A post by Jakub Klímek, Jindřich Mynarz, Lucie Sedmihradská, and Jaroslav Zbranek of the University of Economics, Prague

Tracing EU funds is difficult. The problems with following the money from the EU range from the low availability of budgetary data, which often needs to be obtained on a per request basis, to the ability to identify relevant monetary amounts, which leads to guessing related code list concepts. Due to the complexity of the flows of EU funds, tracing them is a manual and thus error-prone process that involves a lot of copying and pasting from heterogeneous spreadsheets and studying dull regulatory documents along the way. In this post we illustrate the challenges of tracing EU funds on an example of a concrete project.

Imagine you’re sitting on a bench in a park in Děčín, Czech Republic. Next to the bench there’s a plaque acknowledging that the bench was paid for using the European Regional Development Funds. In fact, as the plaque says, the whole park was revitalized using EU funding. Let’s say you have an inquisitive mind curious about public spending and so you want to find out how the revitalization of the park was funded.

Thanks to the plaque we know that the identifier of the park revitalization project is CZ.1.09/1.1.00/32.00709. However, a search for this ID on the Czech portal of EU funded projects yields empty results. Luckily, we know Google and upon searching for the project ID there we discover the project’s description on the site of the Regional Information Service of the Ministry of Regional Development of the Czech Republic. Searching in Google for the project ID restricted to the domain of the official portal of EU funded projects is unsuccessful, but eventually, using various hints, we work our way to the project’s description on the portal.

A bench from the park as shown in a promotional video for the project

The project started on November 1, 2010 and finished on July 31, 2013. The revitalization of the chateau’s park was needed because of its devastation caused by the floods in 2002 and prior to that by the long-term stay of Soviet, German, and Czechoslovak armies during the years from 1932 to 1991 (source). When the project was finished, the park was re-opened on July 5, 2013 (source).

Map of the park under the Děčín’s chateau (Map data ©2016 Google)

According to the project’s description it received 54 609 381 CZK from European Regional Development Fund (ERDF) and was subsidized exclusively from this fund. Even though the ceiling for co-financing from ERDF was set to 75 % (source), shortly after the project started its financing was changed on November 26, 2010, because the Czech government decided to cancel co-financing from the Czech budget for several development projects in the North-West region (source), so the project ended up being 100 % funded from ERDF and the budget of Děčín.

Now, how did this amount of money reach Děčín? As it turns out, the answer is not entirely straightforward. In order to arrive at an answer we need to address three fundamental questions:

1. Which institutions were involved in the flow of EU funds to this project?
2. When were the EU funds transferred?
3. How to identify relevant amounts in budgets of the involved institutions?

To answer the first question we can attempt to reconstruct the flow of EU funds to this project. We will start from the bottom and work our way up the administrative hierarchy to the European Commission. The project funding was received by the municipality of Děčín.

Děčín used the funding in several public contracts, the main one of which, originally estimated to cost 67 250 000 CZK (source), was awarded for 49 948 709 CZK (source) to an association of the companies Viamont DSP and ZEMPRA DC. Unfortunately, there is no reliable way to search for Czech public contracts that received EU funding. Even though the official contract forms feature the section VI with additional information, including the paragraph 2 about financing from the EU, it typically contains only the name of the project without its ID. Moreover, the Czech public procurement register doesn’t allow to specify search in this section. Instead, we found the public contracts related to the project by using full-text search with the project’s name, so it’s entirely possible that we missed some relevant contracts. Interestingly, the main public contract was administered by the ROPRO, a now-defunct company connected to several cases of EU funds fraud. Swietelsky, a competitor of the awarded companies, even protested against the decision to award this tender but the protest was rejected (source). In addition to the main contract we found 5 more contracts related to the project with total cost of 10 081 825.53 CZK (sources: 1, 2, 3, 4, 5).

Using the methodology for the EU funds from the 2007-2013 funding period (in Czech, p. 82) that was issued by the Ministry of Finance of the Czech Republic we find out that Děčín received the funding from the Regional Council of the Northwest Cohesion Region. Funding to the Regional Council was transferred from the Ministry of Regional Development that in turn received it from the Ministry of Finance, which finally got it from the European Commission (see p. 7 of the methodology). In this context the Ministry of Finance plays the role of certifying authority that receives the EU funds in EUR and transfers them further in CZK. The cost of exchange rate difference is covered by the certifying authority (see p. 71 of the methodology).

Now that we know the institutions involved in the flow of funding to this project, we need to discover when the funding was disbursed in order to be able to find it in budget data. EU funds may be paid in advance, via interim payments, or via refunds. The Czech Republic received most of the funds (52.9 %) via interim payments, a smaller fraction (9 %) was pre-financed, and the rest came from co-financing (source). The project we trace EU funds for lasted from 2010 to 2013. During these three years the project was funded via interim payments. Interim payments can be done ex-post (after the project realization) or ex-ante (prior to project realization) based on decisions of the Ministry of Regional Development and its associated Regional Council. A smaller part of the project was funding was refunded in 2014 after the project finished. That’s why we choose 2010-2014 as the relevant period for the flow of the project funding.

Having identified the institutions in the flow of funds to the project and delimited the fiscal periods during which the funds were transferred, we can attempt to identify the budget lines and payments relevant to the tracked project.

Data about beneficiaries of EU funds can tell us how money was spent on the project over time. Unfortunately, doing so requires trawling through many Excel files. First we learn that on March 19, 2011 there was 111 095 623.02 CZK allocated to the project (source), but this allocation was cancelled. After this false start, on April 20, 2011 a smaller amount of 56 747 767.1 CZK was allocated to the project out of which 17 205 502.46 CZK was paid on December 1, 2011 (source). However, it seems this report from November 2011 didn’t get the numbers quite right, because from the subsequent report from October 2012 we learn that the original allocation was in fact 56 745 819.92 CZK (source). A report from November 2013 after the project has finished tells us that the initial allocation was scaled down to 55 661 702.36 CZK and already 54 086 955.15 CZK was paid from it (source). Finally, two years after the project’s end a report from April 2015 quotes 55 635 894.22 CZK as the allocation, out of which 54 609 381.01 CZK was paid up to September 19, 2014 (source).

In the Czech Republic, since 2007 the transfers from the EU are reported by public bodies to the Ministry of Finance using the statement XI. The expenses co-financed from this revenue are reported using the statement XII. These are the official forms that the public bodies must report their expenditure and revenue in. These types of statements aren’t publicly disclosed on the Monitor of the State Treasury that provides budget data for the Czech public bodies, nor they are disclosed as part of the budget data required by the mandatory disclosures arising from the freedom of information law, which demands only aggregated budget data to be published pro-actively. However, few public bodies publish this type of statement voluntarily on their web sites. Děčín is one of such public bodies so we have both statements for all the fiscal years during which the observed project was realized. However, only the statements for the years 2012-2014 were pro-actively published by Děčín. We had to request the statements for the earlier years of 2010 and 2011.

A key instrument for locating relevant monetary amounts in these statements are the code lists standardized for the Czech budgetary bodies. However, in order to pick code list concepts relevant to the tracked project we have to use a bit of guesswork, because the assignment concepts from code lists mostly follows convention rather than rules. In some cases, funds come with explicit instructions on how to classify them, such as when a regional council transfers money to a municipality, but even these rules are not always followed. Additionally, the diverse code lists used to locate the relevant amounts manifest diverse aggregations applied to the funding on its way to Děčín.

We start in the statement XII that describes Děčín’s expenses co-financed from EU funds. We select the “Preservation and restoration of cultural heritage” as the relevant paragraph, “Regional operational programmes” as the instrument, “Buildings, halls, and construction” as the budget item, and both “Revenue and expenditure from domestic sources” and “Revenue and expenditure from foreign sources” as the spatial dimension due to co-financing from the budget of Děčín. Regarding the spatial dimension, the budget data published by Děčín mislabels it as “Source”, which corresponds to another code list. Having located the relevant budget line we select the outturn from the beginning of the year as the actual expenditure. The filtered monetary amounts can be found in the table below.

Moreover, thanks to the freedom of information requests by the Czech Pirate party, the list of Děčín’s invoices from 2010 to 2015 is available. We can narrow it down to the invoices to the company that was awarded with the project tender, filter the relevant invoices based on their textual descriptions, and sum up the invoiced costs per year for the project, so that we can see the actual payments during the project’s realization. Interestingly, there are mismatches between the invoice payment dates and the dates the project payments were reported to be spent. For instance, as can be found in the table, in 2012 Děčín paid 23 million CZK to the company realizing the project, but no project spending was claimed that year.

To locate the relevant revenue in Děčín’s budget we select “Revenue” as the paragraph, “Regional operational programmes” as the instrument, both “Investment transfers from regional councils” and “Non-investment transfers from regional councils” as the budget items, and “Revenue and expenditure from foreign sources” as the spatial dimension. We need to include both investment and non-investment transfers, because we can only assume that the project was funded from investment transfers. We also apply this for the upstream funding. Foreign spatial dimension is chosen because domestic co-financing comes from the budget of Děčín. Note that these amounts include all funding from regional councils, not restricted to cultural heritage. A potentially misclassified revenue may be seen for the year, when Děčín declared to have received 50 million CZK from the Regional Council of the Northwest Cohesion Region, but this year the council did not send any funding to Děčín to spend on “Preservation and restoration of cultural heritage”.

Moving one step up the funding stream we get to the Regional Council of the Northwest Cohesion Region. The council doesn’t publish the statements XI and XII, so we had to obtain them via freedom of information requests. In these statements we can find the relevant monetary amounts if we restrict the paragraph code list to “Preservation and restoration of cultural heritage”, choose “Regional operational programmes” as the instrument, include both “Investment transfers to municipalities” and “Non-investment transfers to municipalities” as the budget items, and select “Revenue and expenditure from foreign sources” as the spatial dimension. The selected amounts are aggregate the transfers to all municipalities that the council funds.

The revenue of the council is aggregated in a yet another way. In order to find it, we pick “Revenue” as the paragraph, “Regional operational programmes” as the instrument, both “Other investment transfers from the state budget” and “Other non-investment transfers from the state budget” as the budget items, and “Revenue and expenditure from foreign sources” as the spatial dimension. At this stage we lose further level of resolution, because the selected amounts are aggregated for all recipients of regional operational programmes.

The next organization involved in the traced monetary flow is the Ministry of the Regional Development. Its budgetary data is available from the Monitor of the State Treasury. We filter the ministry’s expenditure to the budget items “Investment transfers to regional councils” and “Non-investment transfers to regional councils”. In this step we lose the distinctions of the instrument and the receiving regional council, since we can only get the amounts aggregated for all instruments and councils. The Ministry’s revenue is aggregated even more as it includes both the budget items “Investment transfers from the National fund” and “Non-investment transfers from the National fund”.

The matching expenditure can be found for the Ministry of Finance, for which we select the same budget items including “Investment transfers from the National fund” and “Non-investment transfers from the National fund”. The National fund is a department of the ministry that administers EU funds. Unfortunately, the data on revenue of the Ministry from EU funds is unavailable. Although the Ministry shares detailed data on transactions from EU funds with the Czech Statistical Office, the data is shared exclusively with the Statistical Office and cannot be redistributed to third parties.

Finally, we can reach the root of the traced monetary flow, the European Union. While budgets of the EU are available as structured data in XML and CSV from 2013 on, the previous budgets are available only in PDF. For example, the 2012 budget is a 1906 pages long PDF. In order to find the actual payments we need to get outturns for the years during which the tracked project was realized. The outturns are published with two-year delay, so we need EU budgets for 2012-2016. The operational programme of the tracked project pursues the convergence objective (source), which maps to the concept “European Regional Development Fund (ERDF) – Convergence” in the Activity-Based Budgeting Nomenclature that structures the budget of the European Union.

The EU budget provides ERDF funding aggregated for all EU member states. More detailed breakdown can be found in the data for research published by the DG Regional Policy. There we can find the ERDF funds transferred to each member state separately. Moreover, the data available from this page also contains amounts of funds transferred to the Ústecký region, where the tracked project was realized, following the Convergence objective of the ERDF, and classified as “Protection and preservation of the cultural heritage”. Nevertheless, the data provides only cumulative figures for the period from 2000 to 2013. During this time span 6.97 million EUR of relevant funding was transferred. Somewhere in this bulk of money there lies the spending on the project tracked by us.

Conclusions

Tracing EU funds is difficult. Even though this effort did not achieve clear results, it took weeks worth of domain experts’ manual work. In this post we wanted to illustrate the difficulty of tracing EU funds on a concrete example.

The initial problem we encountered was the low availability of relevant data. Data on monetary transfers is rarely disclosed pro-actively on the web. Often you need to request it from the responsible institutions, either informally or via a freedom of information request. This introduces friction and delays into the process of tracing EU funds.

When the data is available, it is often difficult to find. Rather than following explicit links we were following breadcrumbs of hints we discovered on our way to the relevant data sources. Navigation in budgetary data is far from the information superhighway we’ve been promised; it is more like a labyrinthine path in a dark forest. There is a simple solution to improve navigation in the data: use more links.

When we discovered relevant data, it was rarely simple to use. Frequently we had to trawl through heaps of spreadsheets to locate the data of interest. Working with heterogeneous unstructured data required manual effort, including copying and pasting diversely formatted numbers, and was thus error-prone. In fact, we would not be surprised if we made an error.

Perhaps the least reliable part of our effort was the code list guesswork we had to do when locating relevant monetary amounts. In many cases we needed to guess code list concepts of budget lines related to the tracked project, since there are not explicitly stated for the project. Even though we are confident with our guesses, misclassification may still have happened. Moreover, we learnt that some code lists used in the Czech Republic are actually not code lists, but instead they are derived from data, so they are not fixed enumerations.

When we managed to identify the relevant monetary figures, they often had a low level of detail. Monetary amounts are disclosed mostly in aggregated forms, in which individual transactions get lost. Flows of EU money are woven together in the aggregations, so that it is not possible to separate individual payments related to specific projects. Similar problems related to aggregated spending including co-financing were previously reported by Journalism++ in their account of common roadblocks encountered by journalists when working with budget data (see p. 13).

Ultimately, this exercise left us wondering whether the complexity of the flow of EU funds is essential or merely incidental. Incidental complexity may be a manifestation of excessive bureaucracy that opens many opportunities for mismanagement of the EU funds and prevents the mismanagement from being noticed. Better transparency of data about flows of EU funds would help us assess this conundrum more thoroughly.