Government favoritism in public procurement: evidence from ...1430016/...Daniel Pustan Godkänt...
Transcript of Government favoritism in public procurement: evidence from ...1430016/...Daniel Pustan Godkänt...
DEGREE PROJECT IN TECHNOLOGY AND ECONOMICS, SECOND CYCLE, 30 CREDITS
STOCKHOLM, SWEDEN 2019
Government favoritism in
public procurement:
evidence from Romania
DANIEL PUSTAN
KTH ROYAL INSTITUTE OF TECHNOLOGY SCHOOL OF INDUSTRIAL ENGINEERING AND MANAGEMENT
This page intentionally left blank
Government favoritism in public procurement: evidence from Romania
by
Daniel Pustan
Master of Science Thesis TRITA-ITM-EX 2019:728
KTH Industrial Engineering and Management
Industrial Management
SE-100 44 STOCKHOLM
Regeringens favorisering inom offentlig upphandling: fallet Rumänien
Daniel Pustan
Examensarbete TRITA-ITM-EX 2019:728
KTH Industriell teknik och management
Industriell ekonomi och organisation
SE-100 44 STOCKHOLM
Master of Science Thesis TRITA-ITM-EX 2019:728
Government favoritism in public procurement: evidence from Romania
Daniel Pustan
Approved
2019-12-31
Examiner
Matti Kaulio
Supervisor
Kristina Nyström
Commissioner
Contact person
Abstract
In Romania, the consideration that politicians use their influence to control the public
procurement market is axiomatic. It is no surprise that the country ranks high in perception-
based surveys or the low participation of firms on the procurement market. The more difficult
task is to demonstrate the existence of restrictions to procurement contracts in order to benefit
preferred companies. That is, to measure the extent to which the market is captured by favored
companies.
Employing data on all public procurement contracts in Romania for the period 2009 – 2015,
this paper examines government favoritism in public procurement exerted by political parties.
Using a dynamic panel data approach (Dávid-Barrett and Fazekas 2019), the companies are
classified based on their winning pattern with respect to government change. Favoritism is
observed if winning companies within the government period are also associated with a higher
risk of corruption measured by two alternative approaches.
The findings confirm that procurement market is captured in a low to moderate proportion
(24%) and that the market display patterns of systematic favoritism. This may signal certain
progress registered by Romania to combat political corruption. Arguably, the insensitivity of
perception indicators with respect to this progress is, at least partly, due to media coverage of
the on-going corruption investigations related to the past.
Key-words
Grand corruption, High-level corruption, Political corruption, Clientelism, Favoritism,
Governance, Corruption measurement, Public procurement, Romania
Examensarbete TRITA-ITM-EX 2019:728
Regeringens favorisering inom offentlig upphandling: fallet Rumänien
Daniel Pustan
Godkänt
2019-12-31
Examinator
Matti Kaulio
Handledare
Kristina Nyström
Uppdragsgivare
Kontaktperson
Sammanfattning
I Rumänien, finns det en allmän uppfattning om att politiker använder sitt inflytande för att
kontrollera den offentliga upphandlingsmarknaden. Det är ingen överraskning att landet rankas
högt i perceptionsbaserade undersökningar rörande korruption eller att företags deltagande
inom upphandlingsmarknaden är lågt. Ett svårare uppdrag är dock att påvisa och bevisa
förekomsten av begränsningar kring upphandlingskontrakt med syfte att gynna vissa företag.
Med andra ord så föreligger en utmaning att, genom mätning, påvisa i vilken utsträckning
marknaden inkluderar favoriserade företag kontra hur den exkluderar övriga företag.
Med hjälp av uppgifter om samtliga kontrakt gällande offentlig upphandling i Rumänien under
åren 2009 – 2015, undersöker denna avhandling regeringens och dess politiska partiers
favorisering. Företagen i upphandlingarna klassificeras med hjälp av dynamiskpaneldata
(Dávid-Barrett och Fazekas 2019), baserat på des vinnande mönster kopplat till regeringsbyte.
Favorisering kan observeras om vinnande företag inom regeringsperioden även är förknippade
med en högre risk för korruption som mätts genom två alternativa metoder.
Resultaten bekräftar att upphandlingsmarknaden fångas i en låg till måttlig andel (24%) av
favoriserade företag och att marknaden visar mönster av systematisk favorisering. Resultaten
kan dock signalera om visst framsteg som Rumänien har uppnått för att bekämpa politisk
korruption. Det kan argumenteras att perceptionsbaserade indikatorer fångade inte upp dessa
framsteg, åtminstone delvis, på grund av mediatäckningen av pågående korruptionsutredningar
i Rumänien relaterade till det förflutna.
Nyckelord
Stora korruption, korruption på hög nivå, politisk korruption, klientelism, favorisering,
styrning, korruptionsmätning, offentlig upphandling, Rumänien
Government favoritism in public procurement: evidence from Romania 7
Table of Contents
1. Introduction 13
2. Literature review 18
2.1 Definition and classification 18
2.2 Measures of corruption 23
3. Institutional framework 27
3.1 Legislative setting 28
3.2 Public procurement institutions 30
3.3 Government characteristics 31
4. Method 34
5. Data 41
6. Results 44
7. Robustness analysis 54
8. Conclusions 58
References 61
Annexes 70
KTH ROYAL INSTITUTE OF TECHNOLOGY
8 Daniel Pustan
List of figures
Figure 1 The bi-dimensional clientelistic structure
Figure 2 The comprehensiveness of regulation in Romania
Figure 3 Timeline of cabinets and parties for the period 2009 - 2015
Figure 4 Mean regression residuals by two-percentiles of relative procedure period
(contract date - call for tender date), linear prediction
Figure 5 Mean regression residuals by two-percentiles of relative notification of award
period (award notice date - contract date), linear prediction
Figure 7 CRI-arithmetic by company group, half-year period
Figure 8 Combined market shares of company types, half-year period
Figure 9 CRI arithmetic compared to perception-based surveys (CPI, WGI, GCI)
Figure 10 CRI-PCA compared to perception-based surveys (CPI, WGI, GCI)
Figure 11 Single bid compared to perception-based surveys (CPI, WGI, GCI)
Figure 12 Outliers for relative contract value < 2
Figure 13 CRI-PCA patterns by company group, half-year period
Figure 14 CRI-arithmetic patterns by company group, one year period
Figure 15 CRI-PCA patterns by company group, one year period
Figure 16 CRI-arithmetic patterns by company group, changes in legislation
Figure 17 CRI-PCA patterns by company group, changes in legislation
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 9
List of tables
Table 1 Amendments to GEO 34/2006 during the studied period
Table 2 Romania’s government cabinets during 2009-2015
Table 3 Main characteristics of the dataset
Table 4 System GMM linear dynamic panel regression explaining company market
success
Table 5 Logit regression results on contract level, 2009-2015, number of winners 2≥
Table 6 CRI component weights
Table 7 Bivariate Pearson correlation of CRI and single bid with perception survey
indicators
Table 8 Linear regression explaining relative contract value and CRI
Table 9 Summary of selected studies using objective indicators of corruption
Table 10 Corruption techniques and indicators
Table 11 Summary of selected studies on firm’s political connections
Table 12 Descriptive statistics
KTH ROYAL INSTITUTE OF TECHNOLOGY
10 Daniel Pustan
Abbreviations
ANAP National National Public Procurement Agency
ANRMAP National Authority for Regulating and Monitoring Public Procurement
BI Business International
CIN Company Identification Number
CNSC National Council for Solving Complaints
CPI Corruption Perception Index
CPV Common Procurement Vocabulary
CRI Corruption Risk Index
CRI Corruption Risk Index
CVM Cooperation and Verification Mechanism
DLAF Department for the Fight Against Fraud
DNA National Anticorruption Directorate
FDI Foreign Direct Investments
GCI World Economic Forum
GEO Government Emergency Ordinance
ICRG The International Country Risk Guide
IMD The Institute for Management Development
NAPP National Agency for Public Procurement
NGO Non governmental organization
NUTS Nomenclature of Territorial Units for Statistics
PC Conservative Party
PCA Principal Component Analysis
PDL Democrat Liberal Party
PNL National Liberal Party
PSD Social Democratic Party
SEAP Electronic System for Public Procurement
TED Tenders Electronic Daily
UCVAP Unit for Coordination and Verification of Public Procurement
UDMR Democratic Alliance of Hungarians in Romania
UNPR National Union for the Progress of Romania
USL Social Liberal Union
WCY World Competitiveness Yearbook
WEF World Economic Forum
WGI Global Competitiveness Index
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 11
Acknowledgements
On the very outset of this paper, I would like to extend my sincere and heartfelt obligation
towards all the personages who have helped me in this endeavor. Without their active guidance,
help, cooperation and encouragement, I would not have made headway in the project.
I would like to express my gratitude to my supervisor, Kristina Nyström for her patience and
valuable advice for structuring the ideas of this paper. I am extremely thankful to Mihály
Fazekas - co-author of the paper on which the methodology of this study relied on - for the kind
responds to my questions.
I extend my gratitude to Sebastian Buhai for the precious guidance related to the topic of this
research and to Hayk Sargsyan for his useful comments and encouragement when most needed.
I am deeply indebted to my former manager, Larisa Lăcătuș, for creating the conditions to
expand my knowledge alongside my professional development. On the same note, I gratefully
acknowledge the assistance of Karin Asplund with the Swedish translation.
Lastly but not least, my highest appreciation goes to my wife, Maria, who experienced firsthand
this long and strenuous process. Your limitless understanding and unconditional support made
it possible for this journey to end successfully.
Any omission in this brief acknowledgment does not mean lack of gratitude.
Stockholm, December 31st, 2019
Daniel Pustan
KTH ROYAL INSTITUTE OF TECHNOLOGY
12 Daniel Pustan
1. Introduction
Corruption is estimated to cost the European Union economy €120 billion per year which is
almost the equivalent of the EU annual budget (European Commission 2014). The true cost of
corruption, however, includes arguably inestimable social costs such as: poorer income
distribution, environmental protection as well as undermining trust in democratic institutions
(European Commission n.d.).
Thirty years after communism fell in Romania, it seems that the fight against corruption
remains an ongoing battle which affects the country on many levels. In terms of country
perception, Romania ranks amongst the most corrupt countries in the European Union (Hall
2019) while favoritism appears to be the rule for the majority of government decisions (Schwab
2016). Additionally, Romania is held back from joining the free movement Schengen area
because of the lack of progress in combating corruption as reflected by the Cooperation and
Verification Mechanism (CVM) . 1
The civil society became harshly conscient of the fatal consequences of corruption following
Colectiv nightclub fire in November 2015. A fire that tragically ended the lives of many, only
because, allegedly, corrupt practices led to systematic violation of the rules which, if respected,
could have saved lives. Protests that followed denouncing political elites’ tolerance or
association to corruption led to Victor Ponta’s resignation, the first Prime Minister accused of
corruption while still in charge (BBC 2015).
Considering the external and internal pressure - by means of European Commission reports and
civil protests - everything indicated that the message was understood and corruption would be
irremediably defeated in a battle with the government as protagonist.
Nothing could be further from reality. Liviu Dragnea, chief of the ruling Social Democratic Party
(PSD) and president of the Chamber of Deputies in Parliament who already had a suspended
sentence, was risking prison sentence in his ongoing corruption trial. (France24 2018). In an
attempt to spare the PSD chief and other high profile politicians from the ongoing investigations
or prison sentences, the Romanian Government issued an Emergency Ordinance 13/2017 (to
enter into effect 10 days after publication in the Official Gazette).
1 The Cooperation and Verification Mechanism is the mechanism that European Commission
uses to monitor reform implementation in the area of justice and rule of law.
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 13
Massive protests followed against what was perceived as legalization of corruption acts inducing
the government to abolish the controversial Act prior to its entry into force. Should the
Emergency Ordinance came into effect just for one day, Liviu Dragnea could have escaped the
three and a half years of prison that he was sentenced to in May 2019 (BBC 2019).
Unfortunately, political corruption does not represent just a few isolated cases. Only in 2017,
other 12 high level officials were prosecuted among which: ministers, deputies, Secretary of
State, Senators and government Secretary-General National (Anticorruption Directorate Activity
Report 2017).
In this context, public procurement is especially vulnerable to corruption (European
Commission 2014; OECD 2016). Perception based reports advance that public procurement is
manipulated in proportion of 80% (Euractiv 2014). This severe trust deficit is probably one of
the main causes of the low share of firms taking part in the bidding process. In fact, the number
of unique bidding companies during the seven years period of observation was 15,568 (1.3%) out
of a 1.2 million active companies (Oprea 2017). By comparison, in Sweden the number of
companies participating in public procurement bids during 2013 was 212,000 representing
18.80% of the total number of registered companies (Lukkarinen, Morild and Jönsson 2015).
Furthermore, the low number of referrals received from public administrations (9% in 2017) as
well as the rare occasions in which local authorities decide not to become claimants in order to
recover incurred losses (National Anticorruption Directorate 2017), might be an indicator of
participation in a larger network of favoritism or the resignation that such networks exist and
there is not much to do against them.
The overflow of the anti-corruption theme outside the institutional framework into society has
caused deep rifts among many walks of life. In these last years, many have seen in the Social
Democrat Party (PSD) the very essence of corruption, a party that was ripping the resources of
the state shamelessly. At the same time, for others it represented the ultimate guardian for the
ordinary citizen ripped off by multinationals in their run to make profits at any cost.
Perceptions became less reliable while sliding towards a binary, oversimplified division of reality
- between “us” and “they” (Bogdan 2018). Polarization deepened and fact based arguments lost
KTH ROYAL INSTITUTE OF TECHNOLOGY
14 Daniel Pustan
prominence in an ideological battle, arguably, between traditional and modern values (David
2015).
The necessity to regain a common ground between the polarized sides became obvious. This, in
turn, required a measurement tool that focuses on objective behavior rather than opinions and
perceptions.
For a long time, measurement of corruption relied mostly on perception based indicators.
However, perception may not be of much help in this scenario. Furthermore, the little variation
of indicators makes them susceptible of not being responsive to change, in spite of significant
changes in Romania from one year to another. (Oman and Arndt 2006).
Companies associated with high profile politicians are often successful on the procurement
market, earning the label of “companies subscribed to contracts with the State'' (Befu 2015;
Marcovici 2015; Andrei 2018). In fact, Mungiu-Pippidi (2015) highlights that political
connections are the main indicator of favoritism in Eastern Europe. But do high level politicians
use their influential capacity to divert contracts towards favored companies?
According to Gherghina and Volintiru (2017) clientelistic model, parties are interested to engage
in reciprocal relationships by redirecting public contracts towards favored private contractors in
exchange for benefits - which in most cases are monetary. Parties will use this financing to
sustain or increase the electoral support and consequently consolidate their power.
Figure 1 The bi-dimensional clientelistic structure
Source: Gherghina and Volintiru 2017, p. 122
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 15
Therefore, if the answer is affirmative as anticipated, what size of the procurement market is
captured by the political elites?
To my knowledge, the only research on Romania’s procurement market with objective indicators
were performed by Doroftei (2016) and Doroftei and Dimulescu (2015). They examined
contracts over €1,000,000 in the construction sector for the period between 2007 and 2013.
Both papers conclude that single bidding and political connections explained 32% of the number
of contracts awarded during the entire research period and 44% in pre-election years.
Using a more sophisticated measurement of corruption, Dávid-Barrett and Fazekas (2019)
found that 50%-60% of procurement market in Hungary suffer from systematic favoritism while
only 10% of companies were politically favored in the United Kingdom. Their approach
consisted of two steps: (1) analyze the influence that government change has on companies
procurement outcome and (2) investigate the risk of corruption associated with the winning of
contracts. The sample comprised period 2009-2012 and was limited to contracts: (i) awarded by
national authorities, (ii) above the mandatory EU-threshold and (iii) only applied to competitive
markets.
Following Dávid-Barrett and Fazekas (2019) method, this paper explores empirically to what
extent the governmental political parties in Romania exert control over the public procurement
process; in other words, it quantifies the level of government favoritism in Romania. The
analysis extends to all public procurement contracts in Romania during 2009-2015 period. The
sample was not restricted to contracts awarded only by national central administration. This
decision was taken considering the centralization of power in Romania both at the level of public
administration as well as that of party organization.
The contribution to the research literature is fourfold. First, to my knowledge it is the first time
to apply Dávid-Barrett and Fazekas (2019) innovative methodology on public procurement in
Romania. Secondly, it expands the method by performing a cross-governmental analysis with a
focus on the main party in government. Thirdly, this study introduces the Principal Component
Analysis as an alternative to simple average method for weighting the indicators composing the
Corruption Risk Index. Lastly, it provides a corrected large dataset suitable for scientific
research.
KTH ROYAL INSTITUTE OF TECHNOLOGY
16 Daniel Pustan
The method follows three steps. First, companies are classified based on the residuals after
estimating the panel regression using Arellano-Bond system GMM transformation. Second,
Corruption Risk Index is calculated both as: (1) as a weighted average (Dávid-Barrett and
Fazekas 2019) and (2) using the Principal Component Analysis approach. Thirdly, the market
share of aggregated suspicious companies is calculated and companies are aggregated by group
to identify patterns of systematic favoritism.
This paper analyzes a dataset of over 650,000 observations and finds that 24% of the Romanian
procurement market is associated with systematic government favoritism. However, it is
noteworthy, that the study only captures specific clientele for each of the two parties. It does not
reflect favoritism in situations where companies buy influence from both parties or public
administration officials that withstand government change.
Furthermore, highly innovative companies may be mislabeled as “surprise winners” as both
exhibit similar pattern behavior (Dávid-Barrett and Fazekas 2019). This risk is reduced for this
specific dataset by the presence of “lowest price” award criterion in 94% of the procedures
during the analyzed period. Typically, due to the high costs of research, innovative companies do
not compete for price and they are not expected to take part in these tenders. Two other
limitations are discussed further.
The Common Procurement Vocabulary (CPV) codes are the basis for the CPV division (2 digit
level of CPV codes) and CPV product group (3 digit level of CPV codes) which in turn serve as
proxies for the market. Although generally accepted as a market indicator, it is not a perfect
instrument due to flaws in the CPV code structure such as: (1) inconsistent hierarchical
tree-structure, (2) inconclusive classification levels, (3) unmatching between subordinate and
superordinate level and (4) not mutually exclusive codes (Attström et al. 2012).
Aggregating the data based on a half year period “is optimal for retaining a high level of
granularity while also taking into account the erratic character of many public procurement
markets” (Dávid-Barrett and Fazekas 2019, p. 13). However, the random length of the cabinets’
period does not perfectly match the half year aggregation. To circumvent this problem, an
alternative model was analyzed where the cabinet variable was replaced by variables consisting
in legislation changes. Both models yield comparable results.
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 17
In view of the ensuing inefficient distribution and allocation of resources, corruption and in
particular clientelism constitutes, perhaps, the antithesis of sustainable development.
Clientelism skews the distribution of state resources (Ferreira Rubio and Andvig 2019) with
potential negative consequences in all three aspects: environmental, social and economic.
Among the social consequences are the increase of inequality (Ferreira Rubio and Andvig 2019)
and erosion of confidence in institutions. Corruption affects the most vulnerable groups as
infrastructure projects may be vastly financed to steer money towards favorite companies in
detriment of education and health (Mungiu-Pippidi 2015).
Corruption is also negatively correlated to the growth in genuine wealth per capita. This is
clearly illustrated by Aidt (2010) with Denmark. If levels of corruption in Denmark would
suddenly increase to those of the countries that make it to the bottom of Corruption Perception
Index, the development would become unsustainable.
Narrowing it further down, the implications of this paper on sustainability - described as making
decisions in the present so as not to impair future real incomes (Markandya and Pearce 1988) -
are highly practical. The economic impact is given by the importance of public procurement
which rises to almost 50% of public spending in developing countries. Within the social aspect,
it represents an invitation for society to focus on facts in an attempt to find a common ground
where this was lost. It promotes the equality and importance of individuals in the social system
while promoting the ranking of ideas. The positive impact on the environment stems from the
implicit responsible and efficient use of resources as alternative to particularism.
The remaining of the paper is structured in seven sections. First, the literature is reviewed with
special emphasis on the various definitions of corruption and the challenges encountered in
measuring it. Second, the institutional framework is discussed including the changes in
legislation during the period, the specific institutions related to public procurement and the
composition of government. Following, the method section describes the process of suspicious
companies identification based on their market success and the subsequent analysis of the
conditions under which they as shown by the Corruption Risk Index. The remainder of the paper
is organized as follows: data, results, robustness analysis and conclusions.
KTH ROYAL INSTITUTE OF TECHNOLOGY
18 Daniel Pustan
2. Literature review
2.1 Definition and classification
Corruption is a term used to cover a wide range of different practices. It appears that to some
extent the term is framed by prevalent cultural norms (Werner 1983) as well as depending on
the context (Johnston 1996). The difference is obvious when comparing the more developed
American democracy with the Romanian mindset of government. The American view strives to
limit the authority and divide the power (Huntington 1968). By contrast, the Romanian
dominant psychological profile is that of a collectivist person and concentration of power (David
2015).
In an interview at the Romanian Business Leaders Summit (Roșca 2018), David argues that
favoritism and nepotism are embedded in the Romanian culture and what is considered
corruption from a Western perspective, it is simply perceived as a proof of loyalty. David
continues by exemplifying with the implementation of a Western institution in Romania - that of
a parliamentary cabinet - a parliamentary with a collectivist profile will hire a person who they
can trust most - a relative - instead of the best expert (Roșca 2018).
This shows that conflict of interests as a specific form of corruption, may also be the
unintentional result of institution implementation within a culture different than the one where
the institution was born. However, nepotism as a cultural accepted practice is not specific only
to Eastern Europe. In India and Africa it is not only acceptable but expected (Leys 1965; Bayley
1966).
Huntington (1968) goes as far as to suggest that corruption is a hardly avoidable externality of
modernization [of institutions]. Scholars supporting the “grease the wheels” hypothesis
highlight potential malfunctions or inefficiencies that corruption may help circumvent the
functioning of the institutions and government policies. It can (1) increase the quality of the
service (Leys 1965; Bayley 1966) - by keeping the underpaid talented employees in the public
service, (2) speed-up processes like waiting in the queue (Huntington 1968; Lui 1985) and (3)
foster efficiency as an imperfect auction system in competitions for a limited number of
resources or instead of non optimal mechanical quota based systems (Leff 1964; Bayley 1966),
(4) act as an insurance against bad public policies ensuring predictability and give a sense of
being in control during transition period (Leff 1964; Bayley 1966), (5) non-violent manner to
access the establishment and (6) provide political representation or unity in spite of
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 19
irreconcilable ideologies or interests, potentially higher and better investment (Leff 1964; Bayley
1966; Werner 1983).
Most empirical research, however, does not support the claim of positive effects of corruption.
Dependency on grafts acts as an incentive to deliberately slow down or limit the provision of the
service and to block the access to other competent workers in order to secure a large stream of
cash (Lui 1985; Kurer 1993). Furthermore, in order to compensate for the bribe, the firm can
provide products of lower quality (Rose-Ackerman 1997).
The “sand in the wheels” effect of corruption is more obvious when looking at its impact on
economic growth. Various studies confirmed the negative relationship between corruption and
growth consequence of a decrease in private investment (Mauro 1995) or the quality of the
infrastructure (Tanzi and Davoodi 1998) even in countries with a weak rule of law and inefficient
government (Méon and Sekkat 2005). Mo (2001) estimated that 1% increase in corruption led to
a decrease by 0.72% in private investment.
So, what is corruption? Most often, corruption is associated with bribery in the public eyes.
However, it embraces a large palette of forms (bribery, embezzlement, theft and fraud, graft,
revolving door, clientelism, nepotism, cronyism, conflict of interests, patronage, etc.) targeting
different levels of authority and adapting the appropriate methods of sophistication to context
and culture. It is no wonder that there is no consensus among scholars on the definition of
corruption.
The general accepted definition, promoted by international organizations such as Transparency
International, refer to corruption as “the misuse of entrusted power for private gain”
(Rose-Ackerman 2008, p. 551; Andersson and Heywood 2009, p. 746; Xu 2016, p. 86). As
already anticipated, this definition is generally associated with bribery in a bureaucratic context
(Fazekas, Cingolani and Tóth 2016; Fazekas 2017; Mungiu-Pippidi 2013 acc. Mungiu-Pippidi
2015). Scholars classify this type as petty, bureaucratic or low-level corruption.
Another category of corruption - object of this study - takes place at the higher-levels in public
positions, benefits of vast discretion and lacks explicit detailed rules (Fazekas, Cingolani and
Tóth 2016). This type of corruption, defined as a deviation from the universalism norm (Kurer
2005: Mungiu-Pippidi 2015) is known in the literature as high-level corruption, government
favoritism, institutionalised grand or political corruption, (Jain 2001; Graycar 2015; Fazekas,
KTH ROYAL INSTITUTE OF TECHNOLOGY
20 Daniel Pustan
Cingolani and Tóth 2016; Fazekas, Tóth and King 2016; Fazekas 2017). Thus, government
favoritism in public procurement “refers to the allocation and performance of public contracts
by bending universalistic rules of open and fair access to government contracts in order to
benefit a closed network while denying access to all others.” (Fazekas, Cingolani and Tóth 2016,
p. 3; Fazekas, Tóth and King 2016, p. 370).
Ultimately, high-level corruption may deviate into state capture if it reaches the stage where
institutions or laws are expressly created to benefit a private group. At this point corrupt
decisions are implemented in a corrupt-free manner (Søreide 2002; Fazekas 2017). This stage is
more characteristic to economies in transition from communism to capitalism and to some
extent through lobby groups in developed economies (Graycar 2015).
Among the different types of political corruption, clientelism distinguishes as a special form of
particularism because of its contingent or “quid pro quo” nature. Clientelism is the distribution
of all kinds of public goods in exchange for political support (Hicken 2011). The major risk areas
of capture by political elites are the appointments in the public sector, privatization projects,
subsidy programmes and public procurement contracts.
Hicken (2011, p. 291) highlights that the difference between clientelist practices and other types
of exchange consists in the fact that the former “always comes with strings attached”. These
benefits aim to consolidate or increase political power and are typically in the form of voting.
However, it may come in the most diverse forms such as: donations, favorable media coverage,
revolving door and public procurement contracts. It is in such a scenario where principal-agent
models prove their limitations.
Until recently clientelism was understood as a type of principal-agent relationship. However, the
reciprocal character of clientelism does not allow for a clear distinction between the role of the
principal and that of the agent. The relationship may reverse to the point where politicians end
up holding voters accountable for their votes (Stokes 2005).
Instead, Gherghina and Volintiru (2017) propose a bidimensional model of clientelism. Thus,
political elites use state resources, in particular procurement contracts, to favor or promise
favors to private companies in exchange for support, typically donations. These funds are then
used as a competitive advantage against other parties to attract votes (Gherghina and Volintiru
2017).
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 21
Complementing Gherghina and Volintiru (2017) model, Dávid-Barrett and Fazekas (2019)
highlight three stages in which politicians can bend the universalistic norms: formation of the
law, implementation by bureaucracy and monitoring of implementation or evaluation. This
paper focuses on the implementation stage, one that requires a new violation of law with each
transaction (Dávid-Barrett and Fazekas 2019).
One direction of literature sheds light on the important role of political connections on
companies’ performance. Political connections are identified as one or more of the three
indicators: (1) political campaign donations, (2) board seat connections or family ties and (3)
stock holdings by politicians.
With few exceptions, companies benefit greatly of political connections in three main areas: 2
stock value, financing and procurement contracts. Fisman et al. (2012) finds no influence of the
stock value related to Chimney’s personal ties. Furthermore, Jackowicz, Kozłowski and Mielcarz
(2014) study reveals a negative correlation between political connections and firm profitability
which, they suggest, is likely due to the instability of Poland’s politics.
More connected with the subject of this paper, the studies related to public procurement reveal
that political connections have a major impact on contract allocation even in countries with a
solid institutional setting such as Denmark (Amore and Bennedsen 2013) or the United States
(Goldman, Rocholl and So 2013, Luechinger and Moser 2014).
The increase in the value of procurement contracts varies. In the United States, campaign
contribution results in additional procurement contracts of about 2.5 times the value of
donations (Bromberg 2014). The same figure in Brazil is at least 14 times (Boas, Hidalgo and
Richardson 2014) while in the Czech Republic it can reach 100 times (Titl and Geys 2019).
Gürakar and Bircan (2019) shows that closed bids are more expensive than open bids especially
if the winner is a company political connected to the party in power.
Scholars conclude that politically connected firms are on average, less productive and more
likely to win procurement contracts (Zhuravskayay 2012; Brugués F., Brugués J. and Giambra
2018). In particular, donating firms receive more generous contracts or a higher amount of
2 For a detailed list of the studies, please refer to Annex C
KTH ROYAL INSTITUTE OF TECHNOLOGY
22 Daniel Pustan
smaller contracts, flexible completion dates and less bidding competition (Brogaard, Denes and
Duchin 2015; Titl and Geys 2019; Canayaz, Martinez and Ozsoylev 2015) suggesting a potential
clientelistic link between politicians and companies.
In a more targeted study, Hyytinen, Lundberg and Toivanen (2018) identifies favoritism
towards the in-house units of municipalities in public procurement for cleaning services in
Sweden. Nonetheless, they associate favoritism with municipalities’ efforts to support local
employment rather than “outright corruption” (p. 422). This conclusion indicates that
favoritism is viewed from the petty corruption perspective rather than as a form of
particularism.
The only available studies on favoritism with objective indicators on Romania is the paper of
Doroftei and Dimulescu (2015) confirmed by the work of Doroftei (2016). The analysis drew on
public procurement contracts over €1,000,000 on construction sector in Romania between
2007 and 2013. Both studies conclude that single bidding and political connections explain the
number of contracts awarded in proportion of 44% in pre-election years.
It is worth noting that in spite of the clarity of the single bidding indicator, its simplicity makes it
vulnerable to circumvention (Fazekas 2017). One basic technique to adopt is for the connected
firm to agree in advance on the content of the bid with one or two other firms at their
participation in the procurement process (Fazekas, Tóth and King 2016). A corruption case on
the airport runway extension in Cluj-Napoca illustrates such a situation. In this case, the
politically connected business man admitted it was a bid rigging. In fact, to create the
impression of legality, he went so far as to ask one of the participants to dispute the bid. Only
because he knew that the company would not stand any chance as it did not qualify (Bacalu
2018).
Dávid-Barrett and Fazekas (2019) propose a more complex measurement for favoritism. For a
company to be considered favored, it has to have both suspicious winning patterns and higher
risk of corruption as measured by CRI. The CRI is a composite indicator that besides the single
bid includes another 5 input indicators. Their analysis revealed that favored companies control
50%-60% of public procurement market in Hungary, a large number compared to 10% in the
United Kingdom.
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 23
2.2 Measures of corruption
Measuring corruption is a challenging task. Until recently, the corruption measures consisted
mainly in surveys/indicators of attitudes, perception and experience of corruption or a mix of
these, namely composite indices. The respondents are mostly experts or firms and to a lesser
extent citizens (Fazekas, Tóth and King 2016). Reviews of institutions, scientific analyses, audits
and investigations of particular cases complement these measures.
Examples of surveys based on expert opinions are the International Country Risk Guide (ICRG)
indices by Political Risk Services and Business International Corporation (BI). The first was
used by Tanzi and Davoodi (1998) while the former was used for the first time by Mauro (1995).
The Global Competitiveness Index assembled by the World Economic Forum (WEF) and the
Institute for Management Development (IMD) in the World Competitiveness Yearbook (WCY)
are constructed based on business executives opinion. One cross-country experience based
survey among citizens is the Global Corruption Barometer constructed by Transparency
International. Transparency International developed the most famous composite index,
Corruption Perception Index followed by the Corruption Index from the World Bank.
In favor of these indicators stands their extensive use and the strong correlation with each other
as well as with other economic and development indicators (Blackburn, Bose and Emranul
Haque 2010). Nonetheless, in spite of their broad use, critics abound with respect to surveys and
perception indicators.
Andersson and Heywood (2009) emphasize the remarkable similarity in survey questions,
respondents’ group sample and type of corruption targeted (most often bribery) These
characteristics may explain in good measure the high degree of correlation between these
indicators.
Further, perception does not imply having directly experienced corruption in any way. It may be
driven by different factors such as media attention (Golden and Picci 2005). It may also be
inconsistent in time due to a later discovery and broadcasting of past corruption acts (Kenny
2009) or unconsciously using previous reports on corruption as benchmark (Golden and Picci
2005).
KTH ROYAL INSTITUTE OF TECHNOLOGY
24 Daniel Pustan
Other researchers add the fact that answers are shaped by factors such as diverse as individual
characteristics, ethnic diversity, engagement in social activities (Olken 2009), the phrasing of
the questions (Xu 2016), cultural differences between the interviewee and interviewer (Bayley
1966), or expats impressions on their host country (de Maria 2008).
Most interviews are focused on executives or opinion leaders which may indicate a
representativeness bias, reflecting the views of a specific group not the one of the general
population (Fazekas, Tóth and King 2016). In addition, they may reflect only certain types of
corruption while omitting others (Svensson 2005).
Most indices are weakly correlated with reality (Weber Abramo 2008). In a study comparing
villager's perception on corruption with a more objective measure in Indonesia, Olken (2009)
finds that villagers were only able to predict a small part (only 8%) of the real level of corruption
in the project. This was due to the fact that they were better at detecting corruption in
marked-up prices then inflated quantities which generated most corruption in the project.
The main drawback with composite indices is the conceptual imprecision. They are defined by
various indices that compose them which in turn can vary in time (Knack 2007). Their
reductionist approach makes them insensitive to variations in sectors or regions and oblivious to
underlying phenomena indicated by outliers (Lancaster and Motinola 2001).
The Corruption Perception Index (CPI) receives particular criticism as having a Western
business-orientation (de Maria 2008, Andersson and Heywood 2009) and that it is more
reliable to reflect corruption levels in developed rather than developing countries (Golden and
Picci 2005). In spite of these disadvantages, the stakes are high for developing countries by
conditioning financial support on the country’s ranking taken as a proxy for the country's reform
progress (Oman and Arndt 2006). Misinterpretation of the CPI can lead these countries into a
“corruption trap” (Andersson and Heywood 2009, p. 747).
The risk of misunderstanding the true measurement increases exponentially with indices where
corruption is only a part of a larger group of measurement. This is the case of the broadly use
International Country Risk Guide (ICRG) which measures “the political risk involved in
corruption” (Lambsdorff 2006, 82) not the level of corruption.
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 25
Unfortunately, survey of experiences do not provide a higher degree of reliability. Part of the
firms in the sample group may be on the market as a consequence of corruption and may have
no interest to report or even misreport (Svensson 2005; Kurtz and Schrank 2007). Even when
they do report, they mostly reflect petty corruption as citizens do not come in contact often with
high-level corruption (Kenny 2009; Fazekas, Tóth and King 2016).
The reviews of institutions lack measurement precision and are based upon implicit
understanding regarding the functioning of the observed institutions (Fazekas, Tóth and King
2016; Fazekas and Kocsis 2017). The precision of the measurement increases with scientific
analyses, audits and investigation of individual cases. However, their resource-intensive
character makes them available in a limited number and even less useful for comparison
because of their specific goals and risk of state capture (Fazekas, Tóth and King 2016; Fazekas
and Kocsis 2017).
The risk of state capture is circumvented by Escresa and Picci’s (2017) more objective index of
corruption cases based on the Foreign Corrupt Practices Act. Nevertheless, Stephenson (2014)
warns about more or less implicit assumptions that may compromise the validity of this index,
such as: (1) conflict of interest (state ownership in the economy of the targeted country), (2)
biased indicator because of the mix of exports and Foreign Direct Investments (FDI), (3) the
number of investigations in a country being influenced by the USA politics (e.g. high risk
country are investigated more often) and (4) the endogenous character of exports.
Indeed, in the latest years researchers drawn their attention towards the so-called “objective”
indicators of corruption capable to measure behavior instead of relying on perceptions or
experiences. One direction of research analyzed the differences in prices for standardized
products or services (Di Tella and Schargrodsky 2003; Bandiera, Prat and Valletti 2009;
Hyytinen, Lundberg and Toivanen 2018). Other researchers contrasted official reports with
surveys and independent audits to highlight corruption in the award of benefits, grants and
projects in infrastructure (Reinikka and Svensson 2004; Olken 2006; Olken 2007). A different
measure was suggested by Golden and Picci (2005) which consisted in the ratio between the
physical infrastructure and the corresponding cumulative expenditures in 20 regions of the
country. Finally, as we mentioned earlier, using proxies such as donations or connected board of
directors, political connections constitute in a good measure the focus of recent research.
KTH ROYAL INSTITUTE OF TECHNOLOGY
26 Daniel Pustan
Kaufmann, Kraay and Mastruzzi (2011) warns with regards to the two problems that stem from
any kind of data when measuring corruption. First, the “noise” in the measurement, which
represents the difficulty or impossibility to distinguish between corruption, incompetence,
ignorance or other source of noise. This is exemplified in Golden and Picci (2005, p. 37)
statement when describing corruption: “... money is being lost to fraud, embezzlement, waste,
and mismanagement; in other words, corruption is greater.” Second, the imperfect relationship
between specific measures of corruption and overall corruption. There is a risk of generalizing
corruption based on a certain measured type of corruption.
In particular, these indicators reliability is shadowed by their dependency on the context, high
costs to replicate and the use of a single proxy for corruption which only in the most optimistic if
not fanciful scenario coincides with the main driver of corruption (Fazekas, Tóth and King
2016).
To address these shortcomings, Fazekas, Tóth and King (2016) created the Corruption Risk
Index (CRI). This composite indicator of corruption in public procurement is constructed as the
weighted average of various indicators of corruption. The indicators, also called red flags, are
identified through the review of the literature and qualitative interviews. Then, they are
validated through regression analysis (only indicators which prove to be powerful and
significant become part of the index). Further, they are divided between corruption outcomes
such as “single bid” and corruption inputs such as “the procedure type” or “the weight of non
price evaluation criteria”.
The main benefits of this index are not only that it derives from objective data consisting of
more than a single indicator but it is also comparable across time and countries. Furthermore, it
follows directly from the definition of high-level corruption in public procurement, namely, the
bending of prior, explicit rules for the benefit of a limited network (Mungiu-Pippidi 2006;
Rothstein and Teorell 2008) and it can be validated with other proxies (Fazekas and Kocsis
2017).
An important inconvenience is that the indicators are not ranked based on their importance.
Hence, the resulting score is not as explicit as it should be. In fact, the weighted average CRI
score may only reflect the number of indicators composing the CRI, in the absence of categorical
and continuous variables which receive special treatment.
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 27
It might be adventurous to read a higher score as more prone to corruption as it is not evident
that a higher number of indicators equals a higher risk of corruption. The presence of a single
indicator - single bidding - may be sufficient to achieve the corrupt goal while the presence of
other numerous indicators may be insufficient to steer the contracts for the favorite company.
To address this issue, the paper introduces an alternative approach to weighted average,
specifically the Principal Component Analysis.
KTH ROYAL INSTITUTE OF TECHNOLOGY
28 Daniel Pustan
3. Institutional framework
The crucial role that institutions play for the prosperity of a country was demonstrated by
Acemoglu and Robinson (2012). In the same note, most suggested measures to combat
corruption refers to features of institutional environment. Hence, a larger body of research
developed under the umbrella of “good governance”, a notion closely related with that of
institutional environment (Andersson and Heywood 2009, p. 750).
Transparency, meritocracy or the separation of bureaucrat’s careers from political influence as
well as the quality and quantity of regulation are some important areas of investigation (Islam
2006; Lindstedt and Naurin 2010; Bauhr and Grimes 2013; Bauhr et al. 2019).
3.1 Legislative setting
Prior to its adhesion to the European Union as a full member in 2007, Romania transposed the
public procurement Directives 2004/17/EC and 2004/18/EC of the European Parliament and of
the Council into national legislation. The main legislation governing public procurement in
Romania up until 2016 was represented by Government Emergency Ordinance (GEO) 34/2006.
The Ordinance has suffered numerous and substantial modifications while into force. Only
during the analyzed period I identified 17 changes.
Table 1 Amendments to GEO 34/2006 during the studied period
2009 2010 2011 2012 2013 2014
GEO 19 GEO 76 Law 279 Law 76 GEO 4 GEO 51
Law 84 Law 284 Law 114 GEO 44 Law 9
GEO 72 Law 278 GEO 77 GEO 31
GEO 35
Law 166
Law 193
In addition to the primary legislation, secondary and tertiary legislation apply to public
procurement. The secondary legislation refer to Government Decisions consisting of the detailed
arrangements on the application of the primary legislation. The orders and instructions issued
by the National Agency for Public Procurement (NAPP) compose the tertiary legislation. The
existence of this extra layer of legislation highlights the problem of red tape in Romania’s public
procurement (Doroftei and Dimulescu 2015).
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 29
Figure 2 The comprehensiveness of regulation in Romania
Source: Europam.eu
On May 23rd, 2016 a new package of public procurement laws consisting of three laws 98/2016,
99/2016 and 100/2016 replaced the Government Emergency Ordinance (GEO) 34/2006 and its
subsequent modification acts. These laws are meant to transpose Directives 2014/23/EU,
2014/24/EU, 2014/25/EU.
Mungiu-Pippidi (2015) shows that two thirds of the contracts awarded to favored companies
come from non-EU funds. Corruption with EU funds is foreseeable less prevalent considering
that most potential irregularities with these funds are covered by the Law 78/2000 on
prevention, discovering and sanctioning of corrupt acts. The national funds do not benefit of
such a level of protection. In this case, most prosecutions are undertaken based on the law
punishing the abuse or malfeasance of office (Euractiv 2014).
This law was amended later, after the period of analysis, specifically on June 15, 2016, by the
Constitutional Court decision number 405 which specified that the syntagm “performs faulty”
should be interpreted as “performs by breaking the law”. By “the law” is understood only laws,
simple Government Acts or Emergency Acts. This way secondary regulations, such as
Government Decisions or Ministerial orders, are excluded (Toma 2016).
KTH ROYAL INSTITUTE OF TECHNOLOGY
30 Daniel Pustan
To decrease the period of delays in contract execution, on June 28, 2014 the government issued
an Emergency Act to increase the obstacles for submitting complaints. It requires a “legitimate
interest” as well as a guarantee of 1% and up to 100,000 EUR to be paid by complainer. The
guarantee was forfeited in the event that the complaint were rejected (regardless of the reason:
in substance, inadmissible or by way of a plea) (Avocat.net 2014; Euractiv 2014). On March 19,
2015, the Constitutional Court decided that the Emergency Act 51/2014 was unconstitutional.
3.2 Public procurement institutions
Until mid 2015, the institutional framework consisted of three institutions with competences
exclusively in relation to public procurement (Aprahamian and Pop 2016). The main responsible
body was the National Authority for Regulating and Monitoring Public Procurement (ANRMAP)
with a role in policy and regulatory supervision, controls the manner of the contract award
(ex-post) and applies legal sanctions where necessary. The second public body was the Unit for
Coordination and Verification of Public Procurement (UCVAP) with attribution in verification of
the procedural aspects related to the contract award process (ex-ante control). On May 20, 2015
both agencies were merged into one, namely, the National National Public Procurement Agency
(ANAP).
The National Council for Solving Complaints (CNSC) is the third public body specific to public
procurement system and represents the administrative, non-judicial first instance to solve
complaints lodged under public procurement jurisdiction. As a final remedy procedure, CNSC
decision can be challenged before the Court of Appeal located within the same range as the
contracting authority’s headquarters (Directorate-General for Regional and Urban Policy
(European Commission) and PWC 2016).
The Court of Accounts represents the main oversight body. It conducts ex-post audits related to
all types of funds reporting their findings to ANRMAP and suspicious irregularities under the
criminal law to the National Anticorruption Directorate (DNA). Other entities involved in
monitoring public procurement are: the Management Authorities and the Department for the
Fight Against Fraud (DLAF), the Competition Council, the Authority for Certifications and
Payments and the National Integrity Agency (Doroftei and Dimulescu 2015; Directorate-General
for Regional and Urban Policy (European Commission) and PWC 2016)
It is also worth mentioning the Electronic System for Public Procurement (SEAP) operated by
the Romanian Agency for Digital Agenda (formerly known as the National Management Centre
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 31
for the Informational Society). It is an electronic procurement system with the aim to increase
efficiency and transparency in the public procurement system. The platform was upgraded and
integrated with relevant institutions in public procurement on April 2, 2018 (Romanian Digital
Agenda Agency 2018).
3.3 Government characteristics
During the seven years period, Romania had eight Cabinets and three Prime Ministers (though
one of them for a very short period of time). The first party, the Democrat Liberal Party (PDL),
governed between December 22, 2008 and April 26, 2012 when it was dismissed through a
no-confidence motion. The second party, the Social Democratic Party (PSD), was in power
between April 27, 2012 and November 17, 2015.
Following parliamentary elections, Boc I cabinet, made of PDL, PSD and the Conservative Party
(PC), started its mandate on December 22, 2008 and was dismissed by Parliament through a
no-confidence motion on October 13, 2009 (Mediafax 2009). This was soon after PSD and PC
alliance left the government on October 1, 2009. Prime Minister Emil Boc and ministries
belonging to PDL party continued to ensure the interim until December 23, 2009.
On the aforementioned date, Boc II cabinet was invested by the Parliament as Traian Basescu
was re-elected president (Politica Românească 2011). Thus PDL was able to form a
parliamentary majority with the National Union for the Progress of Romania (UNPR) - a party
formed later by independent parliamentarians who either left or were dismissed from their own
party after 2008 parliamentary elections (Ciubotaru 2013).
For clarity purposes, it should be mentioned that 2004 was the last year when Presidential and
Parliamentary elections coincided. This is due to the fact that the Presidential mandate began to
last 5 years while parliamentary continued to be of 4 years. Hence Parliamentary elections took
place in December 2008 while presidential elections were held in December 2009 (Centrul
pentru Integritate în Business 2013).
Cabinet Boc 2 ended its mandate on February 6, 2009 after Emil Boc resigned as prime minister
following street protests. After a three days period of interim, cabinet Ungureanu was invested.
It had an independent prime minister but the ministers were occupied by the National Liberal
Party (PNL), UNPR and the Democratic Alliance of Hungarians in Romania (UDMR). This
KTH ROYAL INSTITUTE OF TECHNOLOGY
32 Daniel Pustan
cabinet only lasted for three months as it was dismissed by the Parliament on April 27, 2012
through a non-confidence motion.
After the dismissal of cabinet Ungureanu, Victor Ponta became prime-minister being in charge
of four consecutive governments between May 7, 2012 until November 5, 2015 when he resigned
following street protests claiming that the tragedy from “Colectiv” club could have been avoided
with more efficient measures to combat corruption (Tran 2015).
The first Ponta cabinet, formed by the Social Liberal Union (USL) formed by PSD and PNL was
in charge until parliamentary elections in December 2012. Cabinet Ponta 2 was formed after the
elections and co-opted UNPR in the government. PNL leaving the government lead to
government restructuring and co-opting the hungarian minority party, UDMR, which resulted
into cabinet Ponta 3. PNL decided to break the coalition suspecting that their ally will not keep
the agreement and will launch their own candidate for presidency instead of supporting PNL’s
candidate.
Cabinet Ponta 4 was formed due to UDMR’s departure from the government as a result of a
massive vote in the presidential election of their electorate in favor of Klaus Iohannis,
contra-candidate of prime minister in charge to presidency.
After a short interim period generated by the prime-minister resignation as consequence of
street protests following Colectiv nightclub fire, the independent Ciolos cabinet was invested.
The investment by the parliament took place on November 17, 2015. Ciolos government was in
power until December 2016 parliamentary elections.
A summary of government’s succession in presented in table 2.
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 33
Table 2 Romania’s government cabinets during 2009-2015
Cabinet Period Prime
minister Party Other parties
Type of
ending
Boc I 2008,
Dec 22
2009,
Dec 23 Emil Boc PD-L PSD+PC (PUR)
No-confidence
motion
Boc II 2009,
Dec 23
2012,
Feb 9 Emil Boc PD-L UDMR, UNPR
Government
resignation
Ungureanu 2012,
Feb 9
2012,
May 7
Mihai
Razvan
Ungureanu
Independent PDL, UNPR,
UDMR
No-confidence
motion
Ponta I 2012,
May 7
2012,
Dec 21
Victor
Ponta PSD PNL, PC (PUR)
Parliamentary
elections
Ponta II 2012,
Dec 21
2014,
Mar 5
Victor
Ponta PSD
PNL, PC (PUR),
UNPR
Government
restructuration
Ponta III 2014,
Mar 5
2014,
Dec 17
Victor
Ponta PSD
PC, UDMR,
UNPR, PLR
Government
restructuration
Ponta IV 2014,
Mar 5
2015,
Nov 17
Victor
Ponta PSD
UNPR, ALDE
(PLR + PC)
Government
resignation
Figure 3 presents the two government periods together with the party composition of
government cabinets during the observation period.
Figure 3 Timeline of cabinets and parties for the period 2009 - 2015
KTH ROYAL INSTITUTE OF TECHNOLOGY
34 Daniel Pustan
The average period time of governing for a cabinet was 12 months during the 7 years observation
period. The shortest period in power was that of cabinet Ungureanu (3 months) while the
longest running cabinet was cabinet Boc 2 (25 months). For simplicity, the interim periods were
considered jointly with the cabinet period that preceded them. Furthermore, this paper
proposed to capture the influence of the main parties (PDL and PSD) along a
multi-governmental period which included various cabinets.
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 35
4. Method
This paper uses a quantitative approach defined as “entailing the collection of numerical data, a
deductive view of the relationship between theory and research ... and an objectivist conception
of social reality” (Bryman 2012, p. 149). In this study, quantitative research is preferred over
qualitative research as it allows testing hypotheses based on directly observable behavior. The
increase in government transparency and availability of big data on public procurement
contracts made this type of research possible.
Further, a particular form of longitudinal analysis, namely, panel analysis was performed after
aggregating the data on a half yearly basis. The main benefits of using panel data are that it
accounts for heterogeneity and allows to study dynamic factors such as the spells of contract
value won by companies (Hill, Griffiths and Lim 2011).
Following Dávid-Barrett and Fazekas (2019) two hypotheses are verified. First, it is identified
for which companies the change in government affects the value of contracts won. Second,
corruption indicators or “red flags” are analyzed to determine whether the change in companies’
winning tender is also associated with a higher risk of corruption.
Companies benefit of favoritism where both conditions are met. Additionally, favoritism is
considered systematic if the patterns of winning contracts in the presence of more red flags than
the rest of the market can be observed at the level of company type.
Fine shifts in government spending or more sophisticated methods of corruption are the most
reasonable explanation when only the first hypothesis is true. If only the second hypothesis is
true, it is probably because firms buy influence from both parties or other authorities capable to
influence the process (Dávid-Barrett and Fazekas 2019).
Controlling for changes in spending preferences as a result of differences in policy priorities
(CVMti), little variation is expected in contract winning tenders as a result of the change in
government.
To identify the companies for which the value of contracts won was affected by the change in
government, I use the Arellano-Bond system GMM transformation of the following dynamic
panel regression model:
KTH ROYAL INSTITUTE OF TECHNOLOGY
36 Daniel Pustan
CVti = α + β
1 CVt-1i + β
2 CVt-2i + β
3 CVMti + β
4 MMi + β
5 Cabt + U
i + Ɛit
The abbreviations of the model stand for the following concepts:
CVti : the contract value won by company in period t
α : the constant term for the sample
CVt-1i : the contract value won by company i in past period (t-1)
CVt-2i : the contract value won by company i in past period (t-2)
CVMti: the contract value spent in main market (CPV level 3)
MMi : the sectoral dummy for the main market (CPV level 3)
Cabt : dummy representing the cabinet in period t
Ui : the company fixed effect
Ɛ it : the error term
I ran the two versions of the above model - fourth root and natural log contract value - on a
subsample of companies winning contracts in at least two different periods (removing 17,819
observations). The reason for using fourth root is the ability to rescale the data while keeping the
observations with zero contract value during the observed period. The two models reveal favored
companies at different stages of maturity. The fourth root version explore the case of both well
established and incumbent companies. The natural log version analyzes only the more
experienced companies in the procurement market as it removes the companies with zero
contract value observations during the analyzed period (Dávid-Barrett and Fazekas 2019).
During the first nine months, PDL and PSD governed together in cabinet Boc I. However, as
important as it may be, given that the period of governing together took place during the first
two half-year periods and the model uses two lags, these observations are not used. They are
automatically dropped because of perfect collinearity as consequence of the fact that all the
observations are zero after the first two periods.
Based on the regressions company-specific error term, companies having ties to government 1
are labeled “surprise losers” and those with government 2 are called “surprise winners”.
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 37
Specifically, companies that have an above period average regression error under one
government and below period average regression error under the other government are labeled
as “surprise losers” and “surprise winners”. Companies with stable residual patterns (above or
below period average in both governments) are considered “stable companies”.
The same terminology as Dávid-Barrett and Fazekas (2019) is used in this paper: “surprise
loser” and “surprise winner”. These terms are defined in relationship to the second government
period and because their use might give rise to misinterpretation, clarification may be necessary.
Thus, the reader is cautioned not to extend the negative and positive connotations of these terms
to the entire period of study. “Surprise losers” while possibly discriminated during the second
government period, might have benefited from preferential treatment during the first
government period. Therefore, it should be borne in mind that “surprise losers” are the
companies suspicious of having ties with government 1 while the “surprise winners” are the
firms suspicious of links with government 2.
The regressions set-up controlled for the following: (i) government spending patterns resulting
from different policy preferences or priorities measured by contract value spent in main market,
(ii) market and technological specifics measured by CPV division (2-digit CPV) and (iii)
regulatory as well as government specificities (e.g. parties composing the government) measured
by cabinet.
Once identified the companies with a suspicious winning pattern, I analyze whether they won
under conditions with a higher potential for corruption. In order to do so, the surprise winners
and losers are cross checked with the Corruption Risk Index (CRI).
The Corruption Risk Index is a relatively recent tool developed by Fazekas, Tóth and King
(2016) aiming to measure the degree of corruption in a more objective manner as a reliable and
cost efficient alternative to current measurements. The index is based on previous research
(Charron et al. 2017, Fazekas and Kocsis 2017, Dávid-Barrett and Fazekas 2019) referring to not
less than 30 quantitative indicators and 20 corruption techniques in public procurement were
identified. A summary table with all these indicators is presented in annex B (Fazekas, Tóth and
King 2013).
KTH ROYAL INSTITUTE OF TECHNOLOGY
38 Daniel Pustan
The model includes one outcome indicator - single bid - and various input indicators. The
elementary corruption risk indicators (red flags) that constitute the CRI are identified based on
the literature review:
1. Single bid (outcome indicator)
2. Call for tender publication (submission phase)
3. Procedure type (submission phase)
4. Weight of non-price evaluation criteria (assessment phase)
5. Procedure period (assessment phase)
6. Notification of award period (assessment phase)
7. Contract type (assessment phase)
Because “submission deadline” was not present in the dataset, it is not possible to build the
continuous variable indicators “length of advertisement period” and “length of decision period”.
However, two similar indicators are used, namely “procedure period” and “notification of award
period”. The “procedure period” indicator represents the number of days between the call for
tender and contract date. The “notification of award period” computes the number of days
between the contract date and notification of award date.
Binary logistic regression is used to determine the suitability of each composite index in the
index excluding those indicators that do not prove to be significant and strong predictors of
single bidding (Fazekas and Kocsis 2017). The logistic model is:
Pr (single bidder = 1) = 11+e−zi
Zi = α + β
1 Sij + β2 Aik + β
3l Dil + β
4m Cim + Ɛi
Where:
Zi - the logit of a contract being a single bidder contract
α - the constant of the regression
Sij
- the matrix of j corruption inputs of the submission phase for the ith contract such as length
of submission period
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 39
Aik
- the matrix of k corruption inputs of the assessment phase for the ith contract such weight
of non-price evaluation criteria
Dil
- the matrix of l corruption inputs of the delivery phase for the ith contract such contract
lengthening
Cim
- the matrix of m control variables for the ith contract such as the number of competitors
on the market
Ɛ i - the error term
β1, β2, β3l, β4m - the vectors of coefficients for explanatory and control variables
Afterwards, component weights for each variable are derived. The outcome indicator - with no
regression coefficient - receives the weight of 1. Also, each significant corruption input receives
the weight of 1 while those variables or variable categories that do not prove significant are
discarded (weight zero).
The categories within the categorical variables received component weights ranked according to
their regression coefficient. For instance, in the case of two significant categories, the one with
the highest coefficient would receive the weight of 1 while the other one receives the weight 0.5.
For continuous variables, threshold were identified using linear regression analysis and the
residual distribution graphs to determine discrete jumps in residual values. The control
variables represent the most important alternative explanation for the outcome.
KTH ROYAL INSTITUTE OF TECHNOLOGY
40 Daniel Pustan
Figure 4 Mean regression residuals by two-percentiles of relative procedure period (contract
date - call for tender date), linear prediction
Figure 5 Mean regression residuals by two-percentiles of relative notification of award period
(award notice date - contract date), linear prediction
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 41
Both figures display three different period-groups. One which overestimates the probability of
single received bid, one which underestimates this probability and, lastly, one that approximate
it.
Moreover, each corruption input was introduced step by step starting with those from the
earliest phases. The inputs that prove to be insignificant were not introduced in the CRI
composite indicator.
The control variables in the regression are: (i) institutional endowments measured by type of
issuer (national, county or municipal), (ii) market and technological specifics measured by CPV
division (2-digit CPV), (iii) contract size measured by log of contract value in euro, (iv)
legislation measured by year and (v) contracting authorities specificities measured by activity
type.
Finally, CRI is calculated in two ways. First, as the arithmetic average of the individual risk
indicators - red flags (Fazekas, Tóth and King 2016):
CRI = j w
j * CIjiΣ
Secondly, CRI is computed using an alternative technique, namely the Principal Component
Analysis (PCA).
The validation of the Corruption Risk Index was done both at macro and micro level. At the
macro-level I compared the Corruption Risk Index with three notorious perception surveys: The
Corruption Perception Index (CPI) from Transparency International, World Bank’s World
Governance Indicators (WGI) and the Global Competitiveness Index (GCI) developed by The
World Economic Forum.
This last indicator, GCI, covers a much broader area besides corruption (e.g. efficiency of
government) for which reason only the most relevant indicator for this purpose was selected,
specifically, favoritism in decisions of government officials (1.07).
The validation of the Corruption Risk Index at micro-level is done by looking at the correlation
of the relative contract value and CRI. The relative contract value is calculated as the ratio
between the actual contract value and the estimated contract value (Fazekas, Tóth and King
2016). A higher ratio of relative contract value represent a higher risk of corruption.
KTH ROYAL INSTITUTE OF TECHNOLOGY
42 Daniel Pustan
A second micro-level indicator, namely, winner’s country of origin (Fazekas and Kocsis 2017)
was not applicable to this study. The purpose of this indicator is to analyze the correlation of
companies’ main office located in a tax haven country related to the CRI. However, no winner of
public procurement contracts during the 2009-2015 reported their main office in a country
included on the EU non cooperative tax jurisdictions and high risk third country list (European
Parliamentary Research Service 2018).
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 43
5. Data
The public procurement data contains all public procurement contracts in Romania between
2009 and 2015. The original datasets were published on data.gov.ro in the form of 11 .csv files by
the Romanian Agency for Digital Agenda (Mihalache 2013). Alternative data was available from
The European procurement journal, Tenders Electronic Daily, which comprised of the contracts
above a certain threshold.
Nevertheless, the data published by the Romanian national agency was prefered because it
allows a better understanding of procurement market in its entirety. Moreover, the complete
dataset is suitable to offset the corruption technique of fragmenting big contracts into smaller
contracts. This technique is used to conceal irregularities or simply to benefit of more permissive
rules.
Unfortunately, the poor quality of the data required a substantial work in order to reach an
adequate quality for scientific research. One obstacle consisted in the fact that the elementary
variable, the Company Identification Number (CIN), included multiple inaccurate observations.
Keeping these observations in the analysis required to manually correct those instances. In spite
of the time consuming process, this effort was carried out with the hope that other researchers
may also benefit from it.
Over 5,000 companies were searched by name on the Ministry of Finance and other specialized
websites (e.g. listafirme.ro and totalfirme.ro). The entities that were still not found, most of
them consisting of self-employed persons, were assigned a unique identification comprising
their names without spaces. In addition, all foreign companies’ CIN was verified. Kompass.com
website together with the official companies registry websites of their countries of origin were
employed for this purpose.
Once the correct CIN was assigned for the winning companies, the names and addresses were
extracted from the dataset published by the Ministry of Finance on the website data.gov.ro. This
dataset consists of all economic operators and public entities. For the operators without a CIN as
well as for foreign companies, the name and address were extracted from a dataset created for
this purpose.
KTH ROYAL INSTITUTE OF TECHNOLOGY
44 Daniel Pustan
Inconsistencies between the CIN and companies’ name in the dataset were tackled by
triangulating with company’s address variable. Whichever between the two - CIN or company
name - matched the address was considered correct while the other was changed accordingly.
Companies that changed their name for branding or legal purposes were assigned the new name
during the entire observation period. Examples of such companies are Engie Romania SRL,
previously GDF Suez Energy and Veolia Energie Prahova SRL which changed from Dalkia
Thermo Prahova SRL due to the acquisition by the former.
The same process was conducted for the contracting authorities related variables. First, the
validity of their CIN was verified. In the event of a mismatch between the entities’ CIN and the
name variable, these were triangulated with the address variable. Subsequently, the name and
address were imported from the Ministry of Finance dataset (Ministry of Finance 2018).
To ensure the quality and integrity of the database, other correction measures were taken for the
existing errors and inconsistencies in all variables. Some examples of the issues encountered
are: broken lines, leading and trailing spaces, consecutive spaces, use of diacritics and non
alphanumeric characters, etc.
The association of companies was treated similarly to Doroftei (2016) and Doroftei and
Dimulescu (2015). Specifically, the contracts awarded to a consortium were assigned to the
leader. This is consistent with the thesis that most frequently it is the leading company with
political connection that enables the consortium to win the contract. Nevertheless, this approach
may generate inflated contract values for some companies.
The regressions were fitted on a subsample of competitive markets. Competitive markets are
defined by product group (CPV group - level 3) and regions (NUTS1) with at least three
participants (Fazekas, Tóth and King 2016). Non-competitive markets under normal
circumstances, such as those for specialized services or defense, were excluded to reduce false
positive due to the underlying assumption that the lack of competition signals corruption.
(Dávid-Barrett and Fazekas 2019, Fazekas and Kocsis 2017).
Albeit the purpose of this paper is to measure governmental favoritism, the sample was not
limited only to government related contracts for two reasons. First, in a centralized country like
Romania the government has significant influence on the local governing authorities through
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 45
the allocation of funds. Secondly, most of the vote buying performed by local party organizations
is carried on behalf of the national party organization.
In any case, the distribution of procurement by the type of contracting authority is as follows:
the government and its related agencies is the main contractor with 45% of the total value of
acquisitions. It is followed by the local administration authorities with 35% and the county level
administration with 11%. The remaining 9% consists of non-governmental organizations
(NGO’s) and companies.
Company level analysis was performed on a half-yearly aggregated database comprising the
companies that won contracts in at least two periods between 2009 and 2015 period. Half-year
period “is optimal for retaining a high level of granularity while also taking into account the
erratic character of many public procurement markets” (Dávid-Barrett and Fazekas 2019, p.13).
The government periods were defined in relation with the government party, the party providing
the political support to the prime minister and holding the majority of the ministerial positions
in the cabinet. Apart from this, one year period of transition covering Ungureanu and Ponta I
cabinets was allowed as a means to capture the differences between two established parties in
power (Dávid-Barrett and Fazekas 2019).
Table 3 Main characteristics of the dataset
2009 2010 2011 2012 2013 2014 2015 Total
Total number of contracts awarded 115,868 121,170 103,342 106,507 65,413 72,560 67,662 652,522
Total number of winners 9,890 10,208 9,144 7,933 7,041 6,509 6,255 16,010
Total number of contract. authorities 3,983 3,647 3,331 3,189 2,814 2,813 2,775 6,826
Total value of awarded contracts (mil. EUR)
11,100 10,400 15,200 13,400 13,900 16,300 16,400 96,700
KTH ROYAL INSTITUTE OF TECHNOLOGY
46 Daniel Pustan
6. Results
The analysis reveals moderate evidence of partisan favoritism in Romanian public procurement
market. Specifically, 24% of the public procurement market in Romania appear to be
dominated by favored companies of which 15% surprise losers and 9% surprise winners. This is 3
above the 10% market capture in the United Kingdom’s but lower than 50%-60% results in
Hungary (Dávid-Barrett and Fazekas 2019).
Outcomes similar to Hungary were expected given the comparability between the two countries
in many aspects including the geographical proximity, the communist background and the
equivalent rankings in the corruption perception index. However, there is a resemblance,
specifically in that, like Hungary,the Romanian public procurement market is characterized by
systematic favoritism. That is to say that favoritism patterns are identifiable at the level of
company type.
Using Arellano-Bond system GMM transformation of the dynamic panel regression, I find
moderate evidence of company performance through the analyzed period (Table 4). The first 4
model has a moderate explanatory power while the second model has a relatively low
explanatory power: 0.69 and 0.38. 5
In Dávid-Barrett and Fazekas (2019) using a smaller sample on Hungary (log model), the
predicting power drops dramatically by almost 90%. This can be a sign of a non-representative
sample for the population or there can be influential observations which are lost in the natural
log model. The fact that in my case the sample decreases much less is one of the reasons for the
reduced decrease in predicting power for the log with respect to the fourth root model.
3 Results for both “fourth root” and “log” contract value regressions were taken into account
4 Sargan test of overidentification of instruments show that the null-hypothesis could not be rejected
5The linear correlation coefficient between predicted and observed outcome was used
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 47
Table 4 System GMM linear dynamic panel regression explaining company market success
fourthrootvaloareeur lnvaloareeur
fourthrootvaloareeur = L1 0.0274***
(0.00453)
fourthrootvaloareeur = L2 0.0220***
(0.00431)
lnvaloareeur = L1 0.0732***
(0.0171)
lnvaloareeur = L2 0.120***
(0.0148)
Control variables
fourthrootCVmainmarket Y N
lnCVmainmarket N Y
main market Y Y
cabinet Y Y
Constant 14.32*** 6.814***
(1.421) (2.054)
Observations 192,120 28,331
Number of panelID 16,010 6,592
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
I identified 3,860 suspicious companies (2,359 surprise losers and 1,501 surprise winners) out of
the total of 16,010 companies. These company groups have a CRI trajectory consistent with
favoritism. Surprise losers win in the presence of more red flags than the rest of the market
during the first government period and surprise winners have a higher CRI than the rest of the
market during government 2 period.
There is a significant positive persistence effect. In particular, a one percent increase in contract
value increases by 7% after one period and by 12% after two periods. For the fourth root model,
significant positive effects with one or two lags are also found in the fourth root model.
KTH ROYAL INSTITUTE OF TECHNOLOGY
48 Daniel Pustan
Figure 7 CRI-arithmetic by company group, half-year period
Figure 8 displays the overall market share trends for each company group. This shows clear
contrasting market outcome trajectories between surprise losers and surprise winners. The
difference is lower during the period in which the two parties governed together.
Figure 8 Combined market shares of company types, half-year period
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 49
In the following, results for the Corruption Risk Index are presented.
First, not publishing the call for tenders strongly predicts single bidding. The average probability
of a single received bid increases by 135%.
Second, contrary to expectations, only the “accelerated negotiation” and “negotiation without
call for tender” procedure types display higher risk of corruption signs. The contracts with a
single bid in an accelerated negotiation tend to be awarded to winning companies 141% more
often as compared to open bid contracts. The “negotiation without call for tender” award 92%
more single bid contracts than open procedure does. All the other non-open procedure types
predict decreases in single bidding. However, “competitive dialogue”, “limited accelerated
tendering” and “limited tendering” are insignificant and according to the theoretical framework
will not be included in the composite indicator.
Third, the indicator “weight of non-price criteria” is a powerful predictor but with a weak to
moderate impact in explaining single bidding contracts. In the complete model the probability
of single bidding decreases by about 13%.
Fourth, an extremely short procedure period (up to 24 days) is a very strong predictor of single
bidding with 218% compared to those of over 77 days. The contracts awarded between 24 and 77
days from the call for notice date have a more moderate predicting capacity of single bid with
63%.
Fifth, the notification of award period is associated with a low corruption risk. The contracts
where the award notice was published within 8 days after the signing of the contract have a 17%
higher risk of corruption while there is a lower risk of corruption roughly by the same
percentage for the contracts signed over 27 days as compared to contracts between 8 and 27
days.
Sixth, compared to supplies, the works/construction contract type decreases the risk of
corruption by 54% whereas the services increases the risk with 35%.
KTH ROYAL INSTITUTE OF TECHNOLOGY
50 Daniel Pustan
Table 5 Logit regression results on contract level, 2009-2015, number of winners 2≥
singleb
Procedure type
Ref. cat: 3, Open tendering
1, Competitive dialogue -0.475
2, Invitation to tender -0.916***
4, Limited tendering -0.130
5, Limited accelerated tendering -0.0659
6, Negotiation -0.336***
7, Accelerated negotiation 1.412***
8, Negotiation without contract notice 0.918***
No call for tenders 1.345***
Weight of nonprice
Ref. cat:Lowest price
Most economically advantageous tender -0.129***
Procedure period
Ref. cat: 3, 77 days - max
1, 0 - 24 days 2.179***
2, 24 - 77 days 0.631***
4, missing 0.223***
Notification of award period
Ref. cat: 2, 8 - 27 days
1, 0 - 8 days 0.176***
3, 27 days - max -0.194***
Contract type
Ref. cat:Supply
2, Works -0.540***
3, Services 0.351***
Control variables
Type of issuer Y
CPV division Y
lnContractValue Y
Year Y
Activity type Y
Constant -1.663***
Observations 362,815
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 51
Table 6 presents the component weights derived based on the regression results.
Table 6 CRI component weights
Variable Component weight
Single bid 1
Procedure type
Ref. cat: 3, Open tendering
7, Accelerated negotiation 1
8, Negotiation without contract notice 0.75
6, Negotiation 0.5
2, Invitation to tender 0.25
No call for tenders 1
Weight of nonprice
Ref. cat:Lowest price
Most economically advantageous tender 1
Procedure period
Ref. cat: 3, 77 days - max
1, 0 - 24 days 1
2, 24 - 77 days 0.66
4, missing 0.33
Notification of award period
Ref. cat: 2, 8 - 27 days
1, 0 - 8 days 1
3, 27 days - max 0.5
Contract type
Ref. cat:Supply
3, Services 1
2, Works 0.5
KTH ROYAL INSTITUTE OF TECHNOLOGY
52 Daniel Pustan
The molding of the CRI on the definition of grand corruption and the fitted regression models
are the main supporters for the validity of the Corruption Risk Index. The macro-level
perception surveys only reinforces this validity.
Table 7 Bivariate Pearson correlation of CRI and single bid with perception survey indicators
Indicator CRI-arithmetic CRI-PCA single bid years
CPI -0.7388 -0.8357 -0.4885 7
WGI (1.07) -0.6862 -0.8321 0.0045 7
GCI -0.4205 -0.3414 -0.3741 7
Figure 9 CRI arithmetic compared to perception-based surveys (CPI, WGI, GCI)
Figure 10 CRI-PCA compared to perception-based surveys (CPI, WGI, GCI)
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 53
Figure 11 Single bid compared to perception-based surveys (CPI, WGI, GCI)
The main reason for the moderate level of correlation stems from the fact that perception-based
indicators are more adequate to compare the levels of corruption across countries for longer
periods of time (Escresa and Pici 2016 acc. Fazekas and Kocsis 2017) as they capture mostly
pitty corruption and they are not sensitive to change.
A more reliable indicator is the relative contract value. However, the estimated contract value
variable present various inexactitudes that may bias the analysis. The following are some of the
observed inaccuracies: (1) use of commas to separate decimals incorrectly, (2) Unrealistically
low estimated contract values (below 20€), (3) the estimated contract value appears for the
entire project but the real contract value is divided in various batches and (4) Total contract
value is repeated because some companies won contracts together but each company is
presented with the total amount as if each of them has won the total amount. The outliers with a
relative contract value below five are presented in figure 12.
Figure 12 Outliers for relative contract value < 2
KTH ROYAL INSTITUTE OF TECHNOLOGY
54 Daniel Pustan
For lack of a better solution, considering the time constraints of this study, the regression was
run on a subset of companies with relative contract value lower or equal to 1 (one) in order to
minimize the number of erroneous observations captured in the analysis.
Table 8 Linear regression explaining relative contract value and CRI
(1) (2) (3) (4) (5) (6)
relativeCV relativeCV relativeCV relativeCV relativeCV relativeCV
cri_arithmetic 0.284*** 0.418***
-0.00323 -0.00408
cri_pca 0.373*** 0.677***
-0.00477 -0.00669
singleb 0.147*** 0.124***
-0.000977 -0.00114
Control variables
TipActivitateAC N Y N Y N Y
AuthorityType N Y N Y N Y
Year N Y N Y N Y
CPV division N Y N Y N Y
lnCV N Y N Y N Y
Constant 0.643*** 0.699*** 0.713*** 0.377*** 0.462*** 0.528***
-0.00121 -0.000791 -0.00044 -0.00841 -0.00868 -0.00823
Observations 289,637 224,711 289,637 212,008 169,687 212,008
R-squared 0.026 0.027 0.072 0.154 0.159 0.16
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
The results confirm the expectations. One additional red flag (1/7 points higher) increases the
price by 4% - 6%. Similar values are obtained for CRI-PCA where the price increases by 5.3 -
9.7%. Single bidding contracts have on average 12.4 - 14.7% higher prices than contracts with at
least two bidders.
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 55
7. Robustness analysis
In this section robustness checks are presented. The CRI-PCA approach and the alternatives in
time periods and the change in control variables confirm the robustness of the model.
First, the CRI weighted using Principal Component Analysis (PCA) display similar results to the
CRI calculated as an arithmetic average.
Figure 13 - CRI-PCA patterns by company group, half-year period
Secondly, changing the aggregated period for the company analysis from half-year to one year
yields similar results. According to this specification, the number of favored companies remains
almost identical with that for the half year period. Specifically, 13.22% surprise losers and 11.15%
surprise winners are identified. In addition, the CRI trajectories are consistent with systematic
favoritism.
KTH ROYAL INSTITUTE OF TECHNOLOGY
56 Daniel Pustan
Figure 14 CRI-arithmetic patterns by company group, one year period
Figure 15 CRI-PCA patterns by company group, one year
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 57
Thirdly, variables representing all the changes to the procurement law during the observation
period were introduced replacing the variable cabinet as a control for changes in legislation.
Again, the results confirmed the robustness of the base model. A very similar number of favored
companies was identified (24.15%) and systematic favoritism was evident in these cases, as well.
Figure 16 CRI-arithmetic patterns by company group, changes in legislation
Figure 17 CRI-PCA patterns by company group, changes in legislation
KTH ROYAL INSTITUTE OF TECHNOLOGY
58 Daniel Pustan
8. Conclusions and suggestions for further research
This paper has examined the political influence at the highest level on public procurement
market in Romania during 2009-2015. Drawing on the most recent research and novel
measurement tools of corruption the study reveals that 24% of the Romanian procurement
market is determined by systematic partisan favoritism.
This number is below the 50% market associated with partisan favoritism in Hungary
(Dávid-Barrett and Fazekas 2019). Considering the geographical proximity as well as cultural
and institutional similarities, a similar score to Hungary was expected.
There are a few possible explanations why that was not the case. First, for comparison purposes,
this paper followed Dávid-Barrett and Fazekas (2019). However, the methodology differed in
two aspects. Specifically, Dávid-Barrett and Fazekas (2019) sample was restricted to contracts
above the EU-threshold and only purchases made by the government and its agencies.
On the other hand, this study included all contracts during the observation period at all levels of
public administration. Not restraining the study only to government level contracts is justified
by the high level of centralization both administratively and politically in Romania. A very
efficient control over the local administration is carried out through budget allocation.
Secondly, the more vibrant and engaged civil society in Romania may have had a significant
impact in mitigating corruption. The analyzed period was especially troubled in which two
governments had fallen as a consequence of street protests.
A third explanation is related to the high frequency of party switching by Romanian politicians,
famous for their lack of ideological attachment. The migration from one party to another also
moves the clientele between parties as the favors will be offered through the party in which the
politician is currently found. This situation hinders the possibility to identify groups of clientele
specific for parties. A good illustration illustration of mass migration from one party to another
occurred in August 2014 with the GEO 457/2014. Then, thousands of mayors were able to
migrate to PSD party without ceasing to hold office as the law stipulated (Bărbulescu 2014).
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 59
The result is more in line to Doroftei (2016) and Doroftei and Dimulescu (2015) findings. Their
study discovered 32% favoritism in construction procurement sector in Romania. It should be
noted though that construction sector is amongst the most susceptible of corruption.
In conclusion, the results of this paper indicate some progress made by Romania in the fight
against high-level corruption in public procurement. However, it invites to moderate optimism
since a decrease in one form of corruption does not necessarily imply a drop in the overall
corruption. The method applied in this paper cannot rule out alternative practices such as
companies buying influence from different parties.
In any case, these results open the way for debates to combat corruption through good
government practices that are based on objective proxy measures. In addition, this study
provides empirical support for EU anti-corruption policies. It may help focus the attention on
specific and quantifiable forms of corruption entailing more precise recommendations by the
European Commission’s in the effort to support the fight against corruption in Romania within
the Cooperation and Verification Mechanism.
In order to be able to fully exploit the potential benefit for the society of this field, further
research is needed. First, as it was shown objective proxy measures are promising tools in
understanding the underlying mechanisms of corruption. Yet they are still in an early phase of
development. Reliable indicators not only in extension but also in depth, ranking the
importance of corruption risk indicators, should be developed.
Another direction of research is represented by the analysis of corruption at local level as well as
the interaction between central and local government related to favoritism. Finally, party
switching in Romania is a far too common practice. Studying political favoritism at a more
granular level, that of the individual politician, may bring important insights on the drives of
political favoritism.
KTH ROYAL INSTITUTE OF TECHNOLOGY
60 Daniel Pustan
References
Acemoglu, D. and Robinson, J. (2012). Why nations fail: the origins of power, prosperity, and poverty. 1st
ed. New York, NY: Crown Business.
Aidt, T. S. (2010). Corruption and Sustainable Development. https://doi.org/10.17863/CAM.5249
Amore, M. and Bennedsen, M. (2013). The value of local political connections in a low-corruption
environment. Journal of Financial Economics, 110(2), pp.387-402.
Andersson, S. and Heywood, P. M. (2009), The Politics of Perception: Use and Abuse of Transparency
International's Approach to Measuring Corruption. Political Studies, 57: 746-767.
Andrei, A. (2018). Topul firmelor care lucrează cu banii statului. Contracte cât bugetul Iaşului. Ziarul de
Iași. [online] Available at:
https://www.ziaruldeiasi.ro/stiri/topul-firmelor-care-lucreaza-cu-banii-statului-contracte-cat-bugetu
l-iasului--181634.html [Accessed 19 Dec. 2019].
Aprahamian, A. and Pop, G. (2016). Building landmarks, smoothing out markets: An Enhanced
Competition Framework in Romania. World Bank Studies. Washington, DC: World Bank. License:
Creative Commons Attribution CC BY 3.0 IGO
Arvate, P., Barbosa, K. and Fuzitani, E. (2018). Party Expertise, Campaign Donation and Government
Contracts: Evidence from an Electoral Quasi-Experiment. INSPER Working paper 377/2018.
Attström, K., Kröber, R., Jähne, J., Bergschneider, J., Thomä, J., Swysen, T., Scholzen, D. and
Trautenmüller, T. (2012). Final Report: Review of the Functioning of the CPV Codes / System.
[online] Berlin: Ramboll. Available at:
https://ec.europa.eu/docsroom/documents/21583/attachments/1/translations [Accessed 9 Dec.
2019].
Auriol, E., Straub, S. and Flochel, T. (2016). Public Procurement and Rent-Seeking: The Case of Paraguay.
World Development, 77, pp.395-407.
Avocatnet.ro. (2014). Garantia de buna conduita obligatorie pentru contestarea licitatiilor publice.
[online] Available at:
https://www.avocatnet.ro/articol_38023/Garantia-de-buna-conduita-obligatorie-pentru-contestarea
-licitatiilor-publice.html [Accessed 20 Dec. 2019].
Bacalu, M. (2018). Uioreanu şi Bene merg la închisoare. Sentinţă finală în dosarul de corupţie aeroport
[online] Muzicainstantelor.ro. Available at:
https://muzicainstantelor.ro/index.php/2018/01/31/uioreanu-si-bene-merg-la-inchisoare-sentinta-f
inala-in-dosarul-de-coruptie-de-la-aeroportul-cluj/ [Accessed 23 Nov. 2019].
Bandiera, O., Prat, A. and Valletti, T. (2009). Active and Passive Waste in Government Spending:
Evidence from a Policy Experiment. American Economic Review, 99(4), pp.1278-1308.
Bărbulescu, O. (2014). Ordonanţa de Urgenţă privind migraţia primarilor a fost adoptată de Guvern.
Mediafax. [online] Available at
httpswww.mediafax.ropoliticordonanta-de-urgenta-privind-migratia-primarilor-a-fost-adoptata-de-g
uvern-13160000 [Accessed 1 Jan. 2020].
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 61
Bauhr, M. and Grimes, M. (2013). Indignation or Resignation The Implications of Transparency for
Societal Accountability. Governance, 27(2), pp.291-320.
Bauhr, M., Czibik, Á., Fine Licht, J. and Fazekas, M. (2019). Lights on the shadows of public procurement
Transparency as an antidote to corruption. Governance, pp.1-29.
Bayley, D. (1966). The Effects of Corruption in a Developing Nation. The Western Political Quarterly,
19(4).
BBC (2015). Romania PM Ponta resigns over Bucharest nightclub fire. 4 November, [online] Available at:
https://www.bbc.com/news/world-europe-34720183 [Accessed 17 Dec. 2019].
BBC (2019). Romania corruption: PSD chief Liviu Dragnea jailed. [online] Available at:
https://www.bbc.com/news/world-europe-48421430 [Accessed 26 Dec. 2019].
Befu, D. (2015). Mama lui Ponta: ”Suntem o mică firmă de familie. Nu avem nimic de ascuns”. România
Curată vă prezintă principalii clienți ai firmei familiei lui Ponta, la care au ajuns azi procurorii DNA.
România Curată. [online] Available at:
http://www.romaniacurata.ro/mama-lui-ponta-suntem-o-mica-firma-de-familie-nu-avem-nimic-de-
ascuns-romania-curata-va-prezinta-principalii-clienti-ai-firmei-familiei-lui-ponta-la-care-au-ajuns-a
zi-procurorii/ [Accessed 19 Dec. 2019].
Blackburn, K., Bose, N. and Emranul Haque, M. (2010). Endogenous corruption in economic
development. Journal of Economic Studies, 37(1), pp.4-25.
Boas, T., Hidalgo, F. and Richardson, N. (2014). The Spoils of Victory: Campaign Donations and
Government Contracts in Brazil. The Journal of Politics, 76(2), pp.415-429.
Bogdan, S. (2018). De ce, în România, ne urâm de moarte pe tema anticorupției. Vice. [online] Available
at:
https://www.vice.com/ro/article/gy3g33/de-ce-in-romania-ne-uram-de-moarte-pe-tema-anticorupti
ei [Accessed 25 Dec. 2019].
Brogaard, J., Denes, M. and Duchin, R. (2015). Political Connections, Incentives and Innovation:
Evidence from Contract-Level Data. Working paper.
Bromberg, D. (2014). Can Vendors Buy Influence? The Relationship Between Campaign Contributions
and Government Contracts. International Journal of Public Administration, 37(9), pp.556-567.
Brugués, F., Brugués, J., & Giambra, S. (2018). Political connections and misallocation of procurement
contracts: evidence from Ecuador. Caracas: CAF.
Bryman, A. (2016). Social Research Methods - 5th Edition. 5th ed. Oxford: Oxford University Press.
Canayaz, M., Martinez, J. and Ozsoylev, H. (2015). Is the revolving door of Washington a back door to
excess corporate returns?. Working paper 1507
Center for Insights in Survey Research (2018). Public Opinion in Romania. [online] International
Republican Institute. Available at:
https://www.iri.org/sites/default/files/final_romania_poll_presentation.pdf [Accessed 25 Dec.
2019].
KTH ROYAL INSTITUTE OF TECHNOLOGY
62 Daniel Pustan
Centrul pentru Integritate în Business (2013). Mandatele politice. [ebook] Available at:
https://www.businessintegrity.ro/sites/default/files/fileuploads/newsletter/91%20-%20Mandatele%
20functiilor%20publice.pdf [Accessed 22 Dec. 2019].
Charron, N., Dahlström, C., Fazekas, M. and Lapuente, V. (2017). Careers, Connections, and Corruption
Risks: Investigating the Impact of Bureaucratic Meritocracy on Public Procurement Processes. The
Journal of Politics, 79(1), pp.89-104.
Ciubotaru, B. (2013). Organizarea şi funcţionarea Guvernului României: legislație, doctrină și practică
politică. Iaşi: Lumen.
Coviello, D. and Gagliarducci, S. (2017). Tenure in Office and Public Procurement. American Economic
Journal: Economic Policy, 9(3), pp.59-105.
David, Daniel (2015), Psihologia poporului român, Editura Polirom, ISBN 978-973-46-5478-9
Dávid-Barrett, E. and Fazekas, M. (2019). Grand corruption and government change: an analysis of
partisan favoritism in public procurement. European Journal on Criminal Policy and Research.
de Maria, B. (2008). Neo-colonialism through measurement: a critique of the corruption perception
index. Critical perspectives on international business, 4(2/3), pp.184-202.
Di Tella, R. and Schargrodsky, E. (2003). The Role of Wages and Auditing during a Crackdown on
Corruption in the City of Buenos Aires. The Journal of Law and Economics, 46(1), pp.269-292.
Directorate-General for Regional and Urban Policy (European Commission) and PWC (2016).
Stock-taking of administrative capacity, systems and practices across the EU to ensure the compliance
and quality of public procurement involving European Structural and Investment (ESI) Funds.
[online] Luxembourg Publications Office of the European Union, pp.180-188. Available at
https://op.europa.euenpublication-detail-publication60bf8ebc-0204-11e6-b713-01aa75ed71a1langua
ge-en [Accessed 7 Dec. 2019].
Doroftei, I. M. (2016). Measuring Government Favouritism Objectively: The Case of Romanian Public
Construction Sector. European Journal on Criminal Policy and Research, 22(3), pp.399-413.
Doroftei, M., and Dimulescu, V. (2015). Romanian Public Procurement in the Construction Sector:
Corruption Risks and Particularistic Links, Deliverable D8.1.5 ANTICORRP. Bucharest: Romanian
Academic Society.
Escresa, L. and Picci, L. (2017). A New Cross-National Measure of Corruption. The World Bank Economic
Review, 31(1), pp.196–219.
Euractiv (2014). Public procurement corruption in Romania. [online] Euractiv.com. Available at:
https://euractiv.eu/wp-content/uploads/sites/2/special-report/euractiv_special_report_public_pro
curement_corruption_in_romania.pdf [Accessed 2 Dec. 2019].
European Commission (2014). EU Anti-corruption report. Report from the Commission to the Council
and the European Parliament. [online] Brussels: COM(2014) 38 final. Available at:
https://ec.europa.eu/home-affairs/sites/homeaffairs/files/e-library/documents/policies/organized-c
rime-and-human-trafficking/corruption/docs/acr_2014_en.pdf [Accessed 26 Dec. 2019].
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 63
European Commission (2017). On Progress in Romania under the Co-operation and Verification
Mechanism. Report from the Commission to the European Parliament and the Council. [online]
Brussels: Com(2017) 44 final. Available at:
https://ec.europa.eu/info/sites/info/files/com-2017-44_en_1.pdf [Accessed 19 Dec. 2019].
European Commission. (n.d.). Corruption. [online] Available at
https://ec.europa.euhome-affairswhat-we-dopoliciesorganized-crime-and-human-traffickingcorrupti
on_en [Accessed 27 Dec. 2019].
European Parliamentary Research Service (2018). Listing of tax havens by the EU. [online] European
Union. Available at:
https://www.europarl.europa.eu/cmsdata/147404/7%20-%2001%20EPRS-Briefing-621872-Listing-t
ax-havens-by-the-EU-FINAL.PDF [Accessed 27 Dec. 2019].
Fazekas, M. (2017) Red tape, bribery and government favouritism: evidence from Europe. Crime Law Soc
Change (2017) 68: 403.
Fazekas, M. and Kocsis, G. (2017). Uncovering High-Level Corruption: Cross-National Objective
Corruption Risk Indicators Using Public Procurement Data. British Journal of Political Science,
pp.1-10.
Fazekas, M., Cingolani, L. and Tóth, B. (2016), A Comprehensive Review of Objective Corruption Proxies
in Public Procurement: Risky Actors, Transactions, and Vehicles of Rent Extraction.
GTI-WP/2016:03, Budapest: Government Transparency Institute.
Fazekas, M., Tóth, I. and King, L. (2013). Corruption Manual for Beginners: 'Corruption Techniques' in
Public Procurement with Examples from Hungary. SSRN Electronic Journal.
Fazekas, M., Tóth, I. and King, L. (2016). An Objective Corruption Risk Index Using Public Procurement
Data. European Journal on Criminal Policy and Research, 22(3), pp.369-397.
Ferraz, C. and Finan, F. (2008). Exposing Corrupt Politicians: The Effects of Brazil's Publicly Released
Audits on Electoral Outcomes. Quarterly Journal of Economics, 123(2), pp.703-745.
Ferreira Rubio, D. and Andvig, E. (2019). Serious about sustainability? Get serious about corruption.
World Economic Forum. [online] Available at:
https://www.weforum.org/agenda/2019/09/serious-about-sustainability-get-serious-about-corrupti
on/ [Accessed 29 Dec. 2019].
Fisman, D., Fisman, R., Galef, J., Khurana, R. and Wang, Y. (2012). Estimating the Value of Connections
to Vice-President Cheney. The B.E. Journal of Economic Analysis & Policy, 12(3).
France24 (2018). Romania party leader denies corruption in trial over fake jobs. [online] Available at
https://www.france24.comen20180321-romania-party-leader-denies-corruption-trial-over-fake-jobs
[Accessed 29 Dec. 2019].
Gherghina, S., & Volintiru, C. (2017). A new model of clientelism: Political parties, public resources, and
private contributors. European Political Science Review, 9(1), 115-137.
Golden, M. and Picci, L. (2005). Proposal for a new measure of corruption, illustrated with Italian data.
Economics and Politics, 17(1), pp.37-75.
KTH ROYAL INSTITUTE OF TECHNOLOGY
64 Daniel Pustan
Goldman, E., Rocholl, J. and So, J. (2013). Politically Connected Boards of Directors and The Allocation of
Procurement Contracts. Review of Finance, 17(5), pp.1617-1648.
Graycar, A. (2015) Corruption: Classification and analysis. Policy and Society, 34(2), 87-96.
Gürakar, E. and Bircan, T. (2019). Political Connections and Public Procurement in Turkey. Crony
Capitalism in the Middle East, pp.236-262.
Hall, B. (2019). Romanian government on brink of collapse. Financial Times. [online] Available at:
https://www.ft.com/content/3a22e218-c8d6-11e9-a1f4-3669401ba76f [Accessed 29 Dec. 2019].
Hicken, A. (2011). Clientelism. Annual Review of Political Science, 14(1), pp.289-310.
Hill, R., Griffiths, W. and Lim, G. (2011). Principles of Econometrics. 4th ed. New York: Wiley.
Huntington, S. (1968). Political order in changing societies. New Haven, CT: Yale University Press.
Hyytinen, A., Lundberg, S. and Toivanen, O. (2018). Design of public procurement auctions: evidence
from cleaning contracts. The RAND Journal of Economics, 49(2), pp.398-426.
Islam, R. (2006). Does more transparency go along with better governance?. Economics and Politics,
18(2), pp.121-167.
Jackowicz, K., Kozłowski, Ł. and Mielcarz, P. (2014). Political connections and operational performance of
non-financial firms: New evidence from Poland. Emerging Markets Review, 20, pp.109-135.
Jain, A. K. (2001), Corruption: A Review. Journal of Economic Surveys, 15: 71-121.
Johnston, M. (1996), The search for definitions: the vitality of politics and the issue of corruption.
International Social Science Journal, 48: 321-335.
Kaufmann, D., Kraay, A. and Mastruzzi, M. (2011). The Worldwide Governance Indicators: Methodology
and Analytical Issues. Hague Journal on the Rule of Law, 3(02), pp.220-246.
Kenny, C. (2009). Measuring Corruption in Infrastructure: Evidence from Transition and Developing
Countries. The Journal of Development Studies, 45(3), pp.314-332.
Knack, S. (2007). Measuring Corruption: A Critique of Indicators in Eastern Europe and Central Asia.
Journal of Public Policy, 27(3), pp.255-291.
Kurer, O. (1993). Clientelism, Corruption, and the Allocation of Resources. Public Choice, 77(2),
pp.259-273.
Kurer, O. (2005). Corruption: An Alternative Approach to its Definition and Measurement. Political
Studies, 53(1), pp.222-239.
Kurtz, M. and Schrank, A. (2007). Growth and Governance: Models, Measures, and Mechanisms. The
Journal of Politics, 69(2), pp.538-554.
Lambsdorff, J., (2006). The validity and precision of subjective indicators (CPI). In: Sampford, C.,
Shacklock, A., Connors, C., Galtung, F. (Eds.), Measuring Corruption. Ashgate, Aldershot, pp. 81–99.
Lancaster, T. and Montinola, G. (2001). Comparative political corruption: Issues of operationalization and
measurement. Studies in Comparative International Development, 36(3), pp.3-28.
Leff, N. H. (1964). Economic Development Through Bureaucratic Corruption. American Behavioral
Scientist, 8(3), pp. 8–14.
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 65
Leys, C. (1965). What is The Problem About Corruption?. The Journal of Modern African Studies, 3(2),
pp.215-230.
Lindstedt, C. and Naurin, D. (2010). Transparency is not Enough: Making Transparency Effective in
Reducing Corruption. International Political Science Review, 31(3), pp.301-322.
Luechinger, S. and Moser, C. (2014). The value of the revolving door: Political appointees and the stock
market. Journal of Public Economics, 119, pp.93-107.
Lui, F. (1985). An Equilibrium Queuing Model of Bribery. Journal of Political Economy, 93(4),
pp.760-781.
Lukkarinen, J., Morild, K. and Jönsson, S. (2015). Facts and figures on public procurement in Sweden.
The Swedish Competition Authority’s report series 2015:9. [online] Stockholm: The Swedish
Competition Authority and the National Agency for Public Procurement. Available at:
http://www.konkurrensverket.se/globalassets/publikationer/rapporter/rapport_2015-9_english.pdf
[Accessed 27 Dec. 2019].
Marcovici, T. (2019). Firma cumnatului premierului Ponta abonată la contracte cu statul. Cotidianul.
[online] Available at:
https://www.cotidianul.ro/firma-cumnatului-premierului-ponta-abonata-la-contracte-cu-statul/
[Accessed 19 Dec. 2019].
Marinas, R. and Dragin, S. (2017). Tens of thousands rally in Romania against planned emergency
decrees. Reuters. [online] Available at:
https://www.reuters.com/article/us-romania-protests/tens-of-thousands-rally-in-romania-against-p
lanned-emergency-decrees-idUSKBN15D118 [Accessed 19 Dec. 2019].
Markandya, A. and Pearce, D. (1988). Natural environments and the social rate of discount. Project
Appraisal, 3(1), pp.2-12.
Mauro, P. (1995). Corruption and Growth. The Quarterly Journal of Economics, 110(3), pp.681-712.
Mediafax (2009). Guvernul Boc 2 a fost demis. [online] Available at:
https://www.mediafax.ro/politic/guvernul-boc-2-a-fost-demis-video-4985350 [Accessed 27 Dec.
2019].
Méon, P. and Sekkat, K. (2005). Does corruption grease or sand the wheels of growth? Public Choice,
122(1-2), pp.69-97.
Mihalache, D., (2013) Achizitii publice 2007-2015 (Contracte). Romanian Digital Agenda Agency. [data]
Available at: http://data.gov.ro/dataset/achizitii-publice-2007-2016-contracte6 [Accessed 17 Dec.
2019]
Ministry of Finance (2018). Date de identificare plătitori actualizate august 2018. [data] Available at:
http://data.gov.ro/dataset/informatii_fiscale_august_2018 [Accessed 17 Dec. 2019]
Mironov, M. and Zhuravskaya, E. (2012). Corruption in procurement and shadow campaign financing:
Evidence from Russia. MIMEO.
Mo, P. (2001). Corruption and Economic Growth. Journal of Comparative Economics, 29(1), pp.66-79.
Mungiu, A. (2006). Corruption: Diagnosis and Treatment. Journal of Democracy, 17(3), pp.86-99.
KTH ROYAL INSTITUTE OF TECHNOLOGY
66 Daniel Pustan
Mungiu-Pippidi, Alina (editor) 2015, Government Favouritism in Europe: Anticorruption report 3,
Barbara Budrich Publishers
National Anticorruption Directorate (2017). Raport privind activitatea desfășurată. [online] Bucharest.
Available at: https://www.pna.ro/bilant_activitate.xhtml?id=40 [Accessed 19 Dec. 2019].
OECD (2016). Preventing Corruption in Public Procurement. [online] Secretary-General of the OECD.
Available at: http://www.oecd.org/gov/ethics/Corruption-Public-Procurement-Brochure.pdf
[Accessed 26 Dec. 2019].
Olken, B. (2006). Corruption and the costs of redistribution: Micro evidence from Indonesia. Journal of
Public Economics, 90(4-5), pp.853-870.
Olken, B. (2007). Monitoring Corruption: Evidence from a Field Experiment in Indonesia. Journal of
Political Economy, 115(2), pp.200-249.
Olken, B. (2009). Corruption perceptions vs. corruption reality. Journal of Public Economics, 93(7-8),
pp.950-964.
Oman, C. and C. Arndt (2006), Uses and Abuses of Governance Indicators, Development Centre Studies,
OECD Publishing, Paris.
Oprea, A. (2017). Registrul Comerţului: În România există 1,4 milioane de firme, din care doar 1,2
milioane sunt active. Cele mai multe insolvenţe sau radieri se înregistrează în comerţ, construcţii şi
transporturi. News.ro. [online] Available at:
https://www.news.ro/economic/registrul-comertului-in-romania-exista-1-4-milioane-de-firme-din-c
are-doar-1-2-milioane-sunt-active-cele-mai-multe-insolvente-sau-radieri-se-inregistreaza-in-comert-
constructii-si-transporturi-1922400517302017051116986291 [Accessed 27 Dec. 2019].
Politica Românească (2011). Guvernul Emil Boc (2). [online] Available at:
http://www.politicaromaneasca.ro/guvernul_emil_boc_2-2638 [Accessed 22 Dec. 2019].
Reinikka, R. and Svensson, J. (2004). Local Capture: Evidence from a Central Government Transfer
Program in Uganda. The Quarterly Journal of Economics, 119(2), pp.679-705.
Romanian Agency for Digital Agenda (2018). Lansarea noului Sistem Electronic de Achizitii Publice.
[online] Available at:
https://www.aadr.ro/-lansarea-noului-sistem-electronic-de-achizitii-publice_227_0.html [Accessed
7 Dec. 2019].
Rosca, A. (2018). Daniel David despre psihologia poporului român. Ce ne definește, ca nație, și cum ne
putem schimba. [Blog] Andreea Rosca Playground. Available at
https://www.andreearosca.ro/daniel-david [Accessed 1 Dec. 2019].
Rose-Ackerman, R. (1997). The political economy of corruption. In: Elliott, K.A. (Ed.), Corruption and the
Global Economy, Institute for International Economics, Washington, DC, pp. 31–60.
Rose-Ackerman S. (2008) Corruption. In: Readings in Public Choice and Constitutional Political
Economy. Springer, Boston, MA.
Rothstein, B. and Teorell, J. (2008), What Is Quality of Government? A Theory of Impartial Government
Institutions. Governance, 21: 165-190.
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 67
Stephenson, M. (2014). Can Foreign Anti-Bribery Enforcement Statistics Help Us Measure Corruption
Levels Objectively?. [Blog] The Global Anticorruption Blog. Available at:
https://globalanticorruptionblog.com/2014/04/15/can-foreign-anti-bribery-enforcement-statistics-h
elp-us-measure-corruption-levels-objectively/ [Accessed 29 Nov. 2019].
Stokes, S. (2005). Perverse Accountability: A Formal Model of Machine Politics with Evidence from
Argentina. American Political Science Review, 99(3), pp.315-325.
Svensson, J. (2005). Eight Questions about Corruption. Journal of Economic Perspectives, 19(3),
pp.19-42.
Schwab, K. (2016). The Global Competitiveness Report 2016–2017. [online] Geneva: the World Economic
Forum. Available at:
http://www3.weforum.org/docs/GCR2016-2017/05FullReport/TheGlobalCompetitivenessReport201
6-2017_FINAL.pdf [Accessed 29 Dec. 2019].
Søreide, T. (2002) Corruption in public procurement: Causes, consequences and cures, Chr. Michelsen
Institute Development Studies and Human Rights
Tanzi V., Davoodi H. (1998) Corruption, Public Investment, and Growth. In: Shibata H., Ihori T. (eds) The
Welfare State, Public Investment, and Growth. Springer, Tokyo
Timu, A. and Zega, R. (2017). Romanian Premier Refuses to Quit as Protests Grow Despite U-Turn.
Bloomberg. [online] Available at:
https://www.bloomberg.com/news/articles/2017-02-04/romanian-government-to-scrap-anti-graft-d
ecree-amid-protests [Accessed 17 Dec. 2019].
Titl, V. and Geys, B. (2019). Political donations and the allocation of public procurement contracts.
European Economic Review, 111, pp.443-458.
Toma, V. (2016). CCR despre abuzul in serviciu: "în mod defectuos" inseamna "incălcare a legii." Revista
22. [online] Available at:
https://revista22.ro/actualitate-interna/ccr-despre-abuzul-in-serviciu-n-mod-defectuos-inseamna-in
clcare-a-legii [Accessed 20 Dec. 2019].
Tran, M. (2015). Bucharest nightclub fire: PM and government resign after protests. The Guardian.
[online] Available at:
https://www.occrp.org/en/daily/4567-romania-prime-minister-ponta-resigns-amid-protests-over-ni
ghtclub-fire [Accessed 22 Dec. 2019].
Weber Abramo, C. (2008). How Much Do Perceptions of Corruption Really Tell Us?. Economics: The
Open-Access, Open-Assessment E-Journal, 2(2008-3), p.1.
Werner, S. (1983). New Directions in the Study of Administrative Corruption. Public Administration
Review, 43(2), p.146-154.
Xu, X. (2016). Corruption and economic growth: an absolute obstacle or some efficient grease?. Economic
and Political Studies, 4(1), pp.85-100.
KTH ROYAL INSTITUTE OF TECHNOLOGY
68 Daniel Pustan
Annexes
Annex A
Table 9 Summary of selected studies using objective indicators of corruption
Source: Fazekas, Tóth and King 2016, p. 2
paper indicator used Country year sector
Auriol, Straub and
Flochel (2016)
Exceptional procedure type Paraguay 2004-200
7
general
procurement
Bandiera, Prat,
and Valletti (2009)
Price differentials for standard goods
purchased locally or through a
national procurement agency
Italy 2000-200
5
various
standardized
goods (e.g.
paper)
Coviello and
Gagliarducci
(2017)
Number of bidders
Same firm awarded contracts
recurrently
Level of competition
Italy 2000-200
5
general
procurement
Di Tella and
Schargrodsky
(2003)
Difference in prices of standardized
products such as ethyl alcohol
Brazil 1996-1997 health care
Ferraz and Finan
(2008)
Corruption uncovered by federal
audits of local government finances
Brazil 2003 federal-local
transfers
Golden and Picci
(2005)
Ratio of physical stock of
infrastructure to cumulative spending
on infrastructure
Italy 1997 infrastructure
Goldman, Rocholl
and So (2013)
Political office holders' position on
company boards
USA 1990-2004 general
procurement
Hyytinen,
Lundberg and
Toivanen (2018)
Number and type of invited firms
Use of restricted procedure
Sweden 1990-1998 cleaning
services
Olken (2006) Difference between the quantity of
in-kind benefits (rice) received
according to official records and
reported survey evidence
Indonesia 1998-1999 welfare
spending
Olken (2007) Differences between the officially
reported and independently audited
prices and quantities of road
construction projects
Indonesia 2003-200
4
infrastructure
(roads)
Klasnja (2015) Single bidder procedures
Non-open procedure types
Romania 2008-2012 general
procurement
Reinikka and
Svensson (2004)
Difference between block grants
received by schools according to
official records and user survey
Uganda 1991-1995 education
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 69
Annex B
Table 10 Corruption techniques and indicators
Source: Fazekas, Tóth and King 2013, p. 63
No. Name Direct indicator Indirect indicator
1 Defining
unnecessary
needs
- -
2 Defining needs to
benefit a
particular
supplier
A) Prevalence of avoiding centralised
procurement
3 Tinkering with
thresholds and
exceptions
A) Proportion of non-open
procedures
D) Contract value according to Public
Procurement Law/total procurement
contract value
B) Average corruption risk score of
procedures followed
C) Frequency of actual contract value
above estimated contract value
4 Tailoring
eligibility criteria
A) Length of eligibility criteria -
5 Abusing formal
and
administrative
requirements
A) Length of eligibility criteria -
B) Proportion of excluded bids
6 Tailoring
evaluation criteria
A) Length of evaluation criteria -
B) Weight of non-price criteria
7 Using long term
complex contracts
A) Combined value of framework
contracts and PPPs / total contract
value
-
B) Average contract duration
8 Tinkering with
submission period
A) Proportion of tenders with
accelerated submission periods
-
B) Proportion of tenders with
extremely short submission periods
C) Average contract value per
weekday available for submission
9 Selective
information
provision
- A) Proportion of tenders with extremely
short submission periods
B) Proportion of procedures with call for
tenders modified within all procedures
10 Avoiding
publication of call
A) Proportion of tenders without call
for tenders in the official journal
-
KTH ROYAL INSTITUTE OF TECHNOLOGY
70 Daniel Pustan
for tenders
11 Strategically
modifying call for
tenders
A) Proportion of procedures with call
for tenders modified within all
procedures
-
12 Excessively pricey
documents,
difficult access to
documents
A) Price of documentation/estimated
contract value
-
13 Deliberate errors
in document
publication
A) Prevalence of extremely erroneous
contract award announcements
-
B) Hiding or erroneously reporting
the final contract value
14 Strategically
annulling
procedures
A) Proportion of annulled procedures
re-launched subsequently
B) Decrease in the number of bids
received in subsequent rounds
15 Repeated
violations of
public
procurement rules
A) Repeated court rulings against the
issuer within the same procedure
-
16 Unfair scoring - A) Average contract value per weekday
available for decision
B) Length of evaluation criteria
C) Weight of non-price criteria
17 Abusing unit
prices in the
contract
A) Proportion of contracts using unit
prices
-
18 Modifying
contracts
strategically
A) Proportion of modified contracts -
B) Difference between the awarded
and final contract value
C) Difference between originally
planned and final completion period
19 Abusing add-on
contracts
A) Proportion of add-on contracts -
B) Proportion of contracts exhausting
the planned reserves
20 Performance
violating contract
- -
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 71
Annex C
Table 11 Summary of selected studies on firm’s political connections
Paper Country Sample
period
Political power
level
Proxy for political
connection
Public procurement
Mironov and
Zhuravskayay (2012)
Russia 6 years regional elections illegal donations to parties
Goldman, Rocholl and So
(2013)
USA 1990-2004 House and Senate
(Parliamentary)
revolving door
Bromberg (2014) USA 2001-2006 Federal
government /
House of
Representative
donation
Straub (2014) Paraguay 2004-2011 government /
parliamentary
*member or sympathisant
Boas, Hidalgo and
Richardson (2014)
Brasil 2004-2010 Chamber of
Deputies
donation
Luechinger and Moser
(2014)
USA 1993-2003 Department of
defense
revolving door
Brogaard, Denes and
Duchin (2015)
USA 2000-2012 federal
government
donation
Doroftei(2016) Romania 2007-2013 donation + connections
Doroftei and Dimulescu
(2015)
Romania 2007-2013 donation + connections
Auriol, Straub and Flochel
(2016)
Paraguay 2004-2007
Titl and Geys (2019) Czech
Republic
2007-2014 13 regional
governments
donation
Arvate, Barbosa and
Fuzitani (2018)
Brasil 2007-2010 state legislatures donation
Brugués F., Brugués J. and
Giambra S. (2018)
Ecuador 2006-2018 bureaucrat (incl.
mayors and local
counselors)
revolving door
Johnson and Roer (2018) USA 2007-2016 House
Appropriations
subcommittee
"relocation" and "selection"
influence
Ferris, Houston and
Javakhadze (2019)
USA 2006-2013 donation
Gürakar and Bircan (2019) Turkey 2004-2011 Parliament and
local
governmental
bodies
revolving door
KTH ROYAL INSTITUTE OF TECHNOLOGY
72 Daniel Pustan
Stock market
Roberts (1990) USA 1983-09-01 congressional
seniority
donation
Fisman (2001) Indonesia 1995-1997 President Suharto Suharto Dependency Index
Faccio (2006) 47
countries
1991 see the proxy for
political
connection
revolving door
Jayachandran (2006) USA 2001-05 donation
Cooper, Gulen and
Ovtchinnikov (2010)
USA 1979-2004 House and Senate donation
Fisman et al. (2012) USA 2000-01-03 -
2001-04-30
revolving door
Amore and Bennedsen
(2013)
Denmark 2002-2008 local municipality family ties
Gropper, Jahera Jr. and
Park (2013)
USA 1989-2010 Congress banks' headquartered in the
same state with the
committee chair
Jackowicz, Kozłowski and
Mielcarz (2014)
Poland 2001-2011 central
government /
national
parliament / local
authorities
revolving door
Coulomb and Sangnier
(2014)
France 2007-01-01 -
2007-05-06
presidential Sarkozy Index (media)
Luechinger and Moser
(2014)
USA 1993-2003 Department of
defense
revolving door
Akey (2015) USA 1998-2010 Senators and
Representatives
donation
Canayaz, Martinez and
Ozsoylev (2015)
USA 1990-2012 government
positions
revolving door
Financing
Khawaja and Mian (2005) Pakistan 1996-2002 national elections if firm's director participates
in an election
Fraser, Zhang and
Derashid (2006)
Malaysia 1990-1999 government political patronage +
informal ties
Faccio, Masulis and
McConnell (2006)
35
countries
1997-2002 revolving door+friendship
Claessens, Feijend and
Laeven (2008) Brasil 1998 & 2002
federal deputy
candidates
donations - contribution to
campaigns
Bliss and Gul (2012) Malaysia 2001-2004 government informal ties
KTH ROYAL INSTITUTE OF TECHNOLOGY
Government favoritism in public procurement: evidence from Romania 73
Blau, Brough and Thomas
(2013)
USA 2008 federal
government
revolving door + lobbying
expenditures
Yeh, Shu and Chiu (2013) Taiwan 1998-2006 revolving door
Infante and Piazza (2014) Italy 2005-09 -
2009-03
national
parliament and all
local councils
revolving door
Houston, Maslar, and
Pukthuanthong (2018)
USA 2004-2012 White House,
Senate and
Congress
donations
Disaster relief
Nikolova and Marinov
(2017)
Bulgaria 2004-2005
Annex D
Table 12 Descriptive statistics
year mean sd min max p50 sum
2009 97276.44 1777334 0 2.79e+08 3751.26 1.19e+10
2010 87914.86 1308800 0 1.78e+08 3610.09 1.09e+10
2011 157587.2 2387378 0 2.61e+08 6346.42 1.66e+10
2012 141756.1 2609191 0 3.07e+08 5766.765 1.53e+10
2013 215950.2 3063701 0 2.91e+08 5611.643 1.44e+10
2014 239164.9 2658020 0 2.30e+08 6515.018 1.76e+10
2015 247728.3 3747858 0 7.76e+08 7246.912 1.73e+10
Total 155311.2 2463122 0 7.76e+08 5093.455 1.04e+11
KTH ROYAL INSTITUTE OF TECHNOLOGY