Monday, 29 October 2018

Calculated overheads rise from the dead

In FP7, the EC encouraged all beneficiaries to use calculated overheads, in particular by scheduling the removal of a 60% flat rate for overheads half way through the programme (with 75% funding of costs). However, they took a decision to keep the 60% for all of FP7 and then provided only a 25% flat rate – no calculated overheads - in H2020 (with 100% funding). The reason for this change of direction is that, when audited, they found many errors in the beneficiaries’ overhead calculations. Early results in H2020 audits suggest the removal of calculated overheads has led to 28% reduction in error rates compared with FP7.

However, the 25% flat rate penalised those using expensive equipment. The EC’s response, called Large Research Infrastructures (LRI), required certification by the EC and was limited to organisations whose “Infrastructures accounted for at least 75% of the total fixed assets”, so excluding most H2020 participants. Uptake of the LRI certificate has been even lower than that of CoMUC/CoMAv. Eventually, the EC changed its mind and added a new cost category called internal invoices, which allowed some overhead costs to be re-classified as direct costs, and no certificate required. 

For Horizon Europe, the EC proposes to add further overheads in costing these internal invoices, which the EC’s supporting documents suggest will involve “fair keys or drivers” to distribute the indirect costs. Is this the EC’s first step in defining standards for overhead cost accounting?

Wednesday, 3 October 2018

EU research grant pays mortgage on castle

In its report for 2017, the EC’s Anti-Fraud Office (OLAF) describes a fraud in which more than €1.4 million of EU research funding was misappropriated. An Italian led consortium with partners in France, Romania and the UK obtained funding for a project to build prototype hovercraft for use as emergency vehicles able to reach remote areas in case of environmental accidents. On-the-spot checks in Italy by OLAF and the Italian Guardia di Finanza discovered various disassembled components of one hovercraft, as well as another hovercraft which was completed after the end of the project.

They also discovered that the British partner only existed on paper: the company was in fact created and owned by the Italian partner. To simulate the development of the project and to divert funds, fictitious costs had been recorded. Analysis of more than 12,000 financial transactions showed that part of the EU funds received by the Italian and UK partners had been used to pay off a mortgage on a castle facing foreclosure.

But fraud in EU research funding is rare. According to OLAF, in 2017 only eight cases of fraud related to payments totalling €520k were detected in the Research and Innovation policy area. This amounts to about 0.01% of payments made that year, similar to the percentage for the period 2013-17. So not much money for castles!

Thursday, 6 September 2018

No Brexit changes but…

While there has been little obvious progress on the main issues which separate the UK and the rest of the EU – trade and movement of people – both sides have published plans for the possibility that negotiations fail. Here’s the UK update for H2020.

In 2016, UK government guaranteed funding for UK organisations included in proposals submitted to H2020 before the UK departed from the EU if H2020 selected the proposal for funding.

Recently, this guarantee has been extended to include successful proposals submitted up to the end of the year 2020 for cases where UK is able to participate as a third country. This means proposals where a consortium is required to carry out the work: most of Societal Challenges and Enabling & Industrial Technologies, and parts of Excellent Science.

UK government is now establishing a register of UK organisations which now or in the next few months will be entitled to claim against this guarantee. If negotiations fail, organisations applying to H2020 after the March 2019 deadline will use the same register so they can claim UK funding for their participation.

Thursday, 26 July 2018

Speedier payments?

In FP7, many were surprised how long it took the EC to make payments against interim and final reports. Often, they were not paid within 90 days of submitting their reports as required by the grant agreement, while the EC maintained that the contractual condition was fulfilled and, indeed, their management target of paying within 60 days was achieved.

The difference of interpretation arises because the EC can ask questions about the reports, so suspending the time limit for payment. The table provides data for two of the EC organisations administering FP7 and H2020, showing the average “net” time to pay after the suspensions or delays are deducted from the gross time to pay which grant recipients experience.

Interestingly, the first substantial data on payments in H2020 shows a large reduction in delays. Is this because the EC is asking fewer questions? Or is it because most of the H2020 data is likely to be for interim payments, which usually suffer shorter delays than final payments?

DG/ Agency
Payments (number)
Average time to pay (days)
RTD = Directorate General for Research and Innovation; REA = Research Executive Agency

Monday, 2 July 2018

H2020 audit results

A major EC motive for using lump sums is that they will normally produce zero cost errors. FP7 audit results up to the end of 2017 showed that errors detected in a representative sample were 5.03% which - after errors are corrected - gives a residual error of about 3%. This should be compared with the 2% target set by the EU. So the EC will again fail its annual audit by the Court of Auditors. 

So far
But the first results for H2020 auditing suggest that the error rate will be below the 2% (see table). The EC audited a statistically representative sample of 142 participations. For the 110 cases finalised, the detected error rate was 1.6%: the corresponding residual error was 1.44%.

However, not all audits in the representative sample are included in these figures, because some are not yet accepted by the organisations audited and so are not finalised. If the EC’s opinion in the draft reports on these outstanding audits is included they produce the result in the pessimistic column – still much lower than FP7.

Interestingly, 28% of the error found in FP7 related to overheads, which has reduced to zero in H2020 by replacing calculated overheads and a choice of two flat rates with the single 25% flat rate. By itself, this would reduce the 3% to 2.16%, close to the EC pessimistic forecast for H2020. So an initial conclusion is that other simplifications in H2020 had either little impact on reducing errors, or even increased them!

Kind regards

Friday, 1 June 2018

Complex but clear

When the first H2020 projects began, EC auditors advised that the cost eligibility rules for third parties were “complex but clear” and should be read carefully. They covered subcontracting (obviously), contracting (what’s the difference?), contributions-in-kind made free of charge (if its free, how can it be an eligible cost?), contributions-in-kind made against payment (isn’t that subcontracting?) and linked third parties (don’t ask). Then there is the possibility to claim the cost of natural persons who are not employees but under direct contract (similar but not identical to FP7’s in-house consultants). Altogether, there is a bewildering array of options and corresponding rules facing an organisation which wants to use the skills of a person not their employee and claim the related cost.

Now add the SME “instrument”: funding up to €2.5 million specifically for individual small companies with large ambitions. This relaxes the usual H2020 rule that subcontracting may cover only a limited part of the project. Mixing small companies in a hurry with complex cost eligibility rules for potentially large outsourcing contracts looks like a recipe for disaster when auditors come to check. But no! The EC anticipated the problem and – exceptionally – allowed applicants for this funding to justify their choice of subcontractors in their project proposal. If the proposal evaluators accepted their justification, then auditors could not question their choices. So all will be OK.

Well, not quite. Some SMEs interpreted acceptance of their outsourcing choices as licence to subcontract the tasks identified, and changed their minds about which supplier they would use. So now the EC agency administering the SME instrument is checking carefully subcontract costs claimed in the progress reports submitted by the SMEs and rejecting those which were not awarded to the supplier identified in the proposal. This could, of course, spell financial disaster for SMEs with large value subcontracts in their projects.

Fortunately, quite a number of them have set up companies specifically for this H2020 funding, separate from their existing business and assets, thus insulating them from death by audit.

Kind regards
Singleimage Limited

Friday, 27 April 2018

#H2020 Simplification fails

In preparing for H2020, the EC proposed measures to simplify claiming costs which they felt would reduce errors found when auditing. One addressed the topic of productive hours, where they proposed that the grant agreement should – unlike its predecessors - contain the number of annual productive hours to be used for the calculation of hourly personnel rates. As a result, the H2020 grant agreement includes three detailed options for calculating productive hours.

At a number of H2020 coordinators’ days held by the EC, their audit staff presented their forecast of errors which they expected to find in financial statements once H2020 auditing began. These were (i) direct costs apportioned, not measured (ii) timesheets missing or unreliable (iii) best value for money in subcontracting and purchasing not achieved (iv) confusion between basic and additional remuneration and (v) cost of in-house consultants not justified. Errors in productive hours did not feature.

In April, EC audit staff presented their first analysis of the H2020 audits (see table). The unit of measurement is the number of audits which found the error type listed. Many of the forecast errors can be seen there. “Direct costs apportioned, not measured” are in both the equipment category and other goods and services, the latter restated as “indirect costs claimed as direct”. Timesheets are in “incorrect time” and contribute to “other” in personnel costs. “Best value for money” appears directly and also as “inadequate supporting documents” and additional remuneration occurs under personnel costs. Consultants don’t feature – maybe they are hiding in “incorrect remuneration”, though this might also cover mistakes in personnel costings using the last closed financial year method– an H2020 simplification not forecast by the EC auditors to produce problems. But the biggest source of errors is “incorrect productive hours calculation”, which is also not in the EC auditors’ forecast top five, with a prize-winning 31% of the total!

Kind regards