to undertake.
      2. The other reason for the time period chosen was that my analysis was originally stimulated by Peter’s post – asking whether or not the change to the procurement regulations in Jan 2015 was driving a shift to the use of frameworks. I didn’t find any evidence of this from the data on Contracts Finder (you might have expected a reduction in the number of contracts, or potentially an increase in their average value). However, it did pop up this oddity – which was the very large gap between advertised spend information and total contract value.
      3. The £17Bn figure is taken from a couple of sources: externally this is the value tracked by industry monitors like Kable and Ovum. This spend figure covers spend by Schedule 1 entities under the Public Procurement Regulations (so, including central, local, devolved and NHS expenditure, but excluding utilities and private companies under government control like the banks). The Cabinet Office’s have cited similar figures on presentations given by both CCS and GDS, but have not formally (to my knowledge) published a breakdown. So, there is no “correct” figure published for aggregate spend – perhaps the topic for a different debate!

  3. Agrajag:

    Your argument, Peter, only works if the contracts are awarded evenly over a period of, say, three years. If, for example, MOD, DWP, NHS and MOJ all award new contracts in the same year, it is likely to skew the distribution. So I stand by my comment that it conflates two separate sets of MI. Contracts Finder is a website for publishing opportunities and newly awarded contracts (albeit with some exceptions). It isn’t intended to be a source of expenditure MI.
    Also it is possible to to have ICT expenditure on a published contract that isn’t categorised as ICT. I would suggest that a third set of data that may help to corroborate or disprove the article would be to analyse the published financial transaction data that each Government Department is required to publish. Every transaction over £25k is published.

    1. Mark Lainchbury:

      Yep, Agrajag has it !!

      Soon as I read this first paragraph, I thought “One of these things is not like the other”

    2. Peter Smith:

      But over time why wouldn’t contracts across government tend to a regular distribution year by year? No reason why not! I know that only works in aggregate but this is across many organizations.

      1. Kelvin Prescott:

        Thanks all for the comments. A quick round up of the methodology here, as there were some good points and questions raised/

        1. Why was only 9 months data used? The reason is that the data for months prior to Feb 2015 was not available in the same format, or the same level of detail, as is available from Contracts Finder today. Doing the same analysis over a longer time period is on the to-do list, but both the quality of the data and the need to adjust formats meant that it was outside the scope of the initial analysis
        2. Is this not a comparison of apples and oranges (i.e. we were comparing total contract values against actual spend estimates)? Yes and no – the problem we were trying to highlight is the very difficulty of getting accurate estimates of either set of numbers.
        3. A couple of readers commented that TCV shouldn’t be expected to match actual spend in year because of the difference in contract terms on advertised opportunities. I would agree that in any one year they are likely to be different, if only because the award of major contracts may be cyclical (e.g. aligned to general elections) leading to some years TCV being higher than actual spend, and some being lower. However, on average they should be the same. The question then is: would you expect a cyclical difference like this to result in a deviation of 60% in any one year? That’s the difference between the £17Bn per year or so of spend, vs. the £6Bn or so of contracts that were advertised. I think that cyclical variations will explain some of the delta, but it seems unlikely to be the main cause (though I will acknowledge, short proper data that is just an opinion).

    3. Kelvin Prescott:

      Hi Agrajag

      you’re absolutely right that, in any one year, total spend will be different to the aggregate value of contracts awarded. However, over time they should be the same (Peter’s point – in year 1 you spend £100, but award contracts for 3 years worth £300, then year 2 and 3 will be zero, year 4 you will re-award again. the average will be £100/year of spend, and £100/year of contract value).

      The question is whether the difference in value can be explained by cyclical factors (e.g. you might expect large contracts to be aligned to general elections, so that more are awarded in the first year of each administration, and fewer in the last)? The difference between £17Bn of spend and £6Bn of advertised contracts is about 60%. I would suggest (though I accept I cannot prove) that this variance is too large to be explained by cyclical factors over as large a population and number of contracts as this.

      I agree that it would be instructive to compare published transaction data vs. published TCV – one for future research I think (the methodological challenges are pretty big…)

  4. Agrajag:

    This conflates two different pieces of MI, the amount spent and the number of contracts. Most organisations will have medium or long term contracts for the provision of their ICT requirements. Therefore most will not need to advertise or publish ICT contracts every year. If most contracts are an average of three long, then your figure of £5.98bn looks about right, assuming about a third come up for renewal each year.

    1. Peter Smith:

      That was my first thought – but on reflection, when you advertise a three-year contract, its quoted value would i think also be 3 years worth of spend if you see what I mean! So I don’t think the argument holds – you should expect to see the higher value every year. So imagine you have total spend of £15 million a year, made up of three contracts, each three years and £5 million a year each. In year one, you advertise contract one – at £15 million. In year two, contract 2, also 15 million … and so on. Each year is £15 million.

  5. Dan:

    Its not just through framework agreements. The obligation to publish on Contracts Finder only applies if you’re advertising it elsewhere (e.g. on your website). If you’re not advertising it and are just picking a handful of companies to invite to tender (whether through framework agreements or just your usual favourites), then Contracts Finder is not involved.

    If I had to guess, I’d say that this attempt to increase transparency has actually had the opposite effect by discouraging public authorities from advertising their tender opportunities.

    1. Kev:

      An accurate insight from Dan. There’s no requirement to publish on CF if you are legitimately ( i.e. via your own CPR) restricting competition to those suppliers you believe can offer you value.
      Not unusual for national level policies have the opposite effect to whatever was intended.

    2. Kelvin Prescott:

      Dan, you are quite right, there are a number of exemptions to the requirement to publish opportunities on Contracts Finder.

      What I have seen (and you may have seen as well) is a number of anecdotal instances where buyers choose to route lower value work either through existing frameworks, or via existing large prime contractors.

      Its quite possible that the effect of the new policy is to discourage authorities from advertising (that was the gist of Peter’s original advert). However, the evidence so far is inconclusive. There isn’t a discernible trend (either up or down) in the number of opportunities published on Contracts Finder since the new policy came in.

      With that said, the data on Contracts Finder has numerous issues of quality/consistency that make it difficult to spot a trend even if there is one!

Discuss this:

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.