IMPROVING THE GOVERNANCE OF STATE-OWNED ENTERPRISES IN THE ENERGY SECTOR
A politically connected company is defined as a private company or state- owned enterprise (SOE), whose decisions could be influenced by politically exposed persons (PEPs). For the purposes of the current research, the main indicator used to identify politically exposed companies will be the participation of a PEP or their immediate relatives in the company or state-owned enterprise (as noted in asset declarations or other public registers). “Immediate relatives” are spouses, parents, siblings, and children (excluding cousins, aunts, etc.). A company is considered “politically connected” if the PEP is participating in the management board/board of directors of a private company or SOE and/or has an ownership of over 50% in the company. We use this definition because PEPs – and their immediate relatives – usually are obliged to report this information in their annual asset declarations which is the primer source of the data.127
The methodology used to measure procurement integrity is developed by the Government Transparency Institute (GTI). Several indicators are tested and validated using rigorous statistical methods to create the composite Integrity Score that measures the overall integrity of each procurement contract – for a detailed description of the methodology and steps of indicator calculation see Fazekas – Kocsis (2020)128 or read the Training Manual written by GTI as part of the R2G4P project.129 The indicators used in this report to create the Integrity Score are reported in Table 9. Each indicator can take the value of 0, 1 or in some cases 0.5, where 0 indicates high corruption risk. The Integrity Score is the simple arithmetic average of the seven indicators. It is important to note that each indicator is calculated and validated on the national datasets, hence each is robust on the national level.
Table 9. Integrity indicator description
Where Y represents an integrity indicator or the composite Integrity Score, is a binary variable indicating whether the supplier of the procurement is politically connected, is a list of control variables including the location of the buyer, the type of the buyer, the location of the supplier, the type of the contract (supplies, works or services), contract value in deciles and market and year fixed effects.
Contract volume share
The expectation behind politically connected firms participating in public procurement procedures implies that political connections can be used for corrupt exchange and favouring certain suppliers based on their relations to PEPs. Yet before analysing whether political connections significantly influence the outcome of public contracting, it is important to check the overall state of market penetration by PCs. The share of political connections is not likely to be very high, albeit the small group of connected firms can introduce a lot of corruption risks. At the same time, PCs in public procurement cannot be considered as signals of corruption per se, only if correlated with lower competitiveness or other corruption risks.
Regional trend shows that the number of contract volume share of politically connected firms varies between 2% (Croatia in 2017) and 16% (Serbia in 2021) with an average share of 5%. Yet in some cases when calculated within the same CPV division the share can significantly differ.130 For instance, in North Macedonia the share of politically connected firms within the same market is usually lower than the overall annual number (e.g. in 2016 the overall share is a little above 2% while within the same CPVs it is 1%). On the contrary, in Bulgaria the distribution looks different: in most of the cases the average volume share within the same economic market is 1-2% higher than the aggregated number per all sectors (Figure 5).
The difference in annual share versus the share of politically connected firms per the same market can reveal whether there is an existing systemic dependency of the market on political connections, as well as potential presence of monopoly or oligopoly in the respective sector. Croatia and Bulgaria appeared to be the countries where the penetration of politically connected companies per sector is on average higher than the annual share, therefore some sectors are more likely to be dominated by firms with political connections.
The Bulgarian case is a good example of how the distribution of contract volume share can be different in aggregated annual numbers and average numbers with the same CPV division (within the same market). While the overall number of contracts distributed among politically connected firms is getting slightly bigger every year, the share within the same CPV also increases but a bit more (Figure 5). For instance, the difference in average share between 2017 and 2019 is less than 0,5%, while increase within the respective CPV division raised from 5,5% to approximately 7,2%. At the same time, there was a reverse dynamic after 2019 – while in 2021 the average annual share increased from 3,5% to almost 7%, the average volume share within the same CPV decreased from 7% to 5%. Therefore, while in general political connections seem to be more prevailing in overall procurement procedures, they are becoming more scattered across different sectors.
Figure 5. Contract volume share of politically connected companies in Bulgaria
Source: GTI and CSD
Contract value share
While the share of contracts won by politically connected firms shows the general picture of their penetration in procurement procedures as well as the distribution across sectors, the contact value share is a more important indicator to assess how much money was actually distributed among such firms and therefore can potentially be prone to corruption risks. It can be the case that while the number of contracts is not high, their value itself is pretty big, meaning that politically connected firms are aimed at large contracts (for instance in the construction sector).131
Similar to contract volume share, one can observe different trends when it comes to aggregated annual numbers and average share within the respective CPV division. Higher numbers in the latter can be a signal of a more intense penetration of politically connected companies within certain economic sectors, and while on average it is not high, some sectors can be dominated more by politically connected firms.
Most of the countries in the region showed a higher contract value share than volume share of politically connected firms. While in some cases the difference is moderate (0,5%-1% difference in Croatia or 2%-3% difference in Bulgaria), other countries show a more drastic distinction. For instance, in Hungary contract value share in 2015-2021 is between 2,5%-14% (EUR 200- 800 million), while the contract volume share of politically connected firms is around 2%-4% which means such firms are mostly aimed at the bigger value contracts.
At the same time there is an observable difference between the average value share of connected firms within their own CPV division and the aggregated annual share. While in some countries contract value share within the same market is significantly lower than the aggregated share (e.g. in North Macedonia the difference builds around 3% on average), in majority of cases the average within market share is higher. This trend confirms the previously observed pattern with contract volume share: politically connect- ed firms are quite frequently concentrated within certain markets rather than scattered across different sectors.
In the case of Croatia, the value of contracts won by politically connected firms steadily increases over the years from 1,5% in 2015 to a little over 5% in 2021 (Figure 6). The 2021 increase is in line with the change in the contract volume share of politically connected companies reaching a bit more than 8% in the respective year. Yet the value share is slightly higher almost every year (around 0,5%-1,5% difference), except for the significant increase in 2016 (the average contract value share of connected firms’ contracts was more than twice higher within the respective CPV division).
Figure 6. Contract value share of politically connected companies in Croatia
Source: GTI and IPF.
Buyer type value share
The number of politically exposed companies’ contracts is expected to be different for different types of buyers depending on the country-specific context, including type of electoral system, size of the country, administrative structure, and the amount of resources available to each level. In larger states with diverse administrative structures and resourceful local-level bureaucracy, one can expect a higher level of political connections being used within regional or municipal authorities due to the higher corruption risks.132 At the same time, a lack of resources of high level of centralization can lead to either a lower level of corruption in municipalities or take a different form.133 For instance, a large number of small contracts are prone to corruption risks at municipal or regional level, but in sum, several of large value contracts on the national level constitute a higher share.
The regional trend is diverse and not straightforward, but in a majority of cases, there is an observable difference between buyer type of politically connected companies’ contracts versus other companies’ contracts. For example, in Romania there are significantly more National Authorities among politically connected companies’ contracts than the others (40% versus 20% among non-connected companies). At the same time, while in Hungary National Authorities are also representing a larger share among connected firms, the prevailing category is Regional Authority (40% versus around 30% among non-connected firms). On the other hand, in the case of Croatia there is no significant difference between the two groups. Some countries show slight differences – for instance, in Bulgaria a slightly more prevailing group among connected firms is non-defined (around 5% difference).
In the case of North Macedonia (Figure 7) among connected companies’ contracts there is a significantly higher number of National Authority buyers in comparison to non-connected firms (around 30% among connected firms and 12-13% among non-connected). At the same time, regional authorities construct around 10% of buyers with connected firms’ contracts and almost 20% with non-connected. While other public bodies prevailing in both groups, non-connected firms’ contracts are more likely to be procured by these buyers than connected ones.
Figure 7. Buyer type value share of politically connected companies’ contracts vs. other companies’ contracts (2015-2021) in North Macedonia
Source: GTI & IDSCS.
In some cases, connections within local authorities can result in the extraction of resources by connected persons from the national level. For example, in Bulgarian city of Kotel, farmers and companies connected to the town’s mayor, Kosta Karanashev, and the now-former Minister of Agriculture, Dessislava Taneva, (both being members of the political party GERB), have been fraudulently extracting agricultural subsidies and procurement funds since as early as 2013.134 The family received funding for lands which were not in active cultivation, constituting fraud.135
Procedure type value share
It is expected that politically connected companies’ contracts will go through less competitive procedures prone to higher corruption risks. For instance, such procedures as negotiated, especially without publication, or restricted, as well as direct award, are less transparent and therefore can be used by the interested party for corrupt exchanges136. On the contrary, open procedures are more difficult to use for corruption because of the more transparent rules and publication requirements. At the same time, while procedure type can serve as a proxy for corruption risks, not all types of restricted procedures necessarily result in it, as interested parties can use other instruments for corrupt exchange even within open procedures: e.g. tailoring tender requirements, deliberately shortening the time for submission, taking longer time for decision period to negotiate terms with supplier.
In the majority of the countries in the region, the share of open procedures among connected firms’ contracts is either not significantly different or trends counter to the expectations (i.e., there are slightly more open procedures among connected firms). For instance, in the case of Bulgaria, there is no substantial difference besides a couple of percentages in negotiated procedures without publication (around 5% among connected firms versus 3% for non-connected firms). Yet a couple of countries showed some expected distributions, for instance in Romania the share of negotiated procedures without prior publication is almost twice higher for connected firms (25% versus 12%). At the same time, Hungary and Serbia showed counterintuitive results with open procedures being more frequent for connected firms’ contracts. In the Hungarian case there is also a difference in negotiated procedure with publication showing a higher share for connected firms (17% versus 8% among non-connected firms). The relatively high share of open procedures among politically connected firms in a majority of the countries might mean that in most of the cases other practices are used for corruption (e.g., tailored tender specifications, short advertisement period, etc).
The case of Bulgaria is more in line with the expected distribution of procedure types in the two groups. Figure 8 shows that both for politically connected and non-connected companies’ majority of the contracts had open procedure types; the share of open contracts was around 87% for politically connected companies and 88% for non-connected companies. Likewise, for both groups 8-10% was distributed between negotiated with and without publication type procedures. However, for connected companies the share of negotiated contracts with no prior call for tender publication was around 3 percentage points higher. This is arguably a riskier procedure type as less information is available for other potential suppliers and for the general public.
Figure 8. Procedure type value share of politically connected companies’ contracts (2015-2021) in Bulgaria
Source: GTI and CSD.
An example of using, for instance, direct awards to get contracts through political connections was found in Romania. Ana Kraus, who was a head of Sibiu County Council’s Administrative Service while her husband was employed first as a driver and then as a personal counsellor to former County Council President Ioan Cindrea, awarded telecommunication service contracts to her son, Robert Kraus through direct award catering. Media Impex SRL, the catering company, therefore initially provided this service free of charge, but was later awarded an actual contract, and sent 58 invoices to Kraus amounting to EUR 63,500 from 2013 to 2014.137
By looking at the individual integrity indicators for politically connected and non-connected firms it is possible to assess whether there is a substantial difference between the two groups when it comes to competitiveness, submission and decision dates, as well as procedure type and publication of the call for tender. Different sets of indicators could have been used for different countries depending on the data availability and quality. All the states in the region had a “call for tender publication” indicator. One of the easiest ways to fix tenders is by avoiding publishing the call for tenders in the official public procurement journal, as this makes it harder for non- connected competitors to prepare bids. Another straightforward way of measuring integrity is through the “procedure type” indicator: using less open and transparent procedure types can indicate the deliberate limitation of competition. Too short of too lengthy “advertisement period” and “decision period” can be signaling the corruption risks too. Extremely short number of days between publishing a tender and the submission deadline or between submission and decision deadlines can be a sign of a tender’s requirements being specified for a particular supplier. Too long time periods could happen due to legal challenges suggesting the outright violation of laws. Finally, “single bidding score” is an easy and direct measurement of the level of competitiveness in procurement procedures.
Across the region, the aggregated integrity score is always slightly lower for connected firms, with some variation across countries. For instance, in Hungary and North Macedonia, the difference adds up to around 0.05 points, while in Croatia or Serbia the difference accounted for around 0.1 points. The single bidding indicator is one of the most factors differentiating the two groups of contracts. For some countries, the single bidding integrity level for politically connected firms is lower than for non-connected in about
0.1 points or more (e.g., Serbia or Croatia). In some cases, call for tender publication indicators also showed significant differences. For instance, in Croatia this indicator shows almost twice less integrity for politically connected firms’ contracts. In most of the cases the procedure type similarly showed better integrity for non-connected firms (e.g., in North Macedonia, Croatia, and Bulgaria, the difference is around 0.05 points). At the same time, in some cases advertisement and submission periods showed higher integrity levels for connected firms. For instance, in Romania and North Macedonia both groups scored either almost the same or with connected firms having slightly higher scores.
The Serbian case is a good example of clear difference between the two groups as demonstrated by integrity indicators (Figure 9). Politically connected firms’ contracts are associated with a significantly lower single bid- ding integrity (0.1 versus almost 0.4 for non-connected firms). Submission and decision periods also significantly contributed to the aggregated integrity score difference: both indicators show around 0.05 points difference with non-connected firms. At the same time, call for tender publication indicator and procedure type integrity both scored very high and do not show any significant difference.
Figure 9. Integrity indicators for politically connected and non-connected companies (2015-2021) in Serbia
Source: GTI and EPC.
The effect of politically connected suppliers on procurement integrity
Besides visual differences between two groups, it is important to check if the observed differentiation is statistically significant. The goal of the regression models is to produce a more robust picture of the effects of politically connect- ed firms on overall procurement integrity. A negative estimated coefficient indicates a lower average level of integrity for contracts of politically connected companies compared to contracts of non-connected companies. Therefore, negative effects may signal an increased risk of corruption. The expected outcome is to observe statistically significant and negative coefficients in the level of integrity for politically connected firms’ contracts at least in the majority of indicators.
Depending on the quality of the data, for some countries the models showed more robust and expected results than for others. For instance, in the case of North Macedonia almost all of the variables are statistically insignificant due to the low number of politically connected firms in the dataset. Similarly, the Serbian model is not sufficiently robust due to a low number of observations. On the other hand, Croatian and Bulgarian models have a majority of the indicators as significant predictors for lower integrity level among politically connected firms’ contracts.
Figure 10 shows the estimated coefficients for six separate models in Croatia. In each model, a politically connected company binary (indicating whether for a given contract the supplier company was owned/controlled by a PEP or family member) is regressed on an integrity indicator. The bars show the 95% confidence interval for the given estimate. The effect is significant for the Call for Tender (Cft) publication rate, submission period integrity, and Procedure Type integrity, indicating that on average connected companies are more likely to win contracts without a prior Cft publication, they often receive contracts where there is less time for potential competitors to submit their interest and they are more likely to win contracts with riskier procedure types (such as outright awards). While the predicted Single Bidding integrity (multi-lot procedures) and Decision Period integrity are not significantly different for politically connected companies, the negative estimates suggest that they are still more likely to be lower. The combined Integrity Score highlights that the composite effect of connected companies winning a contract is significantly negative. Overall, the results suggest that politically connected companies in Croatia often win contracts where the risk of corruption is higher.
Figure 10. The effect of politically connected suppliers on procurement integrity in Croatia (2015-2021)
Post COVID corruption risks in public procurement
In order to analyse the regional trends in post-COVID corruption risks, legislative changes were taken into consideration first- in particular, the introduction of emergency procedures for public procurement (e.g., negotiated procedures for buying COVID-related products). Additionally, to compare countries’ CRI trends, the average scores were calculated for three markets. Figure 11 shows the average score for three types of products: a) COVID-19 related, b) general health related, and c) all other public contracting between 2017 and the end of 2021. The types of products were categorised based on the CPV codes associated with the contracts. By COVID-related goods such products were counted as masks, ICUs, etc. Other medical goods were categorised as healthcare market, and everything outside of the two groups is categorised as non-healthcare market products.
Comparison of all countries’ general trends (drops in integrity level after introduction of emergency period by three markets)
All the countries in the region went through an emergency period due to the COVID-19 outbreak, resulting in the changes in legislative regulations of public procurement. Usually, these changes in the regulations were introducing emergency procedures for COVID-related products, allowing for more expedient procedures but also resulting in less transparent procurement, which was therefore more prone to corruption. The newly introduced procedures usually included direct negotiations under which the procuring body was freed from objective assessment and fair specification and evaluation of bidders. For almost all of the countries in the region, the emergency period due to COVID-19 was lifted over the spring of 2022; in Hungary, however, it was re-introduced due to the war in Ukraine, and the state of emergency decree states that special rules apply to the procurements related to the supply of persons recognized as entitled to temporal protection.
The analysis of the regional trends among three markets (healthcare products, COVID products and non-healthcare products) is presented in Figure 11. It can be observed that the level of integrity either goes down for COVID market only (Romania, Hungary, Croatia), or results in a downward trend for the healthcare market with rapid fluctuations in COVID market (Bulgaria and North Macedonia). In the case of Bulgaria, the significant drop in the level of integrity for COVID products can also be observed with some time lag after the introduction of the state of emergency, which can be attributed to the increase in the number of COVID-market contracts in the second quarter.
Figure 11. Three markets, quarterly integrity trends before and after the introduction of emergency periods in
Bulgaria, Croatia, Hungary, North Macedonia and Romania (2017-2021)
In most of the cases the non-healthcare market is not associated with any significant changes due to the introduction of the state of emergency. Specific case studies prove the point regarding COVID-related products and corruption risks associated with them. For instance, in Serbia, NITES Group signed three contracts with the Ministry of Health worth EUR 3.2 million in April of 2020. This coincides with the period when Darija Kisić Tepavčević, sister of the executive director of the NITES branch in Czech Republic – Bojan Kisić, became a member of the COVID-19 Crisis Task Force before becoming a Minister later on the same year. NITES was usually applying to tenders as part of consortia with other companies that also have politically exposed persons in their top management.
Comparisons to the EU-wide benchmark
Figure 12 shows the average CRI score for three types of products for all EU countries between August 2018 and August 2021. COVID-19 and general health products saw an increase in CRI scores after COVID-19 restrictions in February 2020, with little change in other products. Similarly, whereas the trendline of corruption risks in Europe remained relatively the same for non- health products in the 18 months before and after periods, all health-related products (general and COVID-19) saw a steep increase in CRI scores in the after period.
Figure 12. Three markets quarterly corruption risk trends before and after the introduction of emergency period in Europe (2019-2021)
Source: Fazekas, Sanches, and Abdou, 2022 (GTI).
Therefore, the regional trend shown in Figure 11 is not vastly different from the European averages. While the non-healthcare products remain on average at the same level of integrity, changes in corruption risks for the healthcare market and COVID-related products are markedly visible. However, the scale of the trend differs for the region in comparison to European averages. While Figure 12 shows European corruption risks at the beginning of 2020 on the level of 0.15 (or 0.85 in integrity level), the average integrity for the region Figure 11 in the same time frame is around 0.7. The highest pick of corruption risks for European averages was at 0.24 in corruption risks (or 0.76 in integrity level), which constitutes around 0.10 points in the change of CRI score. This scope of change is similar to the regional trend (within 0.10 points change) with the exception for Bulgaria (the integrity level for COVID products dropped by 40% in the second quarter).