GOVERNANCE GOVERNANCE EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT Evidence on Public Procurement from Firm-Level Surveys Global Statistics from the World Bank Enterprise Surveys and a Novel Public Procurement Survey Module © 2023 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy, completeness, or currency of the data included in this work and does not assume responsibility for any errors, omissions, or discrepancies in the information, or liability with respect to the use of or failure to use the information, methods, processes, or conclusions set forth. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be construed or considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Attribution: Please cite this brief as follows: World Bank. 2023. “Evidence on public procurement from firm-level surveys Global statistics from the World Bank Enterprise Surveys and a novel public procurement survey module” Equitable Growth, Finance & Institutions Insight. Washington, DC: World Bank. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202- 522- 2625; e-mail: pubrights@worldbank.org. >>> Acknowledgments This paper is jointly prepared by the EFI Governance and the Enterprise Survey Unit. Serena Cocciolo (Economist, EPSPA), Domenico Viganola (Economist, DECEA), and Ruggero Doino (Research Analyst, DIME3) authored this paper under the guidance of Adenike Sherifat Oyeyiola (Practice Manager, EGVPF), Jorge Luis Rodriguez Meza (Practice Manager, DECEA), Snezana Mitrovic (Adviser, EGVPF), and Tracey Lane (Practice Manager, EAWG2). The paper uses the World Bank Enterprise Surveys data, in particular the standard surveys and a procurement module piloted in Croatia, Poland, and Romania in May–June 2021. The procurement module was developed jointly by the World Bank Enterprise Analysis Unit and the World Bank Governance Global Practice—in particular, by Serena Cocciolo, Domenico Viganola, and Sushmita Samaddar—with valuable inputs from Elena Carmen Calin, Filip Jolevski, Nona Karalashvili, and Silvia Muzi. The paper also benefited from the inputs received from the participants of the Enterprise Survey Unit seminar and from peer reviewers Manjola Malo (Senior Procurement Specialist, ELCRU), Carmen Calin (Procurement Specialist, EECRU), and Arvind Jain (Senior Economist, DECEA). >>> Contents Executive Summary 8 1. Introduction 9 2. WBES Data and Sample 12 3. Design of Firm-Level Surveys to Generate Evidence on 14 Public Procurement 4. Evidence on Firms’ Participation in Public Procurement 16 4.1. Why firm-level survey data is important 16 4.2. Evidence from the WBES 17 5. Evidence on barriers to and challenges of participation 25 in public procurement from the perspective of firms 5.1. Why Firm-Level Survey Data is Important 25 5.2. Evidence from the WBES 26 6. Evidence on the Experience of Bidders and Suppliers 32 6.1. Why Firm-Level Survey Data is Important 32 6.2. Evidence from the WBES 33 7. Conclusions 36 References 38 Annex A: Public Procurement Module from the Enterprise 40 Survey Questionnaire Annex B: Survey Implementation Methodology 43 Annex C: Definition of Key Variables and Empirical 46 Analysis Annex D: Additional Analysis and Robustness Checks 48 Annex E: Sample Description 67 Annex F: Institutional Background for Croatia, Poland, and 72 Romania >>> Figures Figure 1. Countries with One or Multiple WBES Wave 13 Figure 2. Survey Design Decisions 15 Figure 3. Country-Level Participation Rates in Public 18 Procurement Figure 4. Participation Rates in Public Procurement 18 Across Regions Figure 5. Country-Level Gaps in Participation Rates in 20 Public Procurement between SMEs and Larger Firms Figure 6. Country-Level Trends in Participation Rates in 21 Public Procurement Figure 7. Regional Trends in Participation Rates in Public 22 Procurement Figure 8. Perception of Obstacles (Croatia) 27 Figure 9. Perception of Obstacles (Poland) 27 Figure 10. Perception of Obstacles (Romania) 28 Figure 11. Main Administrative Obstacle before Contract 29 Signature Figure 12. Main Reason for Lack of Trust in the Process 29 Figure 13. Main Geographical Barrier to Securing a 30 Government Contract Figure 14. Biggest Challenge Expected When Working 30 Under a Government Contract Figure 15. Country-Level Rate of Firms Expected to Give 34 Gifts to Secure a Government Contract Figure 16. Rate of Firms Expected to Give Gifts to Secure 35 a Government Contract Across Regions >>> Tables Table 1: Share of Countries with Public Procurement Law 23 Requirement for Awards to SMEs Table 2: Trends in Participation Rates in Public Procurement between 2013 and 2018, by Firm Type 24 (Kenya) Table 3: Share of Firms that Received any Government Contracts Under Affirmative Platform, by Firm Type 24 (Kenya) Table 4: Experience of Bribery and Lack of Trust in the 31 Process >>> Abbreviations WBES World Bank Enterprise Survey eGP E-Government Procurement GPP Green Public Procurement GPPD Global Public Procurement Database PPA Public Procurement Authority SMEs Small and Medium-sized Enterprises EAP East Asia and the Pacific ECA Europe and Central Asia LAC Latin America and the Caribbean MENA Middle East and North Africa SA South Asia SSA Sub-Saharan Africa >>> Executive Summary Public procurement is at the intersection between the public and the private sectors. On one hand, relying on the private sector for the provision of certain goods and services can enhance efficiency; on the other public procurement can be a strategic tool to sustain firms’ growth both in the short and in the long term. Policy makers and practitioners are increasingly paying attention to the potential catalytic role of public procurement to promote economic growth and inclusive and sustainable development, for example through participation of SMEs and women-owned firms in this market. To sustain evidence-based policies in public procurement, firm-level survey data can be a valuable source of information on public procurement market. In particular, it can capture dimensions such as views and perceptions of firms that cannot be observed from e-government procurement data, it allows to study firms that never entered the public procurement market, and it provides data for countries that have not adopted an eGP system yet. First, this paper provides guidance on how to design a high-quality firm-level survey to study public procurement from the perspective of firms. The first set of decisions is on the target population: whether all firms in the private sector should be invited to participate in the survey, or only firms with some previous experience in public procurement. The second set of decisions is on the sampling design, which is critical to ensure that the (weighted) statistics from the firm-level survey data, such as the percentage of firms participating in public procurement, are accurate estimates of the real conditions of the private sector. The third set of decisions is on the mode of data collection, which should be informed by the specific country context and the cost implications. Second, this paper presents some of the statistics and stylized facts that can be generated on public procurement from the existing World Bank Enterprise Surveys data, covering more than 150 countries worldwide. The participation rate of private firms in public procurement is 18 percent, and small firms are 12 percentage points less likely to participate in public procurement than larger firms. Globally, 21 percent of firms that secured or attempted to secure at least one government contract in the 12 months prior to the interview state that firms similar to theirs typically make informal payments or give gifts to secure government contracts. A procurement questionnaire module piloted in Croatia, Poland, and Romania in the period May–June 2021 shows that, in these countries, administrative barriers before contract signature are the biggest obstacle to attempting to secure a government contract, in particular the amount of effort required for bid preparation and the length of the process after bid submission. Together with legislative and institutional reviews, and the analysis of transactional procurement data, firm-level survey data can be used to identify weaknesses of a public procurement system and inform reform efforts. The World Bank Enterprise Surveys microdata are publicly available for research and analytical studies, and the World Bank is continuing expanding its effort to generate global data on public procurement. 1. >>> Introduction Public procurement is a sizeable economic activity, and there is increasing attention on how it can strategically contribute to equitable and sustainable economic growth. Govern- ment procurement spending represents around 15 percent of global GDP (World Bank, 2022b,1 suggesting that improving the efficiency, effectiveness, and integrity of public procurement can generate sizeable savings and create additional fiscal space. Beyond fiscal savings, policy mak- ers and practitioners are increasingly paying attention to the potential catalytic role of public procurement to promote economic growth and inclusive and sustainable development (World Bank, 2022b; World Bank, 2023a). By nature, public procurement is at the intersection between the public and the private sector, and therefore it can be a strategic tool to sustain firms’ growth. For example, in Brazil,2 Ferraz, Finan, and Szerman (2015) demonstrate that public procurement can be an engine for firm growth both in the short and in the long term.3 Despite a growing aca- demic literature, there is still limited evidence on the link between public procurement and firms, which this paper contributes to address in two ways. First, this paper describes how firm-level survey data can be used to generate evidence on firms’ participation in public procurement, and on the perceived barriers and challeng- es when operating in this market. With the increasingly widespread adoption of e-government procurement (eGP),4 an increasing amount of administrative data is available on procurement transactions, and these data sources are typically used for research and for data-driven public procurement diagnostics. However, by design, transactional data cannot capture dimensions such as views and perceptions of firms, and it cannot be used to characterize firms that never entered the public procurement market. Firm-level surveys can be a valuable complementary data source for public procurement diagnostics,5 especially for countries that have not adopted an eGP system yet, and where data collection of administrative data from paper records would represent a costly endeavor. This paper provides guidance to teams interested in using firm- level survey data to generate evidence on public procurement from the perspective of firms. 1. “Public procurement” refers to government spending on goods, services, and works. The procurement and contracting cycle refers to the sequence of related activities, from needs assessment, through competition and award, to payment and contract management, as well as any subsequent monitoring or auditing (OECD, 2021). 2. Similar results are found by Cappelletti, Giuffrida, and Rovigatti (2022) in the context of Italy. Firms that win government contracts are more likely to stay in the market than marginal losers after the award. This boost is associated with improvements in their credit position and earning substitution, and it lasts longer than the contract duration. 3. Ferraz, Finan, and Szerman (2015) show that firms winning at least one government contract increase firm employee growth in that quarter, and in the future, are more likely to participate and win auctions and expand their business in new markets. 4. The World Bank Global Public Procurement Database (GPPD) provides information on eGP adoption and level of maturity of the eGP system across countries globally. As of 2020, 159 countries (74 percent) were reported to have an eGP system, although with varying maturities and functionalities. 5. World Bank (2022a) provides detailed guidance on using procurement transactions data. This evidence, when complemented by legislative and insti- This paper contributes to the literature on the drivers of tutional reviews, and by the analysis of transactional procure- firms’ participation in public procurement. Firms participat- ment data, can be used to identify weaknesses of a public ing in public procurement represent a particular segment of procurement system and inform reform efforts.6 Public pro- the private sector. For example, it is widely believed that al- curement diagnostics based on firm-level survey data can be location of government contracts is skewed toward large firms used as part of standalone public procurement assessments, (Hoekman and Onur Taş, 2022), and in Sub-Sharan Africa, MAPS assessments, or core World Bank diagnostics.7 foreign-owned firms often account for a large share of total public procurement (Hoekman and Sanfilippo, 2020). Given Second, this paper uses the World Bank Enterprise Sur- that public procurement can represent a meaningful source of veys (WBES) data to demonstrate the evidence that can demand for firms, and a stimulus for investments and innova- be generated from firm-level survey data on public pro- tions, policy makers and practitioners are increasingly consid- curement. In doing so, it also provides novel global statis- ering policies to make public procurement more competitive tics and stylized facts on the link between public procure- and accessible for firms , such as access to finance, training, ment and firms. The Enterprise Surveys collect a wide range and information on public procurement (Hjort, Iyer, and de Ro- of information through interviews with formal firms in 151 chambeau, 2020), transparency of the procurement system countries,8 including on participation in public procurement for (Knack, Biletska, and Kacker, 2017), timeliness of payments 150 countries,9 making it possible to investigate a few char- (Barrot and Nanda, 2020; Conti et al., 2021), or slicing bigger acteristics of the public procurement market across a large contracts into smaller ones (di Giovanni et al., 2022).13 This sample of countries.10 This paper uses the WBES data to pro- paper contributes to this literature by analyzing the participa- vide an overview of firms’ participation in public procurement tion of firms in public procurement globally, and the barriers at the global level, exploring differences in participation rates and challenges perceived by firms in Croatia, Poland, and between different types of firms (e.g., small versus large firms) Romania when participating in public procurement or working and sectors, and the drivers of participation. Additionally, it ex- under a government contract. ploits a new procurement module piloted in Croatia, Poland, and Romania in 2021.11 This data is used to demonstrate how The WBES data indicates that the participation rate firm-level survey data can generate valuable and actionable of private firms in public procurement is 18 percent, information on the barriers and challenges perceived by firms and small firms are 12 percentage points less likely to when participating in public procurement or working under a participate in public procurement than larger firms. For a government contract. Beyond demonstrating the potential of subsample of countries with Enterprise Surveys data available using firm-level data for public procurement diagnostics, the over multiple years, the trends show that participation rates analysis presented in this paper provides novel stylized facts in public procurement have not been increasing over time. In and global statistics on public procurement. These stylized Croatia, Poland, and Romania, the majority of firms indicate facts and statistics are not intended to provide policy recom- that administrative barriers before contract signature are mendations to specific countries or globally, but they can be the biggest obstacle to attempting to secure a government used as starting point for further assessments and to stimulate contract, policy dialogue.12 6. A comprehensive methodology for assessing public procurement systems (MAPS) has already been developed and implemented in more than 45 countries, and World Bank (2022a) provides a framework for the application of data analytics in public procurement. Using firm-level survey data to analyze public procurement should be considered as complementary to these existing approaches, and it is intended to contribute to expanding the evidence base for procurement reforms. 7. For example, the Croatia Country Economic Memorandum 2022 (World Bank, 2023b) and the “Supporting the Efficiency of the Romanian Public Procurement System” project (World Bank, 2022c) use firm-level survey with a focus on public procurement. 8. The Enterprise Surveys focus on firms registered in the country, regardless whether the ownership is national or international. 9. All 159 firms interviewed in Guinea-Bissau did not respond to the question about their participation in public procurement. 10. Further details on the Enterprise Surveys are provided in Annex B.1. 11. The module was developed jointly by the World Bank Enterprise Survey Unit and the World Bank Governance Global Practice, in particular, by Serena Cocciolo, Dome- nico Viganola, and Sushmita Samaddar, with inputs from Elena Carmen Calin, Filip Jolevski, and Nona Karalashvili. The opportunity of piloting this new procurement module in Croatia, Poland, and Romania emerged from the ongoing engagements with government counterparts in the area of public procurement, and the timing of the implementation of the COVID-19 Enterprise Surveys. In particular, at the time of launching the first round of the COVID-19 Enterprise Surveys in May 2020, the World Bank was working on a public procurement assessment in Croatia, a public procurement assessment in Romania, and the preparation of a public procurement focused RAS in Poland. The data for Croatia was used in the Croatia Country Economic Memorandum 2022 (World Bank, 2023b). 12. The evidence presented in this paper is limited by the available WBES data, and it does not intend to provide policy recommendations to specific countries or globally. Rather, it can be used as a starting point for further assessments and to stimulate policy dialogues. To inform specific national reforms, it is recommended to adopt a problem-driven approach and complement the empirical findings from firm-level survey data with legislative and institutional reviews, and, when possible, by the analysis of transactional procurement data. An example is provided by the ongoing study on public procurement in Romania, conducted under the “Supporting the Efficiency of the Romanian Public Procurement System” project (World Bank, 2022c). 13. Even non-targeted policies might be distortionary. For example, as demonstrated by di Giovanni et al. (2022), slicing bigger contracts into smaller ones can help small firms to overcome financial constraints and participate in public procurement, but they can also reduce saving incentives for large firms. contract, citing in particular the amount of effort required for The paper is organized as follows: The next section de- bid preparation and the length of the process after bid submis- scribes the Enterprise Surveys data and the sample used to sion. Lack of trust in the process is another deterrent to par- generate the global statistics and stylized facts presented in ticipation in public procurement, as 21 percent of firms globally the remainder of the paper. Section 3 explains how to design state that they are expected to give gifts to secure government firm-level surveys to generate evidence on public procure- contracts. Overall, the paper demonstrates which empirical in- ment. Section 4 shows how firm-level survey data can be used sights on public procurement can be drawn from the analy- to estimate participation rates in public procurement in the pri- sis of firm-level surveys and provides additional resources for vate sector. Section 5 elaborates on the need to use firm-level teams interested in replicating this approach and data collec- survey data to investigate the barriers and challenges experi- tion (i.e., public procurement module in Annex A and survey enced by firms in the public procurement market, and Section implementation methodology in Annex B). The novel stylized 6 highlights what can be learned from firm-level surveys with facts and global statistics on public procurement presented in bidders or contracted firms. Each of these sections include this paper do not intend to generate policy recommendations reflections on why firm-level surveys data can be useful and to specific countries or globally. Rather, they are viewed as the necessary design considerations, a demonstration using starting points for further assessments and to stimulate policy the publicly available Enterprise Surveys data, and novel styl- dialogues. ized facts and global statistics. Section 7 concludes with some reflections on the research agenda on the link between public procurement and firms, and additional resources. 2. >>> WBES Data and Sample The estimates presented in Section 4 and Section 6 are derived from the World Bank En- terprise Surveys data, covering 174,297 firms from 150 countries in 301 survey waves.14 Firms’ participation rates in public procurement are estimated using the standard WBES waves where firms are asked whether they have secured or attempted to secure a government contract in the 12 months prior to the interview.15 For the analysis that uses only the last available survey wave for each country, the subsample covers 93,294 firms from 150 countries over the period 2009–2021 (see details in Annex E), of which 90,199 (97 percent) responded to the question on participation in public procurement. Table E.1 provides the year of the latest standard WBES wave for each country, and Table E.2 describes the main firm characteristics by region. The analysis of time trends (Section 4.2.4) is restricted to countries where multiple waves are avail- able, and the subsample covers 46,096 firms from 67 countries. Figure 1 depicts countries with one or multiple WBES waves.16 The statistics presented in Section 5 are derived from the World Bank Enterprise Surveys data collected for Croatia, Poland, and Romania in May–June 2021, which include a pro- curement module. The procurement module was developed jointly by the World Bank Enter- prise Surveys Unit and the World Bank Governance Global Practice, with the objective of better understanding the barriers and challenges experienced by firms with respect to public procure- ment. This module was implemented as part of the COVID-19 Enterprise Surveys in Croatia, Poland, and Romania and it is reported in Annex A.17 The COVID-19 WBES was designed to estimate the effects of the COVID-19 pandemic on the private sector, and therefore the sample was the panel of firms interviewed during the latest standard WBES wave for each country.18 The COVID-19 WBES sample for Croatia, Poland, and Romania covers, respectively, 324, 500, and 983 firms, of which only 22 in total (1 percent) did not respond to the questions in the pro- curement module. Table E.4 summarizes the main characteristics of firms interviewed during the COVID-19 surveys in Croatia, Poland, and Romania, showing that the COVID-19 WBES sample is quite similar to the 2019 sample for these countries. 14. This paper uses the Enterprise Surveys data as of October 2021. 15. See the Enterprise Survey standard questionnaire. In the Manufacturing module, the question on participation in public procurement is question J.6a. The Enterprise Surveys consider only national public procurement. 16. Further details on the Enterprise Surveys are provided in Annex B.1. 17. The COVID-19 Enterprise Surveys were implemented between May 2020 and August 2022 in 46 countries. See additional information in the World Bank Enterprise Surveys, COVID-19 webpage. 18. Muzi et al. (2022) provide an example of how the COVID-19 Enterprise Surveys can be used to estimate the exit rates of establishments from the economy. Another example is provided in the study by Hyland et al. (2021), which provides descriptive evidence on the differential effect of the COVID-19 crisis on female- and male-owned firms. > > > F I G U R E 1 - Countries with One or Multiple WBES Wave Source: World Bank Enterprise Survey. 3. >>> Design of Firm-Level Surveys to Generate Evidence on Public Procurement Firm-level surveys can be most informative if they are based on a robust design. Beyond the design of the questionnaire, the realization of high-quality firm-level surveys requires taking a series of design decisions, of which the primary ones are summarized in Figure 2. The first set of decisions is on the target population, which depends on the objective of the analysis and the type of policy decisions that the analysis intends to inform. This strictly depends on the objective of the analysis and the type of policy decisions that it intends to inform. In the case of public procurement, one of the main decisions is whether all firms in the private sector should be invited to participate in the survey, or only firms with some previous experience in public procurement. This decision has implications for the analytical insights that could be generated from the survey data. If the objective of the analysis is to generate evidence on the participation of firms in the public procurement market, then the target population should consist of all firms in the private sector. If the objective of the analysis is to focus on the behaviors and experiences of bidders or suppliers, then the target population should consist of firms with some recent experience in public procurement. Respectively, Section 4 and Section 6 further elaborate on these cases. The second set of decisions is on the sampling design. To be informative, firm-level surveys should be designed to be representative of the target population, so that the (weighted) statis- tics, such as the percentage of firms participating in public procurement, are accurate estimates of the real conditions of the private sector. The quality of survey data and analysis crucially depends on whether the survey design allows for the generation of a representative sample of firms and/or representative statistics, and therefore avoids the risk of capturing only selected and biased views from selected firms, which is a typical concern for qualitative studies and focus group discussions. Achieving a representative survey sample and/or representative statistics re- quires having data on (1) the size of the target population, and (2) how the population of interest is distributed over key firm characteristics (e.g., number of small firms in the capital region in the retail sector). For example, if the target population consists of firms with some recent experience in public procurement, then it is necessary to have data on bidders and suppliers (i.e., their total number and the distribution of key firm characteristics, such as the number of contracts won in recent years). Another element of sampling design is whether the survey should be based on a census or should employ a sampling approach. A census consists of inviting all the firms in the target population to participate in the survey, while a sampling option would be to consider only bidders and suppliers in the approach consists of inviting only a group of firms (selected target population, but no data exists (or no data is available to in order to ensure sample representativeness). This decision the research team) on their total number and key characteris- closely relates to the dissemination strategy, the mode of de- tics, then the choice of the target population could be recon- ployment of the survey, and the budget available to run the sidered. The survey design should be adjusted to the specific study. country context and objective of the analysis, ensuring coher- ence across decisions. For example, the Enterprise Surveys The third set of decisions is on the mode of data collec- consider all firms in an economy as their target population, tion. There is a variety of approaches to data collection. The follow a sampling approach, define the sampling strategy on questionnaire can be completed online, on the phone, or in the basis of official statistics on the private sector and official person. It can be completed autonomously by respondents or contact lists, contact firms directly, and administer the ques- with the assistance of an interviewer. Firms can be invited to tionnaire with interviewers on the phone or in person. As an- participate in the survey through a variety of ways, such as other example, for an ongoing study on public procurement through the Chamber of Commerce, through the eGP portal, in Romania (World Bank, 2022c), the target population is all or directly by the research team. The deployment strategy of firms registered on the eGP portal, statistics from official data the questionnaire should be informed by the specific country on the size and key characteristics of the target population are context and the cost implications. available, the survey uses a census approach, and firms re- ceive an invitation to participate in the survey through the eGP Decisions on the target population, on the sampling de- portal and autonomously complete the questionnaire online. sign, and on the mode of data collection are closely re- Further details on these two examples and the specific condi- lated and inform each other. These decisions are closely tions that enable them are provided in Annex B. interlinked and usually are not linear. For example, if the best > > > F I G U R E 2 - Survey Design Decisions Target population Sampling design Mode of data collection • Which information do we have • How can we reach all firms in • Which types of firms should be on firms in the target the target population to invite invited to participate in the population? them to the survey? survey? • Will the survey use a census or • How will the survey be sampling approach? implemented? Source: Authors’ elaboration. 4. >>> Evidence on Firms’ Participation in Public Procurement 4.1. Why Firm-Level Survey Data is Important Firms’ participation rate in public procurement represents the share of firms that are ac- tive in the public procurement market. Measures of competitiveness in public procurement typically consider indicators such as the number of bidders, the share of new bidders, the share of international bidders, and the winning rebate (World Bank, 2022a). While these indicators characterize the procurement process, the participation rate in public procurement considers all firms in an economy to represent the extent to which public procurement is a relevant market for the private sector, and it can be further disaggregated by sector or other firm characteristics (e.g., firm size). Participation rates in public procurement could be estimated from administrative micro- data, but only in specific circumstances which are currently met in only a small set of countries. Procurement transactions’ microdata only captures data on bidders and firms award- ed a contract, and therefore cannot be used to understand the extent to which firms participate in public procurement or the barriers and challenges preventing firms from entering this market. Participation rates can be estimated by connecting microdata on procurement transactions with firms’ registry data, such as tax registries, which can be done reliably only when both databases include firm-level identifiers. At the moment, only a few countries meet these requirements. Microdata on procurement transactions with good or complete coverage of all government con- tracts is typically available for countries with an eGP system. However, not all countries have implemented an eGP system yet, fewer capture official firm-level identifiers (for example, cor- responding to tax registries), and even fewer make administrative data (e.g., public procurement and tax registries) publicly available for research. Given these challenges in using administrative microdata, World Bank Enterprise Surveys data offers the additional firm-level representative surveys are viable alternatives advantage of comparability across countries, which allows for estimating participation rates. Firm-level surveys allow for the estimation of statistics at the global level and for for the collection of data from firms irrespective of their past comparisons across countries and over time.21 activities in the public procurement market, and therefore the inclusion of firms that have never been awarded a government contract as well as firms that have never submitted a bid for such. To be informative, firm-level surveys should be designed 4.2. Evidence from the WBES to be representative of the population of interest, so that the (weighted) statistics, such as the percentage of firms participating in public procurement, are accurate estimates 4.2.1. Estimation of Participation Rates in of the real conditions of the private sector. As described in Public Procurement Section 3, achieving a representative survey sample requires (1) data on the size of the population of interest (e.g., number Globally, the firms’ participation rate in public of firms in the economy); (2) data on how the population of procurement is 18 percent. Figure 3 and Table D.2 depict interest is distributed based on key firm characteristics (e.g., the distribution of participation rates in public procurement number of small firms in the capital region in the retail sector); across the world. The share of firms that secured or attempted and (3) the ability to contact the population of interest, or a to secure a government contract in the 12 months prior to the stratified sample of it. These requirements can be fulfilled interview ranges from 12 percent in South Asia to 22 percent through different sampling strategies, as elucidated by two in Sub-Saharan Africa (Figure 4). The participation rate of examples from the Enterprise Surveys and an ongoing study private firms in public procurement is below 20 percent in 60 on public procurement in Romania (World Bank, 2022c), percent of countries. Differences in participation rates across described in detail in Annex B. The next sections use the countries may depend on a variety of factors, such as the World Bank Enterprise Surveys data to demonstrate how public procurement regulatory framework; the existence of firm-level survey data can be used to generate empirical an eGP system;22 the structure of the economy; the sectoral insights on the links between the private sector and public composition of the private sector; the prevalence of SOEs procurement. Section 4.2.1 provides statistics on firms’ in the economy; the accessibility of the public procurement participation rates, estimated from a global representative market; access to finance, training, and information on public sample of firms.19 Section 4.2.2 explores differences in procurement (Hjort, Iyer, and de Rochambeau, 2020); the participation rates for different types of firms, focusing on transparency of the procurement system (Knack, Biletska, small and medium-sized enterprises (SMEs) versus larger and Kacker, 2017); the timeliness of payments (Barrot and firms.20 Section 4.2.3 studies trends in firms’ participation in Nanda, 2020; Conti et al., 2021); the availability of small value public procurement over time. Section 4.2.4 demonstrates contracts (di Giovanni et al., 2022); and trust towards the how firm-level survey data can be used to provide initial state. The procurement regulatory framework is only one of evidence on the impact of procurement policies or strategies, the factors influencing the propensity of firms’ participation in such as preference policies for specific groups of firms. The public procurement, and indeed, despite the common system World Bank Enterprise Surveys are designed to construct a of laws, regulations, and policies applicable to all European representative sample of firms at the national level, sector Union countries, a similar variation in procurement participation level, firm-size level, and regional level, thereby ensuring rates across countries is observed within the European Union the reliability of (weighted) statistics by firm characteristics and within other regions. in terms of those dimensions (see details in Annex B.1). The 19. The participation rate in public procurement is calculated from the WBES using question j6a in the standard Enterprise Survey questionnaire. The exact question in the standard Enterprise Survey questionnaire reads: “Over the last year, has this establishment secured or attempted to secure a government contract?”. The standard Enterprise Survey questionnaire is available at the World Bank Enterprise Surveys website. 20. Although Section 4.2.2 focuses on differences in participation rates across firm size, the participation of firms in public procurement is also influenced by other firm char- acteristics, such as the age of the firm, whether the firm is female- or male-run, and whether the firm is based in the capital city or not (Table D.6). 21. The Enterprise Surveys are implemented across the world using a common methodology and a common set of core questions, which enhances cross-country compa- rability of statistics and indicators. 22. In line with various findings from the literature, Table D.7 provides suggestive evidence that firms’ participation rate in public procurement is higher in countries with an eGP system in place, even after controlling for GDP per capita and other country-level characteristics. This analysis was based on the global indicators drawn from the World Bank GovTech Maturity Index (World Bank, 2022e). This analysis could not be extended to other potential factors explaining variations across countries, such as the public procurement regulatory framework, the prevalence of SOEs in the economy, or the accessibility to the public procurement market, because of lack of global indicators on these dimensions. > > > F I G U R E 3 - Country-Level Participation Rates in Public Procurement Source: World Bank Enterprise Survey. Note: N = 90,199. The intervals are computed using Jenks natural break. Statistics are computed using the latest ES wave available for each country and using sample weights (“Wt”). > > > F I G U R E 4 - Participation Rates in Public Procurement Across Regions Source: World Bank Enterprise Survey. Note: N = 90,199. Each value indicates the share of firms. Statistics are computed using the latest ES wave available for each country and using sample weights (“wt”). Participation rates in public procurement vary significantly SMEs and larger firms is in part attributed to the different across sectors.23 Contracting with the government is much sectors in which they operate. However, differences across more relevant for certain sectors (e.g., construction) than others sectors do not fully explain the differences in participation (e.g., tourism). When firm-level survey data are designed to rates of SMEs relative to larger firms. Even after taking into be representative at the sector level, it is possible to calculate account differences due to firm age, firm productivity, and the participation rates for each sector. Exploiting this feature of the main sector, small firms are on average 12 percentage points World Bank Enterprise Surveys, Table D.1 summarizes sector- less likely to participate in public procurement than larger firms specific participation rates in public procurement globally and (Table D.3).26 The results are robust to using a sector-specific across regions.24 Not surprisingly, the sectors with the largest definition of SMEs,27 and are consistent across sectors with participation rates in public procurement are construction and different structural characteristics, such as manufacturing of computer and related activities. There is no sector for which basic metals and provision of computer and related activities the participation rate is negligible at the global level, and for (Table D.5). At the country level, after taking into account 60 percent of sectors, at least 10 percent of firms in that sector differences due to age of the firm and the sector, SMEs are attempted to secure or secured a government contract at least statistically less likely to participate in public procurement than once in the 12 months prior to the interview (Table D.1). larger firms in 36 percent of countries (Figure 5).28 The lower participation of SMEs in public procurement may be explained by a variety of factors, such as differences in efficiency and 4.2.2. Estimation of Participation Rates of competitiveness, access to finance and credit, administrative SMEs in Public Procurement and bureaucratic barriers, knowledge of public procurement rules and procedures, delays in payments, and awareness of economic opportunities in the procurement market. Section 5 Relative to larger firms, SMEs participate significantly less demonstrate the application of a procurement module that can in public procurement. Table D.2 reports the participation be integrated into firm-level surveys to generate evidence on rates by firm size for each country and shows that they are the barriers and challenges experienced by firms in the public lower for SMEs than for larger firms for the large majority procurement market in Croatia, Poland, and Romania. of countries.25 The difference in participation rates between 23. In the WBES, firm sector is defined as the (self-reported) main sector of operations of the firm. 24. In this paper, sectors are defined using ISIC codes (Annex C). Unless specified otherwise, the two-digit ISIC Rev. 3.1 codes are used. Additional information on the International Standard Industrial Classification of All Economic Activities (ISIC) is available at the SICCODE website. 25. The participation rate in public procurement among small firms is larger than that of large firms in only 18 percent of countries (27 countries out of 148). 26. The regression analysis controls for firms’ age (log), whether the firm is women-managed, firm productivity, country fixed effects, and sector fixed effects (two-digit ISIC Rev. 3.1 codes). The baseline is given by “Large Enterprises” (omitted category). 27. In Table D.4, the definition of small, medium, and large enterprises is specific to each sector and based on the distribution in the number of employees for each sector: “small enterprises” are firms in the first tercile in distribution of number of employees within each sector, “medium enterprises” are firms in the second tercile, and “large enterprises” are firms in the third tercile. 28. The regression analysis at country level controls for firms’ age and sector fixed effects at aggregate level (manufacturing, retail services, and other services). > > > F I G U R E 5 - Country-Level Gaps in Participation Rates in Public Procurement between SMEs and Larger Firms Source: World Bank Enterprise Survey. Note: The map shows the statistical differences in public procurement participation of Small and Medium Enterprise compared to large firms. The results are computed using a Linear Probability Model and controlling for the log of the age of the firm and fixed sector effects (Retail services, Manufacturing and Others), with robust standard errors clustered at the sector and firm size level. The linear probability model is estimated for each country separately. Statistics are computed using the multiple ES waves available for each country and using sample weights (“wt”). 4.2.3. Evaluation of Trends in Participation Rates in Public Procurement Overall, participation rates in public procurement have from 19 percent to 18 percent.29 This trend varies significantly slightly dropped over the last decade. The Enterprise across countries, with countries split approximately equally Surveys are regularly implemented across countries, and 47 into those with increasing participation rates and those percent of countries have completed at least two standard with decreasing participation rates (see Figure 6 and Table WBES waves. This allows studying trends and changes over D.8). On average, over time the participation rate in public time in public procurement participation rates for a large sample procurement increased only in LAC and SSA countries, of countries. Between 2007–2013 and 2014–2020, the global whereas it decreased in the other regions (Figure 7). participation rate in public procurement decreased slightly 29. For this report we use the last and second last Enterprise Survey available for each country with multiple Enterprise Surveys conducted. The second last Enterprise Surveys available cover the following periods: 2015-2019 for EAP; 2018-2020 for EAC; 2016-2018 for LAC; 2013-2020 for MENA; 2013-2015 for SA; 2013-2020 for SSA. The last Enterprise Survey available cover the following periods: 2009-2014 for EAP; 2007-2014 for EAC; 2010 for LAC; 2010-2013 for MENA; 2007-2009 for SA; 2007-2011 for SSA. > > > F I G U R E 6 - Country-Level Trends in Participation Rates in Public Procurement Source: World Bank Enterprise Survey. Note: N = 33,853 (first wave). N = 58,339 (last wave). The graph shows the difference in procurement participation rates across countries, calculated for each country as the simple difference between participation rates in the second and first wave. Each country appears in each wave only once. First waves range from 2009 to 2014 for EAP, from 2007 to 2014 for EAC, 2010 for LAC, from 2010 to 2013 for MENA, from 2007 to 2009 for SA, and from 2007 to 2011 for SSA. Second waves range from 2015 to 2019 for EAP, from 2018 to 2020 for EAC, 2016 to 2018 for LAC, from 2013 to 2020 for MENA, from 2013 to 2015 for SA, and from 2013 to 2020 for SSA. Statistics are computed using the latest ES wave available for each country and using sample weights (“Wt”). > > > F I G U R E 7 - Regional Trends in Participation Rates of Private Firms in Public Procurement Source: World Bank Enterprise Survey. Note: N = 33,853 (first wave). N = 58,339 (last wave). First waves range from 2009 to 2014 for EAP, from 2007 to 2014 for EAC, 2010 for LAC, from 2010 to 2013 for MENA, from 2007 to 2009 for SA, and from 2007 to 2011 for SSA. Second waves range from 2015 to 2019 for EAP, from 2018 to 2020 for EAC, 2016 to 2018 for LAC, from 2013 to 2020 for MENA, from 2013 to 2015 for SA, and from 2013 to 2020 for SSA. Statistics are computed using the latest ES wave available for each country and using sample weights (“wt”). 4.2.4. Analysis of the Impact of Preference Policies on Firm Participation Policy makers and practitioners are increasingly paying Public Procurement Database,30 32 percent have some attention to the accessibility and inclusiveness of public provision in their procurement legal framework to incentivize procurement. When used strategically, public procurement SMEs’ participation in public procurement (Table 1). In can play a significant role in supporting national policy particular, countries have been adopting two types of policies. priorities and addressing local and global challenges, such as The first one consists of policies targeted to specific groups of in developing local industries, reducing income inequalities, firms, such as SMEs or domestic firms, for example through and creating local jobs (World Bank, 2022b). Among countries set-aside quotas or price preferences.31 for which information is available in the World Bank Global 30. World Bank Global Public Procurement Database features an extensive set of country-level information about public procurement practices. These statistics are based on the data field “Public Procurement Law requirement for awards to SMEs (small and medium-sized enterprises) clause.” 31. There are various tools that can be used to implement preference policies to specific groups of firms. One is to employ contract threshold preferences for SMEs, with contracts below a certain threshold value designated specifically for SMEs, as in the case of Brazil, Colombia, Indonesia, Malta, Russia, and Trinidad and Tobago. A sec- ond approach is the use of set-asides or quotas, where a reserved portion of the public procurement budget is set aside for certain types of firms, such as SMEs. Quotas may range from as high as 25 percent in Angola, Kenya, Brazil, and Liberia to as low as 10 percent in Argentina, Australia, Chad, Egypt, Philippines, and Uruguay (World Bank 2021a). A third way is based on monetary incentives for SMEs or other targeted groups, such as through waivers of participation fees (e.g., India and Honduras) or through preference margins that—for evaluation purposes only—apply a discount to the offered price (e.g., Bolivia). > > > T A B L E 1 - Share of Countries with Public Procurement Law Requirement for Awards to SMEs Share of countries with Public Procurement Region Law requirement for awards to SMEs clause EAP 24% ECA 47% LAC 26% MENA 20% SA 29% SSA 37% Source: World Bank. Global Public Procurement Database (accessed March 2022). Note: The indicator “Public Procurement Law requirement for awards to SMEs clause” considers whether the country has adopted public procurement related laws including clauses, exceptions, or requirements for SMEs. However, there is still limited evidence on the costs and surveys with firms can be used to investigate the barriers and benefits of policies aimed at encouraging the participation challenges experienced by firms in the public procurement of specific groups of firms in public procurement. On the market, and the results from three pilot countries—Croatia, one hand, these policies can align private and public sector Poland, and Romania. incentives and reduce market failures that prevent certain groups of firms from accessing the procurement market and Firm-level survey data can be used to provide some from effectively competing for government contracts. On the initial insights on the application of preference policies, other hand, they can be distortionary, prone to capture, and for example, by evaluating their take-up and therefore difficult to monitor, and can suffer from poor implementation the extent to which they are able to reach the intended and compliance (World Bank, 2021a). The existing evidence segment of the economy. For example, Kenya has an on the impact of these policies is still limited and mixed, affirmative action law on public procurement that stipulates even when conducted on the same preference program.32 A that 30 percent of the value of all government procurement second type of policies consists of non-targeted policies that contracts should be set aside for firms owned by women, make public procurement competitive and accessible for all persons with disabilities, and/or youth.34 Several services firms, and that can have a larger impact on firms traditionally have been introduced and expanded in Kenya to support excluded from the public procurement market, such as progress toward this target, such as the Access to Government access to finance, training, and access to information on Procurement Opportunities Program,35 which certifies firms public procurement (Hjort, Iyer, and de Rochambeau, 2020); that are owned by the disadvantaged groups articulated in the transparency of the procurement system (Knack, Biletska, policy. A rigorous evaluation of this policy is complicated by and Kacker, 2017); timeliness of payments (Barrot and Nanda, the lack of procurement contracts data disaggregated by the 2020; Conti et al., 2021); or slicing bigger contracts into smaller characteristics of the supplier, therefore creating a stronger ones (di Giovanni et al., 2022).33 Section 5 demonstrates how need for survey-based evidence.36 The 2013 and 2018 World 32. For example, even two studies conducted on the same program for small firms in California draw opposite conclusions. While Marion (2007) found that procurement costs are 3.8 percent higher on auctions using preferential treatment, Krasnokutskaya and Seim (2011) found those distortionary effects are not huge when compared with the benefits of firm growth. 33. Even non-targeted policies might be distortionary. For example, as demonstrated by di Giovanni et al. (2022), slicing bigger contracts into smaller ones can help small firms to overcome financial constraints and participate in public procurement, and they can also reduce saving incentives for large firms. 34. Government of Kenya, Public Procurement Regulatory Authority: Kenya Public Procurement and Asset Disposal Act, Revised Edition 2016. 35. Government of Kenya, The National Treasury and Economic Planning: Access to Government Procurement Opportunities (AGPO). 36. IDRC (2021) reports the findings from a 2019 research study which indicate that 41 percent of the businesses registered with the Access to Government Procurement Opportunities (AGPO) program are women-owned Enterprises (WoEs), and of these, 52 percent were able to obtain a government contract in the period under review. However, this research study does not consider the selection of firms into the AGPO program since firms must register into the AGPO in order to be eligible for procure- ment through the affirmative action law. Related to this, IDRC (2021) highlights that “the WoEs that obtain public contracts under AGPO are led by urban, well-educated, and privileged women.” Bank Enterprise Surveys for Kenya allow for the tracking and 10 percentage points for large firms, and by 5 percentage of trends in firms’ participation rates in public procurement, points for women-owned firms and 14 percentage points for disaggregated by the groups of firms targeted by the policy, men-owned firms (Table 2). In 2018, among firms that secured in particular women-owned firms and SMEs. In addition, the or attempted to secure a government contract, 11 percent did Enterprise Survey implemented in Kenya in 2018 included a so through an affirmative action scheme, with similar shares question to identify firms that received a government contract for small, medium, and large firms (Table 3) and for women- under an affirmative action scheme. Between 2013 and 2018, and men-owned firms (Table 3). participation rates increased by 8 percentage points for SMEs > > > T A B L E 2 - Trends in Participation Rates in Public Procurement between 2013 and 2018, by Firm Type (Kenya) “Over the last year, has this establishment secured or at- Firm type tempted to secure a government contract?” 2013 2018 SMEs 16% 24% Large 28% 38% Women-owned 16% 21% Men-owned 22% 36% Source: World Bank, 2013 and 2018 Kenya Enterprise Survey. Note: Statistics are based only on firms that secured or attempted to secure a government contract in the year prior the survey. > > > T A B L E 3 - Share of Firms that Received Any Government Contracts Under Affirmative Platform, by Firm Type (Kenya) Share of countries with Public Procu- Firm type rement Law requirement for awards to SMEs clause Small 11% Medium 10% Large 12% Women-owned 15% Men-owned 14% Source: World Bank, 2018 Kenya Enterprise Survey. Note: Statistics are based only on firms that secured or attempted to secure a government contract in the year prior the survey. 5. >>> Evidence on Barriers to and Challenges of Participation in Public Procurement from the Perspective of Firms 5.1. Why Firm-Level Survey Data is Important Beyond estimating participation rates in public procurement, firm-level survey data can be used to investigate the barriers to participation and challenges experienced by firms in the public procurement market. Section 4 demonstrated how to use firm-level survey data to estimate firms’ participation in public procurement, overall and among different types of firms, which is a useful first step to characterize the public procurement market. The next step for policy design is to understand the causes of the observed patterns and trends and identify the existing constraints, which is necessary in developing the appropriate strategies to make public procure- ment more competitive and inclusive. For example, there might be different barriers to participa- tion in public procurement experienced by firms in different countries (e.g., access to finance, lack of information, administrative and bureaucratic costs, expected delays in payments), and therefore different solutions would be appropriate to stimulate firms’ consideration of expanding their businesses in each market. In-depth, firm-level surveys can be tailored to investigate these topics in detail with a representative sample of firms, therefore avoiding the risk of capturing only selected and biased views from selected stakeholders, which is a typical feature of qualita- tive studies and focus group discussions. Critically, firm-level surveys can target firms with and without prior experience in the public procurement market, thereby allowing understanding of the different barriers and challenges experienced by existing bidders and new entrants. The next sections use a procurement questionnaire procurement and the private sector. This deep-dive on public module piloted in Croatia, Poland, and Romania in the procurement is based on the results from the procurement period May–June 2021 to demonstrate how firm-level module for Croatia and from a report on the performance of survey data can be used to understand firms’ perspectives public procurement in Croatia which analyzed transactional on participation in public procurement. Section 5.2.1 procurement data from the eGP system (World Bank, 2022d). provides statistics on the areas that firms perceive as more critical with respect to public procurement and how these perceptions are formed. Section 5.2.2 investigates the 5.2. Evidence from the WBES differences in perceptions between different types of firms, focusing on SMEs versus larger firms and firms with and without previous participation in public procurement, which is 5.2.1. Assessment of the Main Gaps and helpful to target priority actions to different segments of the Identification of Priority Actions private sector.37 These empirical insights are obtained from a novel procurement module developed by the World Bank In Croatia, Poland, and Romania, firms perceive Governance Global Practice and the World Bank Enterprise administrative procedures before contract signature Surveys Unit, which was piloted in Croatia, Poland, and as the biggest obstacle to attempting to secure a Romania (see details in Annex A).38 The module can be government contract.39 In 2021, the percentage of firms implemented as part of a broader questionnaire, and used that stated that administrative barriers are at least a moderate to compare the views of firms across sectors, firm sizes, obstacle to participation in public procurement is 47 percent experiences in the public procurement market, and other firm in Croatia (Figure 8), 81 percent in Poland (Figure 9), and characteristics. In Croatia, Poland, and Romania, the module 38 percent in Romania (Figure 10).40 Lack of trust in the was administered to all firms participating in the survey, process and expected challenges during contract execution allowing for comparisons on the perceptions of firms with are also perceived as an obstacle by a large share of firms, and without previous participation in public procurement. The while geographical barriers are perceived to be less of an results shown in Section 5.2, other than providing interesting obstacle. The procurement module captures interesting analytical insights for these countries, demonstrate that the patterns across and within countries as well. Firms in Poland procurement module is able to capture different trends and express much more negative views of the existing obstacles patterns even in countries with relatively similar legislation to attempting to secure a government contract than firms in and regulatory framework, thereby supporting the validity and Croatia and Romania.41 In Croatia, administrative barriers usefulness of this survey instrument. before contract signature are perceived as a much stronger obstacle to participation in public procurement than other Results from firm-level survey data are most informative potential challenges,42 while this is less the case for Romania when combined with additional diagnostics on the and Poland.43 procurement system, such as legislative and regulatory reviews, analysis of transactional procurement data, or surveys of procurement officers. For example, the Croatia Country Economic Memorandum 2022 (World Bank, 2023b) contains a chapter on institutions and firm productivity, of which a subsection is a deep-dive on the intersection between public 37. Other firm characteristics may play a role in shaping the perceived barriers to and challenges of participation in public procurement, such as the age of the firm, whether the firm is domestic or international, or whether the firm has ever been awarded a government contract. For demonstration purposes, in this report we focus only on two firm characteristics: the size of the firm and prior participation in public procurement. 38. The opportunity of piloting this new procurement module in Croatia, Poland, and Romania emerged from the ongoing engagements of regional teams in the Governance Global Practice with government counterparts in the area of public procurement, and the timing of the implementation of the COVID-19 Enterprise Surveys. In particular, at the time of launching the first round of the COVID-19 Enterprise Surveys in May 2020, the World Bank Governance Global Practice regional teams were working on a public procurement assessment in Croatia, a public procurement assessment in Romania, and the preparation of a public procurement focused RAS in Poland. 39. This paragraph, Figure 8, Figure 9, and Figure 10 are based on the answers provided to question Q3 in the procurement module piloted in Croatia, Poland, and Romania (Annex A). 40. The difference is statistically significant between Croatia and Poland and Croatia and Romania (at 99% confidence level), and at 90% between Poland and Romania. 41. The difference is statistically significant between Poland and Romania and Poland and Croatia (at 99% confidence level), and at 90% between Croatia and Romania. 42. For Croatia, the share of firms that report administrative barriers before contract signature as a moderate-very severe obstacle is statistically significantly higher (at 95% confidence level) than the share of firms that report the other potential challenges as moderate-very severe obstacles 43. A wide range of factors might explain these differences across countries, such as public procurement laws, eGP systems, public sector institutions, and trust toward the state, of which investigation is beyond the scope of this paper. > > > F I G U R E 8 - Perception of Obstacles (Croatia) Source: World Bank Enterprise Survey. Note: Weighted results. N = 324. Statistics are computed using follow-up sample weights (“wmedian_COVID”). The percentages on the left panel report the share of firms that indicate each challenge as a “no obstacle”. The percentages on the right panel report the share of firms that indicate each challenge as a moderate – very severe obstacle. > > > F I G U R E 9 - Perception of Obstacles (Poland) Source: World Bank Enterprise Survey. Note: weighted results. N = 592. Statistics are computed using follow-up sample weights (“wmedian_COVID”). The percentages on the left panel report the share of firms that indicate each challenge as a “no obstacle”. The percentages on the right panel report the share of firms that indicate each challenge as a moderate – very severe obstacle. > > > F I G U R E 1 0 - Perception of Obstacles (Romania) Source: World Bank Enterprise Survey. Note: weighted results. N = 252. Statistics are computed using follow-up sample weights (“wmedian_COVID”). The percentages on the left panel report the share of firms that indicate each challenge as a “no obstacle”. The percentages on the right panel report the share of firms that indicate each challenge as a moderate – very severe obstacle. The procurement module piloted in Croatia, Romania, in Romania (57 percent), and other forms of unfair competition and Poland in May–June 2021 collects granular and in Croatia (44 percent) (Figure 12).46 The main geographical detailed data that allows for the identification of specific barrier to securing a government contract is the costs to provide priority actions.44 The main administrative obstacle before goods and services outside the firms’ locality in Romania (62 contract signature is the amount of effort needed for bid percent) and Poland other geographical barriers in Croatia preparation in Croatia (44 percent) and Romania (55 percent), (40%) (Figure 13).47 The biggest challenge when working and the length of the process after bid submission in Poland under a government contract is delays in payments (51 (40 percent) (Figure 11).45 Lack of trust in the procurement percent) in Romania and Poland, and too many administrative process is mainly due to requests for informal payments (37 procedures during contract execution (37 percent) in Croatia percent) and other forms of unfair competition (40 percent) in (Figure 14).48 Poland, contracts being typically awarded to connected firms . 44. This paragraph, Figure 11, Figure 12, Figure 13, and Figure 14 are based on the answers provided to question Q4–Q7 in the procurement module piloted in Croatia, Poland, and Romania (Annex A). 45. The difference in the share of firms that report that the main administrative obstacle before contract signature is the amount of effort needed for bid preparation is sta- tistically significant (at 99% confidence level) between Romania and Poland and between Croatia and Poland, but not between Croatia and Romania. The difference in the share of firms that report that the main administrative obstacle before contract signature is the length of the process after bid submission is statistically significant (at 90% confidence level) between Poland and Romania and between Poland and Croatia, but not between Croatia and Romania. 46. The difference in the share of firms that report that requests for informal payments are the main reason for lack of trust is statistically significant (at 90% confidence level) between Croatia and Poland and between Croatia and Romania, but not between Poland and Romania. The difference in the share of firms that report that contracts being typically awarded to connected firms are the main reason for lack of trust is statistically significant (at 99% confidence level) between Romania and Poland and between Romania and Croatia, but not between Croatia and Poland. The difference in the share of firms that report that other forms of unfair competition are the main reason for lack of trust is not statistically significant (at 90% confidence level) between the three countries. 47. The difference in the share of firms that report that the costs to provide goods and services outside the firms’ locality are the main geographical barrier is statistically significant (at 95% confidence level) between Croatia and Poland, and between Poland and Romania, and between Croatia and Romania. The difference in the share of firms that report “other” as the main geographical barrier is statistically significant (at 90% confidence level) between Croatia and Poland, and between Croatia and Romania, but not between Poland and Romania. 48. The difference in the share of firms that report delays in payments as the main challenge when working under a government contract is statistically significant (at 99% confidence level) between Croatia and Poland, and between Croatia and Romania, but not between Poland and Romania. The difference in the share of firms that report too many administrative procedures during contract execution as the main challenge when working under a government contract is statistically significant (at 99% confidence level) between Croatia and Poland, and between Croatia and Romania, but not between Poland and Romania. > > > > > > F I G U R E 1 1 - Main Administrative Obstacle F I G U R E 1 2 - Main Reason for Lack of Trust in Before Contract Signature the Process Source: World Bank Enterprise Survey. Source: World Bank Enterprise Survey. Note: N = 784 (Croatia: 197; Poland: 459; Romania: 128). Statistics are Note: N = 670 (Croatia: 120; Poland: 429; Romania: 121); 1137 missing computed using follow-up sample weights (wmedian COVID). This question values (Croatia: 204; Poland: 554; Romania: 379). Statistics are computed was asked only to firms that indicated that administrative procedures before using follow-up sample weights (wmedian COVID. This question was asked contract signature are an obstacle. only to firms that indicated that lack of trust in the process is an obstacle. > > > > > > F I G U R E 1 3 - Main Geographical Barrier to F I G U R E 1 4 - Biggest Challenge Expected when Securing a Government Contract Working Under a Government Contract Source: World Bank Enterprise Survey. Source: World Bank Enterprise Survey. Note: N = 530 (Croatia: 60; Poland: 387; Romania: 83); 1277 missing values Note: N = 695 (Croatia: 121; Poland: 465; Romania: 109); 1112 missing values (Croatia: 264; Poland: 596; Romania: 417). Statistics are computed using (Croatia: 203; Poland: 518; Romania: 391). Statistics are computed using follow-up sample weights (wmedian COVID. This question was asked only to follow-up sample weights (wmedian COVID. This question was asked only firms that indicated that geographical barriers to secure a government contract to firms that indicated that expected challenges during contract execution are are an obstacle. an obstacle. 5.2.2. Target Priority Actions to Different Groups of Firms Different types of firms have different perceptions of the for lack of trust in the process (Figure D.6).50 In Romania, obstacles to participation in public procurement, and this small firms are more likely than medium and large firms to information is important to target the right policies to the identify the costs to provide goods or services outside their right segment of the private sector. For example, small locality as the main geographical barrier (Figure D.11).51 firms have different views than medium and large firms on a variety of barriers and challenges related to the public Depending on their experience with the procurement procurement market. In Croatia, small firms are more likely system, firms that secured or attempted to secure a than medium and large firms to identify unclear guidelines on government contract may have different perceptions of how to participate as the main administrative barrier (Figure existing obstacles than firms that have never entered D.1).49 In Poland, small firms are more likely than medium and the public procurement market. Firms may have negative large firms to indicate unfair competition as the main reason perceptions of how to operate in the public procurement market 49. In Croatia, the difference in the share of firms that report unclear guidelines on how to participate as the main administrative barrier is statistically significant (at 95% confidence level) between small and medium firms, and between small and large firms. 50. In Poland, the difference in the share of firms that report unfair competition as the main reason for lack of trust in the process is statistically significant (at 90% confidence level) between small and medium firms, and between small and large firms. 51. In Romania, the difference in the share of firms that report the costs to provide goods or services outside their locality as the main geographical barrier is statistically significant (at 99% confidence level) between small and medium firms, and between small and large firms. because they are not familiar with the system, or because firms from getting discouraged once they enter the public they experienced first-hand the challenges that it implies. procurement market. Speaking to the first explanation, in Croatia, firms that did not participate in public procurement in the three years prior to the Similarly, expectations of bribery in public procurement survey perceive geographical barriers as a bigger obstacle, may (or may not) shape firms’ perceptions and trust suggesting that firms might underestimate their probability towards the public procurement system. In the 2019 of success in this market (Table D.9). In line with the second standard Enterprise Surveys, the share of firms that explanation, in Poland, firms that secured or attempted to participated in public procurement and believe that private secure a government contract in the three years prior to the operators similar to themselves make informal payments or survey have more negative views of the procurement process, give gifts to public officials to secure government contracts is as they perceive administrative procedures before contract 12 percent in Croatia, 14 percent in Poland, and 19 percent signature, geographical barriers, lack of trust in the process, in Romania, higher than the regional average of 8 percent and expected challenges during contract executions as bigger in Europe and Central Asia.52 Exploiting the panel nature of obstacles (Table D.10). In Romania, previous participation in the COVID-19 Enterprise Surveys, it is possible to estimate public procurement does not seem to have an encouraging the association of these expectations of bribery with trust nor discouraging effect on firms (Table D.11). These findings towards the public procurement system. Table 4 shows that, have various policy implications. For example, in the case of in 2019, Croatian firms that expected giving a bribe to secure Croatia, it might be useful to create more awareness in the a government contract giving are more likely, in 2021, to cite private sector of the procurement procedures and existing lack of trust in the process as a more severe obstacle (but business opportunities in order to reduce the perception this mechanism does not hold for Polish and Romanian firms). of obstacles among firms that never entered the public These results indicate that the business environment can procurement market. In the case of Poland, it may be useful shape firms perceptions, and these in turn can have important to improve the existing procurement system and the quality implications on their participation in public procurement. of interactions with firms, and reduce complexities, to prevent > > > T A B L E 4 - Experience of bribery and lack of trust in the process Source: World Bank Enterprise Survey. Note: * p < 0.1; ** p < 0.05; *** p < 0.01. Ordinary Least Square. Dependent variables: To what degree Lack of Trust in the process is an obstacle to attempting to secure a government contract, on a scale between 0 (no obstacle) and 4 (very severe obstacle)? The dependent variable takes value 0 if the answer is “no obstacle,” and 100 if the answer is between “Minor obstacle” to “Very severe obstacle.” Controls: firm age (log), number of employees (log), a dummy variable if the firm is run by a woman, country fixed effects and sector fixed effects (manufacturing, retail services, and other services). Results are weighted so that the sample is representative at national level. Sample (1) refers to only Croatian firms, (2) to Polish firms, and (3) to Romanian firms. Robust standard errors clustered at firm size and sector level. 52. World Bank Enterprise Surveys, Corruption. 6. >>> Evidence on the Experience of Bidders and Suppliers 6.1. Why Firm-Level Survey Data is Important Firm-level surveys can be used to generate evidence on the specific experiences of bid- ders and suppliers, especially for areas that are difficult to observe from administrative data, such as instances of corruption and practices during contract implementation. As demonstrated in Section 5.2, firms that already participated in public procurement have different experiences, views, and perceptions than firms that never entered this market. Data on firms that never participated in public procurement is necessary to estimate participation rates and understand the causes that prevent them from exploring these business opportunities with the government. At the same time, depending on the objectives of the analysis and the type of policy decisions that it intends to inform, in some cases, it may be more relevant to focus on the behav- iors and experiences of bidders or suppliers rather than all firms (for example, when the policy interest is on investigating aspects related to the submission of complaints, contract manage- ment or contract implementation). In these cases, it is not necessary to analyze data from the universe of firms or from a representative sample of the private sector, but it would be sufficient to focus on suppliers and bidders.53 Section 6.2 uses the World Bank Enterprise Surveys data to demonstrate how to use firm-level survey data to investigate an area that is typically difficult to observe – corruption risks for firms that participate in public procurement. 53. When focusing on the experiences of bidders and suppliers, data from the eGP system can capture relevant information for the analysis, especially for countries where bidders have to register on the eGP portal to operate in the procurement market, and where official firm-level identifiers are available to connect eGP data with other administrative data, such as tax registries. However, firm-level survey data remain relevant for two reasons. First, only few countries currently meet these conditions. Second, there are areas that are difficult to observe from administrative data, such as corruption risks, perceptions, and views of firms. When available from the eGP data, the list of registered 6.2. Evidence from the WBES suppliers and bidders can be used to precisely define the target population of interest and develop efficient strategies to invite them to take part in the survey. Firm- 6.2.1. Assessment of Corruption Risks in level surveys that are representative of the entire private sector Public Procurement are, by design, able to capture only a small share of firms that ever participated in public procurement, and therefore they Globally, 21 percent of firms are expected to give gifts to may not be the most efficient solution for studies that focus on secure a government contract.54 The Enterprise Surveys the experiences and perceptions of suppliers and bidders. In include a question on whether firms consider that firms similar these cases, there are excellent opportunities for combining to theirs make informal payments or give gifts to secure survey data and administrative data; in particular, the list of government contracts, which can be interpreted as a proxy of suppliers and bidders registered on the eGP portal can be used self-reported incidence rate of bribery in public procurement.55 to identify the target population and the sampling strategy. For Globally, 21 percent of firms that secured or attempted to example, this strategy is being applied to an ongoing study on secure at least one government contract in the 12 months public procurement in Romania, which relies on a firm-level prior to the interview state that firms similar to theirs typically survey targeted to firms registered on the Romanian eGP make informal payments or give gifts to secure government portal (World Bank, 2022c). Annex B discusses in more detail contracts, ranging from 0 percent in Sweden to 88 percent in the methodology for implementing a survey representative Pakistan (Figure 15).56 of the private sector overall, when it is most appropriate, and how to adapt it to cases when it is sufficient to collect data from a representative sample of suppliers or bidders. 54. The self-reported incidence rate of bribery in public procurement is calculated from the WBES using question j6 in the standard Enterprise Survey questionnaire. The exact question in the standard Enterprise Survey questionnaire reads: “When establishments like this one do business with the government, what percent of the contract value would be typically paid in informal payments or gifts to secure the contract?”. This question is asked only to those firms that report having secured or attempted to secure a government contract in the year prior to the interview (question j6a). In line with the Enterprise Survey methodology to calculate the indicator “Percent of firms expected to give gifts to secure a government contract”, it is assumed that firms that refuse to answer this question consider that firms with characteristics similar to theirs are making informal payments or giving gifts to public officials to secure a government contract. 55. The indicator “Percent of firms expected to give gifts to secure government contract” is available at the Enterprise Surveys website. 56. At country-level, the rate of firms expected to give gifts to secure a government contract does not exhibit clear correlations with other government corruption indexes, such as the Worldwide Governance Indicators – Control of Corruption index (World Bank, 2022f), suggesting that this index captures specific features of the public procurement process rather than overall corruption levels. > > > F I G U R E 1 5 - Country-Level Rate of Firms Expected to Give Gifts to Secure a Government Contract Source: World Bank Enterprise Survey. Note: N = 13,666. The intervals are computed using Jenks natural break. Statistics are computed using the latest ES wave available for each country and using sample weights (“wt”). The percent of firms expected to give gifts to secure a South Asia (29 percent). The rate of firms expected to give government contract varies significantly across regions gifts to secure a government contract is lowest in Europe and (Figure 16). This is more likely among firms in East Asia and Central Asia region (7 percent) and Latin America and the Pacific (34 percent), Sub-Saharan Africa (31 percent), and Caribbean (14 percent). > > > F I G U R E 1 6 - Rate of Firms Expected to Give Gifts to Secure a Government Contract Across Region Source: World Bank Enterprise Survey. Note: weighted results. N = 13,666. Each value indicates the share of firms. Statistics are computed using the latest ES wave available for each country and using sample weights (“wt”). 7. >>> Conclusions This paper provides guidance on how to use firm-level survey data to generate evidence on the public procurement market from the firms’ perspective. It elaborates on why firm- level survey data are necessary to better understand the public procurement market and firms’ perspectives, especially in the absence of high-quality administrative microdata. Sections 4, 5, and 6 detail the dimensions that can be measured – such as participation rates in public procure- ment, barriers to and challenges of participation in public procurement, and experiences of bid- ders and suppliers – and the type of analysis that can be conducted, including assessing trends and comparing segments of the economy. It also discusses how firm-level survey data can complement administrative data from procurement transactions, as well as the various consid- erations on survey design, such as targeting of all firms in an economy versus targeting of firms already active in the public procurement market (i.e., bidders and suppliers). This paper is part of a broader effort to continuously expand the available data, statistics, and tools for evidence- based policy making in public procurement (World Bank, 2022a). This paper also reports novel global statistics on firms’ and SMEs’ participation in public procurement using the World Bank Enterprise Surveys data. This paper examines firms’ participation rate in public procurement for 150 countries, and evidence from a public procure- ment module piloted in Croatia, Poland, and Romania in May–June 2021. Globally, firms’ partici- pation rate in public procurement is 18 percent, and small firms are 12 percentage points less likely to participate in public procurement than larger firms. For a subsample of countries with Enterprise Surveys data available over multiple years, the trends show that participation rates in public procurement have not been increasing over time. The evidence from a novel procure- ment module piloted in Croatia, Poland, and Romania provides additional evidence on some of the barriers and challenges experienced by firms, such as the amount of effort required for bid preparation, the length of the process after bid submission, lack of trust in the procurement pro- cess, and delays in payments. This evidence suggests that technological innovations and eGP adoption have the potential of stimulating participation and competition in public procurement, but they can be most effective when accompanied by systemic changes addressing the wide range of obstacles experienced by firms, especially SMEs. The increasing attention toward the strategic role of The Governance Global Practice and the Development public procurement policies in promoting inclusive and Economics Group of the World Bank made available sustainable growth calls for further research on the various resources to researchers and World Bank task interactions between private and public actors in the teams working on these topics. The World Bank Enterprise public procurement market, which this paper contributes Surveys microdata are publicly available for research and to. An emerging academic literature is generating evidence on analytical studies.59 This paper provides resources for which policies can be effective in attracting firms to participate the application of firm-level surveys in the area of public in public procurement. This literature encompasses studies procurement, in particular the public procurement module on preference policies targeted to specific groups of firms (Annex A) and the survey implementation methodology (Annex (e.g., SMEs or domestic firms), as well as non-targeted B). Additional support, such as additional references and policies, such as access to information on public procurement lessons from previous applications,60 assistance in the design (Hjort, Iyer, and de Rochambeau, 2020), transparency of the and implementation of analytical projects on the links between procurement system (Knack, Biletska, and Kacker, 2017), public procurement and the private sector, and connections timeliness of payments (Barrot and Nanda, 2020; Conti et with data analysts and researchers from the World Bank or al., 2021), and slicing bigger contracts into smaller ones (di from external research institutions, are also available to task Giovanni et al., 2022).57 The World Bank has also been actively teams upon request. promoting this research agenda, for example, through a series of research studies on the link between public procurement and private sector growth.58 57. Even non-targeted policies might be distortionary. For example, as demonstrated by di Giovanni et al. (2022), slicing bigger contracts into smaller ones can help small firms to overcome financial constraints and participate in public procurement, but they can also reduce saving incentives for large firms. 58. For example, the Governance Global Practice and the DIME-Governance are working on a series of research studies under the World Bank project Public Procurement and Firm Behavior (P177551). 59. See World Bank Enterprise Surveys microdata. 60. For example, the design choices and lessons drawn from a firm-level survey targeted to firms registered on the Romanian eGP portal, conducted as part of an ongoing study on public procurement in Romania, “Supporting the Efficiency of the Romanian Public Procurement System” (World Bank, 2022c). >>> References Barrot, Jean-Noël, and Ramana Nanda. 2020. “The Employment Effects of Faster Payment: Evidence from the Federal Quickpay Reform.” The Journal of Finance. Bosio, Erica, Simeon Djankov, Edward L. Glaeser, and Andrei Shleifer. 2022. “Public Procure- ment in Law and Practice.” American Economic Review. Cappelletti, Matilde, Leonardo M. Giuffrida, and Gabriele Rovigatti. 2022. “Procuring Survival.” CESifo Working Paper Series. Conti, Maurizio, Leandro Elia, Antonella Rita Ferrara, and Massimiliano Ferraresi. 2021. “Gov- ernments’ Late Payments and Firms’ Survival: Evidence from the European Union.” The Journal of Law & Economics. di Giovanni, Julian, Manuel García-Santana, Priit Jeenas, Enrique Moral-Benito, and Josep Pijoan-Mas. 2022. “Government Procurement and Access to Credit: Firm Dynamics and Aggre- gate Implications.” CEPR Discussion Paper Series. Ferraz, Claudio, Frederico Finan, and Dimitri Szerman. 2015. “Procuring Firm Growth: The Ef- fects of Government Purchases on Firm Dynamics.” NBER Working Paper Series. Government of Kenya, Public Procurement Regulatory Authority. 2016. Public Procurement and Assets Disposal Act, Revised Edition 2016. Kenya Gazette Supplement No. 207 (Act No. 33). Government of Kenya, The National Treasury and Economic Planning. 2015. Access to Govern- ment Procurement Opportunities (AGPO). Hjort, Jonas, Vinayak Iyer, and Golvine de Rochambeau. 2020. “Informational Barriers to Market Access: Experimental Evidence from Liberian Firms.” NBER Working Paper Series. Hoekman, Bernard, and Marco Sanfilippo. 2020. “Firm Performance and Participation in Public Procurement: Evidence from Sub-Saharan Africa.” Review of World Economics. Hoekman, Bernard, and Bedri Kamil Onur Taş. 2022. “Procurement Policy and SME Participa- tion in Public Purchasing.” Small Business Economics. Hyland, Marie Caitriona, Nona Karalashvili, Silvia Muzi, and Domenico Viganola. 2021. Female- Owned Firms during the COVID-19 Crisis. World Bank Global Indicators Briefs. Washington, DC: World Bank. IDRC. 2021. “Public Procurement and Women’s Economic Empowerment in East Africa – Challenges and Opportunities.” IDRC Evidence Review. Knack, Stephen, Nataliya Biletska, and Kanishka Kacker. 2017. “Deterring Kickbacks and En- couraging Entry in Public Procurement Markets: Evidence from Firm Surveys in 88 Developing Countries.” World Bank Policy Research Working Paper Series. Krasnokutskaya, Elena, and Katja Seim. 2011. “Bid Preference Programs and Participation in Highway Procurement Auctions.” American Economic Review. Marion, J. 2007. “Are Bid Preferences Benign? The Effect of Small Business Subsidies in High- way Procurement Auctions.” Journal of Public Economics. Muzi, Silvia, Filip Jolevski, Kohei Ueda, and Domenico Viganola. 2022. “Productivity and Firm Exit during the COVID-19 Crisis: Cross-Country Evidence.” Small Business Economics. OECD. 2021. Government at a Glance 2021. Paris: OECD. World Bank. 2020. e-Procurement Reform in Bangladesh. In Enhancing Government Effective- ness and Transparency: The Fight Against Corruption. Washington, DC: World Bank. World Bank. 2021a. Strengthening World Bank SME-Support Interventions: Operational Guid- ance Document. Washington, DC: World Bank. World Bank. 2021b. Croatian Women and the COVID-19 Pandemic – The Coronavirus Is Not Gender-Blind. Washington, DC: World Bank. World Bank. 2022a. Using Data Analytics in Public Procurement: Operational Options and a Guiding Framework. Washington, DC: World Bank. World Bank. 2022b. An International Stocktaking of Developments in Public Procurement: Syn- thesis Report. Washington, DC: World Bank. World Bank. 2022c. Supporting the Efficiency of the Romanian Public Procurement System. World Bank project (P178991). World Bank. 2022d. Measuring Performance of Public Procurement in Croatia – Report on the Analysis of the Effectiveness of the Public Procurement System in Croatia. Washington, DC: World Bank. World Bank. 2022e. GovTech Maturity Index (GTMI) Data Dashboard. Washington, DC: World Bank. World Bank. 2022f. Worldwide Governance Indicators. Washington, DC: World Bank. World Bank. 2023a. The Expanding Role Of Public Procurement In Africa’s Economic Develop- ment. Governance for Development blogs. Washington, DC: World Bank. World Bank. 2023b. Croatia Country Economic Memorandum. Washington, DC: World Bank. >>> Annex A – Public Procurement Module from the Enterprise Survey Questionnaire > > > Q.1 Is public procurement a relevant market segment for this establishment? Yes 1 No 2 Don’t know -9 > > > Q.2 Over the last three years, has this establishment secured or attempted to secure a government contract? Yes, attempted but did not secure any contract or in process 1 Yes, attempted and secured at least one contract 2 No 3 Don’t know -9 > > > Q.3 To what degree are each of the following an obstacle to attempting to secure a government contract, on a scale between 0 (no obstacle) and 4 (very severe obstacle)? No Minor Moderate Major Very Severe Does not Don’t know obstacle obstacle obstacle obstacle Obstacle apply Administrative procedures before 0 1 2 3 4 -9 -7 contract signature Q3A Lack of trust in the process 0 1 2 3 4 -9 -7 Q3B Geographical bar- riers 0 1 2 3 4 -9 -7 Q3C Expected challen- ges during contract 0 1 2 3 4 -9 -7 execution Q3D > > > Q . 4 A S K O N LY I F Q 3 A > 0 What is the main administrative obstacle before contract signature? Unclear guidelines on how to participate 1 Too much effort for bid preparation 2 Length of the process after bid submission 3 Risk of appeals by other firms 4 Other 5 Don’t know -9 > > > Q . 5 A S K O N LY I F Q 3 B > 0 What is the main reason for lack of trust in the process? Contracts typically awarded to politically connected firms 1 Other forms of unfair competition 2 Requests of informal payments or gifts to secure the contract 3 Other 4 Don’t know -9 > > > Q . 6 A S K O N LY I F Q 3 C > 0 What is the main geographical barrier to secure a government contract? Lack of government contracts in this establishment’s locality 1 Too costly to provide goods or services outside this establishment’s locality 2 Lack of connections outside this establishment’s locality 3 Other 4 Don’t know -9 > > > Q . 7 A S K O N LY I F Q 3 D > 0 What is the biggest challenge that is expected when working under a government contract? Too many administrative processes during contract execution 1 Risks of renegotiations or changing orders 2 Delays in payments 3 Complicated regulations on subcontracting 4 Other 5 Don’t know -9 >>> Annex B – Survey Implementation Methodology B.1. Representative Sample of of the firm and its constraints. During the fieldwork phase, usually performed by an independent contractor, the data all Firms: The Experience of the collection is constantly monitored and validated by the World World Bank Enterprise Surveys Bank Enterprise Surveys Unit. The World Bank standard Enterprise Surveys are nationally The sampling methodology designed for and followed by representative surveys of registered (formal) private firms the Enterprise Surveys allows providing a representative with at least five employees. Owners and top managers of picture of the (formal) private sector within a country. Within establishments operating in the manufacturing or services each country, firms are allocated to different strata based on sectors are interviewed on a broad range of business their sector, size, and location within the country. Then, the environment topics, spanning from access to finance to sampling weights provided in the surveys can be employed to interactions with the governments. All the Enterprise Surveys get indicators representative at the country level that correct are characterized by a common sampling methodology for the unequal probability of selection across strata and for (stratified random sampling) and by a standardized ineligibility in the fieldwork. questionnaire, and therefore they allow full comparability across countries and over time. The standard Enterprise When the COVID-19 pandemic hit, building on the frames Surveys are collected in different rounds across countries, identified for the standard WBES, the World Bank Enterprise typically 4 to 6 years apart, with part of the interviewed firms Survey Unit performed phone interviews to monitor the impact drawn from the previous round of participants. The Enterprise on the private sector of the crisis induced by the pandemic. Surveys data is published on the WBES platform.61 On top of safety concerns for interviewers and interviewees, the interviews were administered over the phone as they The standard Enterprise Surveys are implemented as it were meant to be simple follow-ups to the broader WBES follows. First, a survey design is defined by using inputs from questionnaire, with mostly categorical and yes/no types the best possible sampling frame available and the most of questions. The COVID-19 WBES builds on the WBES accurate estimate of the universe of the private sector of an methodology and therefore it is also fully comparable across economy. Sampling frames are the most up-to-date listing of countries. For more than forty countries worldwide, up to three available establishments’ contacts, usually generated with the different rounds of data collection were performed: the first support of statistical offices or private firms; survey designs one in the immediate aftermath of the crisis (spring 2020), the can be thought of as the target of completed interviews for others 5 to 6 months apart, on average.62 The procurement each combination of size, sector, and region stratification. module used for the analysis in Section 5.2 and reported in Then, establishments are randomly selected for the interviews, Annex A was included in the questionnaire implemented in the which are customarily completed face to face with the owners third round of COVID-19 WBES data collection in Romania, of the businesses or by top managers, i.e., with the persons Croatia, and Poland between May and June 2021. who have the broadest and deepest knowledge of the insights 61. https://www.enterprisesurveys.org/en/enterprisesurveys 62. https://www.enterprisesurveys.org/en/covid-19 B.2. Representative Sample of However, they do not necessarily represent the experiences of bidders and suppliers, which can be a quite selected Suppliers and Bidders: Drawing the sample and different from other firms. The appropriateness of Sampling Frame from Procurement the sample depends on the objectives of the analysis and the type of policy decisions that it is intended to inform. Focusing Data on bidders and suppliers is relevant, for example, when the objective of the analysis is to understand aspects related to Firm-level surveys representative of the full private sector in contract management and contract implementation. an economy, such as the standard WBES and the COVID-19 WBES surveys, are not necessarily representative of specific In these cases, a firm-level survey can be designed in order to subsamples that were not used to define the survey design provide a representative sample of suppliers or bidders. The and sampling frame. For example, the selection of firms survey methodology of the World Bank Enterprise Surveys participating in the WBES or COVID-19 WBES surveys does can be adjusted in order to build a representative sample of not depend on whether firm participated in public procurement suppliers or bidders. The key input for the definition of the or not, and therefore these surveys are representative at sampling frame is the list of firms in the universe of interest, country level and in the strata used for the survey design such as the list of suppliers or the list of bidders from the eGP (size, sector and location) but not in the subsample of system. Depending on the firm characteristics included in the firms that previously participated in public procurement. As available lists, the survey design can be adjusted in order to specified in Section 4.2.1, on average 18 percent of firms build a representative sample of suppliers or bidders over participated in public tenders in the twelve months preceding some strata, defined for example by firm size, sector, and their participation in the survey, and participation rates vary location. The granularity of the stratification can be improved considerably across countries. Because only a minority of by connecting the list of suppliers or bidders from the eGP firms secured or attempted to secure a government contract system with firm registries (e.g., tax registries), which typically in each country, representativeness in this subsample is an include more firm characteristics than in public procurement important consideration. data. As an example, an ongoing study on public procurement for Romania (World Bank, 2022c) relies on a firm-level survey Firm-level surveys whose survey design was not stratified targeted to firms registered on the Romanian eGP portal. The for ensuring representativeness in the subsample of firms list of firms registered on the eGP portal includes a few firm that participated in public procurement, such as the standard characteristics that can be used for the sampling strategy, WBES and the COVID-19 WBES surveys, are best used to such as year of registration on the eGP portal, firm size and estimate participation rates in public procurement in the full location. private sector, and to investigate the perceived difficulties in securing a government contract for all firms, including those that never participated in these processes. >>> Annex C – Definition of Key Variables and Empirical Analysis > > > C . 1 Variable Definitions • Firm age: The time between the date of the survey and • Sectors: The firm sector is defined as the (self-reported) the date of firm’s establishment. main sector of operations of the firm. For analysis at global or regional level, sectors are identified by the two • Employees: Total number of full-time equivalent workers digits ISIC Rev. 3.1 codes.64 For analysis at country level, in the establishment recorded on the date of the survey. due to constraints in sample size, the following three macro-sectors are used: Retail, Manufacturing and Other • Female-run: The establishment is considered female-run services. if the respondent declares that the top-manager is female. > > > • Expected to give a gift: Dummy variable that takes value 1 if the firm reports that similar establishments would C . 2 Weighting typically pay more than 0 percent of contract value For each country, the sample is stratified by industry, firm size, in informal payments or gifts to secure a government and location within the country. Based on this sampling design, contract, or if the firm refused to answer this question.63 sampling weights are constructed for each combination When establishments like this one do business with the of industry, firm size, and location within each country. The government, what percent of the contract value would sampling weights are used to correct for unequal probability of be typically paid in informal payments or gifts to secure selection as well as for ineligibility, and therefore they are used the contract? In the standard Enterprise Surveys, this to obtain statistics that are representative of the underlying variable is defined only for firms that secured or attempted population within each country.65 to secure a government contract in the 12 months prior to the interview. In addition, to account for the pooling of data across countries, survey weights are re-scaled to sum to 1, so that each country • Firm size categories: Using the number of equivalent is equally considered in the estimations. full-time employees in the establishment, the following categories are defined following the WBES Unit standards: > > > • Small: employees ≤ 20 • Medium: employees > 20 and ≤ 99 C . 3 Regression Analysis • Large: employees > 99 All the regression models are specified with an Ordinary Least Square estimator. Outcome variables that take value 0 or 1 • SME: Dummy variable that takes value 1 if the firm is a are transformed to take value 0 or 100: thus, we can interpret Small or Medium enterprise, according to the Firm size the effects estimated as effects in percentage points on the categories. outcome of interest. All regression models include a set of firm-level characteristics, such as firm size, firm productivity, • Participation in public procurement: For the standard and firm age, plus country and sector fixed effects to control Enterprise Surveys, dummy variable that takes value 1 for unobserved heterogeneity across these dimensions. if the firm secured or attempted to secure a government Standard errors are robust and clustered at the country, contract in the 12 months prior to the interview. For the sector, and firm size level. Survey weights are used to make COVID-19 Enterprise Surveys in Romania, Croatia and the sample representative within each country and to correct Poland, dummy variable that takes value 1 if the firm for the pooling of data across countries so that each country is secured or attempted to secure a government contract in equally considered in the estimations. the three years prior to the interview. 63. The exact question in the standard Enterprise Survey questionnaire reads: “When establishments like this one do business with the government, what percent of the contract value would be typically paid in informal payments or gifts to secure the contract?”, and it is asked only to those firms that report having secured or attempted to secure a government contract in the year prior to the interview. In line with the indicators computed by the Enterprise Surveys, it is assumed that firms that refuse to answer this question consider that firms with characteristics similar to theirs are making informal payments or giving gifts to public officials to secure government contract: https://www.enterprisesurveys.org/content/dam/enterprisesurveys/documents/Indicator-Description.pdf. 64. Additional information on the International Standard Industrial Classification of All Economic Activities (ISIC) is available at: https://siccode.com/page/what-is-an-isic- code. 65. Additional information on the Enterprise Surveys sampling methodology and sampling weights can be found here: https://www.enterprisesurveys.org/content/dam/enter- prisesurveys/documents/methodology/Sampling_Note-Consolidated-2-16-22.pdf. >>> Annex D – Additional Analysis and Robustness Checks > > > D . 1 Global Sample > > > T A B L E D . 1 - Sector-Specific Participation Rates in Public Procurement Participation rates in Public Procurement Latin Europe East Middle East Sector America Sub-Saha- and Asia and South Overall and ran Central and North Afri- Asia The Africa Asia Pacific ca Caribbeans Air transport 5% 6% 1% 15% 18% 0% 5% Computer and rela- 27% 28% 19% 42% 20% 9% 16% ted activities Construction 26% 26% 29% 31% 29% 17% 17% Hotels and restau- 6% 4% 9% 7% 8% 11% 5% rants Land transport; transport via pipe- 6% 4% 13% 17% 7% 20% 16% lines Manufacture of 18% 53% 10% 3% 11% 19% 3% basic metals Manufacture of che- micals and chemical 11% 13% 8% 11% 22% 10% 7% products Manufacture of coke, refined petro- 7% 16% 3% 2% 4% 43% 14% leum products and nuclear fuel Manufacture of elec- trical machinery and 14% 16% 13% 6% 23% 23% 9% apparatus n.e.c. Manufacture of fabricated metal products, except 13% 8% 7% 39% 12% 17% 7% machinery and equipment Manufacture of food products and 10% 6% 7% 18% 7% 18% 13% beverages Manufacture of furniture; manufac- 10% 7% 17% 12% 4% 10% 10% turing n.e.c. Manufacture of ma- chinery and equip- 11% 9% 10% 21% 21% 11% 5% ment n.e.c. Manufacture of me- dical, precision and 23% 43% 5% 19% 12% 19% 4% optical instruments, watches and clocks Manufacture of mo- tor vehicles, trailers 11% 14% 9% 5% 1% 16% 17% and semi-trailers Participation rates in Public Procurement Latin Europe East Middle East Sector America Sub-Saha- and Asia and South Overall and ran Central and North Afri- Asia The Africa Asia Pacific ca Caribbeans Manufacture of office, accounting 19% 8% 0% 55% 0% 0% 100% and computing ma- chinery Manufacture of other non-metallic 14% 11% 20% 12% 8% 16% 16% mineral products Manufacture of oth- er transport equip- 7% 8% 3% 3% 33% 3% 11% ment Manufacture of paper and paper 14% 27% 3% 5% 12% 9% 4% products Manufacture of radio, television and communica- 5% 15% 3% 0% 11% 3% 0% tion equipment and apparatus Manufacture of rubber and plastics 7% 9% 5% 3% 11% 10% 5% products Manufacture of 8% 5% 13% 8% 1% 7% 7% textiles Manufacture of 10% 0% 0% 2% 0% 15% 0% tobacco products Manufacture of wearing apparel; 7% 5% 10% 9% 7% 5% 10% dressing and dyeing of fur Manufacture of wood and of prod- ucts of wood and cork, except furni- 7% 8% 3% 3% 10% 7% 4% ture; manufacture of articles of straw and plaiting materials Post and telecom- 10% 13% 4% 27% 27% 3% 43% munications Publishing, printing and reproduction of 11% 14% 4% 9% 17% 16% 20% recorded media Recycling 21% 24% 1% 12% 7% 5% 25% Retail trade, except of motor vehicles and motorcycles; re- 8% 8% 8% 9% 7% 4% 10% pair of personal and household goods Participation rates in Public Procurement Latin Europe East Middle East Sector America Sub-Saha- and Asia and South Overall and ran Central and North Afri- Asia The Africa Asia Pacific ca Caribbeans Sale, maintenance and repair of motor vehicles and motor- 10% 9% 10% 9% 4% 16% 12% cycles; retail sale of automotive fuel Supporting and auxiliary transport 7% 7% 4% 7% 2% 15% 11% activities; activities of travel agencies Tanning and dress- ing of leather; manufacture of 2% 2% 1% 6% 1% 5% 6% luggage, handbags, saddlery, harness and footwear Water transport 4% 1% 6% 17% 1% 44% 3% Wholesale trade and commission trade, except of 16% 21% 11% 17% 8% 7% 17% motor vehicles and motorcycles Source: The World Bank, Enterprise Survey. Note: This table uses two-digits ISIC Rev 3.1 codes. > > > T A B L E D . 2 - Country-Level Participation Rates in Public Procurement, Overall and by Firm Size Participation rates in Public Procurement Country Overall Small Medium Large Afghanistan 13% 11% 18% 18% Albania 11% 9% 15% 18% Angola 7% 3% 7% 28% Antigua and Barbuda 14% 14% 15% 18% Argentina 12% 7% 21% 18% Armenia 10% 9% 9% 16% Austria 8% 7% 12% 12% Azerbaijan 25% 18% 30% 42% Bahamas 23% 21% 22% 37% Bangladesh 4% 6% 3% 3% Barbados 18% 15% 24% 22% Belarus 30% 26% 34% 41% Belgium 17% 14% 24% 27% Belize 15% 14% 16% 17% Benin 30% 16% 40% 49% Bhutan 33% 27% 45% 62% Bolivia 23% 20% 29% 35% Bosnia and Herzegovina 14% 14% 17% 7% Botswana 53% 50% 58% 55% Brazil 13% 12% 16% 11% Bulgaria 5% 3% 8% 23% Burkina Faso 42% 35% 54% 62% Burundi 47% 37% 64% 37% Cambodia 10% 9% 14% 5% Cameroon 21% 16% 36% 41% Cape Verde 5% 5% 6% 18% Central African Republic 23% 23% 24% 10% Chad 16% 16% 10% 51% Chile 16% 20% 17% 12% China 10% 7% 12% 20% Colombia 28% 26% 32% 23% Congo 29% 22% 43% 14% Costa Rica 25% 18% 28% 54% Côte d'Ivoire 18% 12% 33% 26% Croatia 8% 7% 8% 22% Cyprus 15% 16% 9% 28% Czech Republic 29% 28% 33% 24% Denmark 27% 21% 37% 46% Djibouti 12% 5% 18% 56% Dominica 12% 13% 11% 0% Dominican Republic 8% 7% 11% 9% DRC 10% 7% 23% 11% Participation rates in Public Procurement Country Overall Small Medium Large Ecuador 27% 23% 30% 39% Egypt 5% 2% 9% 20% El Salvador 14% 12% 15% 39% Eritrea 12% 13% 8% 12% Estonia 35% 36% 31% 41% Eswatini 13% 7% 21% 32% Ethiopia 36% 27% 54% 42% Fiji 14% 13% 13% 27% Finland 18% 15% 20% 39% Gabon 17% 15% 16% 35% Gambia 34% 36% 27% 25% Georgia 12% 12% 11% 17% Ghana 24% 21% 29% 31% Greece 15% 14% 16% 23% Grenada 18% 18% 17% 24% Guatemala 13% 11% 16% 19% Guinea 20% 20% 20% 23% Guyana 41% 43% 41% 39% Honduras 14% 8% 28% 29% Hungary 7% 9% 5% 1% India 10% 11% 9% 12% Indonesia 5% 5% 7% 6% Iraq 12% 12% 19% 3% Ireland 17% 15% 23% 23% Israel 20% 17% 27% 33% Italy 3% 3% 4% 6% Jamaica 12% 9% 15% 32% Jordan 22% 16% 28% 36% Kazakhstan 21% 19% 21% 47% Kenya 25% 22% 29% 38% Kosovo 13% 9% 28% 40% Kyrgyz Republic 24% 22% 28% 24% Lao PDR 9% 6% 11% 70% Latvia 39% 33% 52% 42% Lebanon 11% 8% 17% 8% Lesotho 13% 8% 19% 53% Liberia 25% 22% 16% 79% Lithuania 15% 11% 28% 21% Luxembourg 26% 16% 33% 65% Madagascar 6% 4% 8% 8% Malawi 27% 25% 33% 23% Malaysia 16% 12% 28% 21% Mali 51% 42% 61% 71% Malta 34% 27% 43% 32% Participation rates in Public Procurement Country Overall Small Medium Large Mauritania 35% 43% 28% 25% Mauritius 15% 11% 21% 22% Mexico 12% 10% 16% 18% Micronesia 18% 14% 25% - Moldova 36% 37% 34% 34% Mongolia 33% 26% 49% 46% Montenegro 25% 18% 36% 62% Morocco 20% 18% 22% 24% Mozambique 23% 17% 39% 26% Myanmar 7% 6% 11% 11% Namibia 28% 24% 41% 56% Nepal 3% 2% 9% 1% Netherlands 15% 14% 14% 35% Nicaragua 21% 10% 43% 12% Niger 21% 16% 27% 39% Nigeria 13% 12% 14% 20% North Macedonia 3% 2% 5% 13% Pakistan 13% 8% 18% 15% Panama 1% 2% 0% 2% Papua New Guinea 32% 20% 22% 50% Paraguay 17% 15% 16% 24% Peru 20% 18% 19% 33% Philippines 11% 11% 13% 7% Poland 15% 15% 16% 20% Portugal 12% 9% 21% 15% Romania 16% 14% 20% 22% Russian Federation 17% 15% 22% 16% Rwanda 17% 12% 23% 59% Samoa 37% 31% 47% 74% Senegal 21% 14% 29% 47% Serbia 31% 28% 34% 44% Sierra Leone 15% 14% 19% 44% Slovak Republic 4% 2% 8% 12% Slovenia 33% 29% 47% 27% Solomon Islands 24% 19% 26% 34% South Africa 2% 2% 2% 2% South Sudan 17% 15% 30% 37% Sri Lanka 11% 10% 16% 14% St Kitts and Nevis 21% 19% 22% 32% St Lucia 15% 18% 8% 13% St Vincent and Grenadines 21% 19% 32% 15% Sudan 20% 13% 28% 44% Suriname 6% 3% 12% 20% Sweden 16% 11% 18% 56% Participation rates in Public Procurement Country Overall Small Medium Large Tajikistan 11% 8% 18% 12% Tanzania 4% 3% 6% 20% Thailand 5% 4% 8% 1% Timor-Leste 36% 34% 41% 30% Togo 30% 22% 39% 42% Tonga 15% 12% 53% - Trinidad and Tobago 22% 19% 28% 40% Tunisia 18% 19% 15% 24% Turkey 11% 9% 15% 20% Uganda 8% 5% 15% 40% Ukraine 16% 14% 18% 28% Uruguay 18% 16% 18% 45% Uzbekistan 12% 9% 19% 23% Vanuatu 21% 17% 27% 50% Venezuela 13% 8% 38% 6% Vietnam 20% 16% 26% 20% West Bank and Gaza 11% 10% 17% 20% Yemen 16% 15% 12% 48% Zambia 25% 21% 26% 33% Zimbabwe 17% 15% 23% 26% Source: The World Bank, Enterprise Survey. > > > T A B L E D . 3 - Procurement Participation and SME Status Dependent variable: Over the last year, has this establishment secured or attempted to secure a government contract? Global ECA EAP LAC MENA SA SSA (1) (2) (3) (4) (5) (6) (7) ⇤⇤⇤ ⇤⇤⇤ ⇤⇤⇤ ⇤⇤⇤ ⇤⇤⇤ Small Enterprises 11.982 12.351 13.845 7.179 16.194 0.767 15.545⇤⇤⇤ (0.538) (1.046) (1.413) (1.323) (1.843) (1.048) (1.364) Medium Enterprises 5.365⇤⇤⇤ 6.931⇤⇤⇤ 6.834⇤⇤⇤ 2.073 6.916⇤⇤⇤ 2.053⇤⇤ 6.782⇤⇤⇤ (0.557) (1.089) (1.422) (1.359) (1.922) (1.043) (1.409) Sector fixed e↵ects Yes Yes Yes Yes Yes Yes Yes Controls Yes Yes Yes Yes Yes Yes Yes Observations 78,311 23,489 9,624 12,443 7,523 12,004 13,228 R2 0.126 0.128 0.147 0.107 0.109 0.207 0.146 Adjusted R2 0.124 0.125 0.142 0.102 0.103 0.204 0.141 Source: World Bank Enterprise Survey. Note: * p < 0.1; ** p < 0.05; *** p < 0.01. Ordinary Least Square. Dependent variable: it takes value 100 if the establishment has secured or attempted to secure a government contract in the 12 months prior to the interview, otherwise 0. Controls: firm age (log), a dummy variable if the firm is run by a woman, productivity computed as sales per worker, country fixed effects and sector fixed effects (two-digits ISIC Rev. 3.1 codes). The baseline is given by “Large Enterprises” (omitted category). Results are weighted so that the sample is representative within each country, and each country is equally considered in the estimations. Six different samples: (1) Global pooled sample, (2) Europe and Central Asia, (3) East Asia and the Pacific; (4) Latin America and the Caribbean; (5) Middle East and North Africa; (6) South Asia; (7) Sub-Saharan Africa. Robust standard errors clustered at firm size, sector and country level. > > > T A B L E D . 4 - Procurement Participation and SME Status, with Sector Specific Definition of SMEs Dependent variable: Over the last year, has this establishment secured or attempted to secure a government contract? Global ECA EAP LAC MENA SA SSA (1) (2) (3) (4) (5) (6) (7) ⇤⇤⇤ ⇤⇤⇤ ⇤⇤⇤ ⇤⇤⇤ ⇤⇤⇤ Small Enterprises 10.785 11.042 8.894 7.825 11.891 1.374 15.074⇤⇤⇤ (0.404) (0.728) (1.127) (0.986) (1.329) (0.857) (1.081) Medium Enterprises 4.184⇤⇤⇤ 4.996⇤⇤⇤ 3.164⇤⇤⇤ 0.946 5.421⇤⇤⇤ 0.709 7.115⇤⇤⇤ (0.404) (0.743) (1.082) (0.976) (1.345) (0.818) (1.070) Sector fixed e↵ects Yes Yes Yes Yes Yes Yes Yes Controls Yes Yes Yes Yes Yes Yes Yes Observations 78,311 23,489 9,624 12,443 7,523 12,004 13,228 R2 0.127 0.131 0.143 0.110 0.105 0.206 0.148 Adjusted R2 0.125 0.128 0.138 0.105 0.099 0.203 0.143 Source: World Bank Enterprise Survey. Note: * p < 0.1; ** p < 0.05; *** p < 0.01. Ordinary Least Square. Dependent variable: it takes value 100 if the establishment has secured or attempted to secure a government contract in the 12 months prior to the interview, otherwise 0. Controls: firm age (log), a dummy variable if the firm is run by a woman, productivity computed as sales per worker, country fixed effects and sector fixed effects (two-digits ISIC Rev. 3.1 codes). The baseline is given by “Large Enterprises” (omitted category). The definition of small, medium and large enterprises is specific to each sector and based on the distribution in the number of employees for each sector: “Small Enterprises” are firms in the first tercile in distribution of number of employees within each sector, “Medium enterprises” are firms in the second tercile, and “Large Enterprises” are firms in the third tercile. Results are weighted so that the sample is representative within each country, and each country is equally considered in the estimations. Six different samples: (1) Global pooled sample, (2) Europe and Central Asia, (3) East Asia and the Pacific; (4) Latin America and the Caribbean; (5) Middle East and North Africa; (6) South Asia; (7) Sub-Saharan Africa. Robust standard errors clustered at firm size, sector and country level. > > > T A B L E D . 5 - Procurement Participation and SME Status, Focus on Two Topical Sectors Dependent variable: Over the last year, has this establishment secured or attempted to secure a government contract? Manufacture of Basic Metals Computer and Related Activities (1) (2) Small Enterprises 17.986⇤⇤⇤ 13.093⇤⇤ (2.799) (5.134) Medium Enterprises 0.560 1.018 (2.917) (5.461) Controls Yes Yes Observations 1,425 1,551 R2 0.058 0.019 Adjusted R2 0.055 0.015 Source: The World Bank, Enterprise Survey. Note: * p < 0.1; ** p < 0.05; *** p < 0.01. Ordinary Least Square. Dependent variable: it takes value 100 if the establishment has secured or attempted to secure a government contract in the 12 months prior to the interview, otherwise 0. Controls: firm age (log), a dummy variable if the firm is run by a woman, productivity computed as sales per worker. The baseline is given by “Large Enterprises” (omitted category). Participation rate in public procurement is relatively high in both sectors with a share of 18% and 27%, respectively. Manufacture of Basic Models has only 30% of small and 40% of medium enterprises compared to Computer and Related Activities with 40% of small enterprises and 36% of medium enterprises. Robust standard errors clustered at firm size and country level. > > > T A B L E D . 6 - Procurement Participation and Firm Characteristics Dependent variable: Over the last year, has this establishment secured or attempted to secure a government contract? Global ECA EAP LAC MENA SA SSA (1) (2) (3) (4) (5) (6) (7) ⇤⇤⇤ ⇤⇤⇤ ⇤⇤ ⇤⇤ ⇤⇤⇤ Log(Firm Age) 0.828 1.274 1.230 0.305 1.284 3.761 0.803⇤ (0.181) (0.305) (0.524) (0.623) (0.600) (0.336) (0.436) Log(Employees) 4.292⇤⇤⇤ 4.037⇤⇤⇤ 3.798⇤⇤⇤ 4.153⇤⇤⇤ 5.387⇤⇤⇤ 0.787⇤⇤⇤ 5.288⇤⇤⇤ (0.128) (0.223) (0.350) (0.416) (0.446) (0.217) (0.325) Female-run 0.634⇤ 3.445⇤⇤⇤ 0.853 2.862⇤⇤ 8.313⇤⇤⇤ 0.377 1.080 (0.360) (0.602) (0.811) (1.128) (1.673) (0.887) (0.924) Capital City 2.899⇤⇤⇤ 2.507⇤⇤⇤ 0.887 4.889⇤⇤⇤ 0.170 2.629⇤⇤⇤ 4.074⇤⇤⇤ (0.335) (0.558) (0.937) (0.989) (1.272) (0.690) (0.855) Sector fixed e↵ects Yes Yes Yes Yes Yes Yes Yes Observations 76,413 25,846 9,847 6,656 6,757 12,951 14,356 R2 0.128 0.131 0.143 0.123 0.108 0.089 0.147 Adjusted R2 0.126 0.129 0.139 0.117 0.102 0.086 0.143 Source: The World Bank, Enterprise Survey. Note: * p < 0.1; ** p < 0.05; *** p < 0.01. Ordinary Least Square. Dependent variable: it takes value 100 if the establishment has secured or attempted to secure a government contract in the 12 months prior to the interview, otherwise 0. Controls: country fixed effects and sector fixed effects (two-digits ISIC Rev. 3.1 codes). Results are weighted so that the sample is representative within each country, and each country is equally considered in the estimations. Six different samples: (1) Global pooled sample, (2) Europe and Central Asia, (3) East Asia and the Pacific; (4) Latin America and the Caribbean; (5) Middle East and North Africa; (6) South Asia; (7) Sub-Saharan Africa. Robust standard errors clustered at firm size, sector and country level. > > > T A B L E D . 7 - Procurement Participation and eGP System Dependent variable: Over the last year, has this establishment secured or attempted to secure a government contract? Global Is there an e-Procurement portal? 6.664⇤⇤⇤ (2.581) Controls Yes Observations 135 R2 0.156 Adjusted R2 0.130 Source: The World Bank, Enterprise Survey. Note: * p < 0.1; ** p < 0.05; *** p < 0.01. Ordinary Least Square at country level. Dependent variable: share of firms in the country that secured or attempted to secure a government contract in the 12 months prior to the interview. The main explanatory variable is an indicator (as of 2022) of whether the country has a functioning e-Procurement portal, available from the World Bank GTMI dataset (World Bank, 2022e). Controls: the log of GDP per capita, “Government Effectiveness” (World Governance Indicators), and “Control of corruption” (World Governance Indicators). Robust standard errors. > > > T A B L E D . 8 - Country-Level Participation Rates in Public Procurement, First and Latest WBES Waves Participation rate in Public Procurement Country Region First wave Last wave Afghanistan SA 2008 22% 2014 13% Albania ECA 2007 14% 2019 11% Armenia ECA 2009 24% 2020 10% Azerbaijan ECA 2009 18% 2019 25% Bangladesh SA 2007 14% 2013 4% Belarus ECA 2008 29% 2018 30% Benin SSA 2009 40% 2016 30% Bhutan SA 2009 43% 2015 33% Bosnia and Herzegovina ECA 2009 22% 2019 14% Bulgaria ECA 2007 19% 2019 5% Cameroon SSA 2009 15% 2016 21% Chad SSA 2009 21% 2018 16% Côte d'Ivoire SSA 2009 9% 2016 18% Croatia ECA 2007 23% 2019 8% Czech Republic ECA 2009 24% 2019 29% Dominican Republic LAC 2010 15% 2016 8% Egypt MENA 2013 5% 2020 5% Estonia ECA 2009 33% 2019 35% Ethiopia SSA 2011 22% 2015 36% Georgia ECA 2008 21% 2019 12% Ghana SSA 2007 16% 2013 24% Hungary ECA 2009 5% 2019 7% Indonesia EAP 2009 5% 2015 5% Jordan MENA 2013 12% 2019 22% Kazakhstan ECA 2009 18% 2019 21% Kosovo ECA 2009 11% 2019 13% Kyrgyz Republic ECA 2009 22% 2019 24% Lao PDR EAP 2009 0% 2018 9% Latvia ECA 2009 37% 2019 39% Lebanon MENA 2013 11% 2019 11% Lesotho SSA 2009 42% 2016 13% Liberia SSA 2009 17% 2017 25% Lithuania ECA 2009 15% 2019 15% Madagascar SSA 2009 24% 2013 6% Malawi SSA 2009 26% 2014 27% Mali SSA 2007 11% 2016 51% Moldova ECA 2009 27% 2019 36% Mongolia EAP 2009 40% 2019 33% Montenegro ECA 2009 6% 2019 25% Morocco MENA 2013 21% 2019 20% Mozambique SSA 2007 16% 2018 23% Myanmar EAP 2014 2% 2016 7% Participation rate in Public Procurement Country Region First wave Last wave Nepal SA 2009 8% 2013 3% Niger SSA 2009 42% 2017 21% North Macedonia ECA 2009 10% 2019 3% Philippines EAP 2009 9% 2015 11% Poland ECA 2009 26% 2019 15% Romania ECA 2009 13% 2019 16% Russian Federation ECA 2009 36% 2019 17% Senegal SSA 2007 6% 2014 21% Serbia ECA 2009 8% 2019 31% Slovak Republic ECA 2009 4% 2019 4% Slovenia ECA 2009 29% 2019 33% South Africa SSA 2007 9% 2020 2% Suriname LAC 2010 9% 2018 6% Sweden ECA 2014 16% 2020 16% Tajikistan ECA 2008 29% 2019 11% Timor-Leste EAP 2009 46% 2015 36% Togo SSA 2009 19% 2016 30% Tunisia MENA 2013 30% 2020 18% Uzbekistan ECA 2008 10% 2019 12% Vietnam EAP 2009 20% 2015 20% West Bank and Gaza MENA 2013 12% 2019 11% Zambia SSA 2007 19% 2019 25% Zimbabwe SSA 2011 23% 2016 17% Source: The World Bank, Enterprise Survey. > > > D . 2 Follow Up Sample > > > > > > F I G U R E D . 1 - Main Administrative Obstacle F I G U R E D . 2 - Main Reason for Lack of Trust in before Contract Signature by Firm Size (Croatia) the Process by Firm Size (Croatia) Source: The World Bank, Enterprise Survey. Source: The World Bank, Enterprise Survey. Note: N = 197. Statistics are computed using follow-up sample weights Note: N = 120. Statistics are computed using follow-up sample weights (wmedian COVID). This question was asked only to firms that indicated that (wmedian COVID). This question was asked only to firms that indicated that administrative procedures before contract signature are an obstacle. lack of trust is an obstacle. > > > > > > F I G U R E D . 3 - Main Geographical Barrier to F I G U R E D . 4 - Biggest Challenge that is Expected Secure a Government Contract by Firm Size When Working Under a Government Contract by (Croatia) Firm Size (Croatia) Source: The World Bank, Enterprise Survey. Source: The World Bank, Enterprise Survey. Note: N = 60. Statistics are computed using follow-up sample weights Note: N = 121. Statistics are computed using follow-up sample weights (wmedian COVID). This question was asked only to firms that indicated that (wmedian COVID). This question was asked only to firms that indicated geographical barriers are an obstacle. expected challenges when working under a government contract are an obstacle. > > > > > > F I G U R E D . 5 - Main Administrative Obstacle F I G U R E D . 6 - Main Reason for Lack of Trust in Before Contract Signature by Firm Size (Poland) the Process by Firm Size (Poland) Source: The World Bank, Enterprise Survey. Source: The World Bank, Enterprise Survey. Note: N = 459. Statistics are computed using follow-up sample weights Note: N = 429. Statistics are computed using follow-up sample weights (wmedian COVID). This question was asked only to firms that indicated that (wmedian COVID). This question was asked only to firms that indicated that administrative procedures before contract signature are an obstacle. lack of trust is an obstacle. > > > > > > F I G U R E D . 8 - Biggest Challenge that is Expected F I G U R E D . 7 - Main Geographical Barrier to When Working Under a Government Contract by Secure a Government Contract by Firm Size (Poland) Firm Size (Poland) Source: The World Bank, Enterprise Survey. Source: The World Bank, Enterprise Survey. Note: N = 387. Statistics are computed using follow-up sample weights Note: N = 465. Statistics are computed using follow-up sample weights (wmedian COVID). This question was asked only to firms that indicated that (wmedian COVID). This question was asked only to firms that indicated that geographical barriers are an obstacle. expected challenges when working under a government contract are an obstacle. > > > > > > F I G U R E D . 9 - Main Administrative Obstacle F I G U R E D . 1 0 - Main Reason for Lack of Trust in Before Contract Signature by Firm Size (Romania) the Process by Firm Size (Romania) Source: The World Bank, Enterprise Survey. Source: The World Bank, Enterprise Survey. Note: N = 128. Statistics are computed using follow-up sample weights Note: N = 121. Statistics are computed using follow-up sample weights (wmedian COVID). This question was asked only to firms that indicated that (wmedian COVID). This question was asked only to firms that indicated that administrative procedures before contract signature are an obstacle. lack of trust is an obstacle. > > > > > > F I G U R E D . 1 1 - Main Geographical Barrier F I G U R E D . 1 2 - Biggest Challenge that is to Secure a Government Contract by Firm Size Expected When Working Under a Government (Romania) Contract by Firm Size (Romania) Source: The World Bank, Enterprise Survey. Source: The World Bank, Enterprise Survey. Note: N = 83. Statistics are computed using follow-up sample weights Note: N = 109. Statistics are computed using follow-up sample weights (wmedian COVID. This question was asked only to firms that indicated that (wmedian COVID). This question was asked only to firms that indicated that geographical barriers are an obstacle. expected challenges when working under a government contract are an obstacle. > > > T A B L E D . 9 - Perception of Obstacles and Participation Rates - Croatia Dependent variable: To what degree are each of the following an obstacle to attempting to secure a government contract? (Croatia) Administrative Trust Geography Contract Execution (1) (2) (3) (4) ⇤⇤ ⇤⇤⇤ Participated in Public Procurement 0.172 0.063 0.181 0.094 (0.074) (0.074) (0.060) (0.072) Sector fixed e↵ects Yes Yes Yes Yes Controls Yes Yes Yes Yes Observations 324 324 324 324 R2 0.077 0.048 0.038 0.049 Adjusted R2 0.060 0.030 0.019 0.031 Source: The World Bank, Enterprise Survey. Note: * p < 0.1; ** p < 0.05; *** p < 0.01. Ordinary Least Square. Dependent variables: To what degree are each of the following an obstacle to attempting to secure a government contract, on a scale between 0 (no obstacle) and 4 (very severe obstacle)? – (1) Administrative procedures before contract signature (q3a); (2) Lack of trust in the process (q3b); (3) Geographical barriers (q3c); (4) Expected challenges during contract execution (q3d). Each dependent variable tables value 0 if the answer is “no obstacle”, and 1 if the answer is between “Minor obstacle” to “Very severe obstacle”’. Controls: firm age (log), number of employees (log), a dummy variable if the firm is run by a woman, country fixed effects and sector fixed effects (two-digits ISIC Rev. 3.1 codes). Results are weighted so that the sample is representative at national level. The sample refers to only Croatian firms. Robust standard errors clustered at firm size and sector level. > > > T A B L E D . 1 0 - Perception of Obstacles and Participation Rates - Poland Dependent variable: To what degree are each of the following an obstacle to attempting to secure a government contract? (Poland) Administrative Trust Geography Contract Execution (1) (2) (3) (4) ⇤⇤⇤ ⇤⇤⇤ ⇤⇤⇤ Participated in Public Procurement 0.095 0.085 0.139 0.095⇤⇤⇤ (0.028) (0.030) (0.037) (0.031) Sector fixed e↵ects Yes Yes Yes Yes Controls Yes Yes Yes Yes Observations 563 567 564 563 R2 0.057 0.090 0.076 0.117 Adjusted R2 0.047 0.080 0.066 0.107 Source: The World Bank, Enterprise Survey. Note: * p < 0.1; ** p < 0.05; *** p < 0.01. Ordinary Least Square. Dependent variables: To what degree are each of the following an obstacle to attempting to secure a government contract, on a scale between 0 (no obstacle) and 4 (very severe obstacle)? – (1) Administrative procedures before contract signature (q3a); (2) Lack of trust in the process (q3b); (3) Geographical barriers (q3c); (4) Expected challenges during contract execution (q3d). Each dependent variable tables value 0 if the answer is “no obstacle”, and 1 if the answer is between “Minor obstacle” to “Very severe obstacle”’. Controls: firm age (log), number of employees (log), a dummy variable if the firm is run by a woman, country fixed effects and sector fixed effects (two-digits ISIC Rev. 3.1 codes). Results are weighted so that the sample is representative at national level. The sample refers to only Polish firms. Robust standard errors clustered at firm size and sector level. > > > T A B L E D . 1 1 - Perception of Obstacles and Participation Rates - Romania Dependent variable: To what degree are each of the following an obstacle to attempting to secure a government contract? (Romania) Administrative Trust Geography Contract Execution (1) (2) (3) (4) ⇤ Participated in Public Procurement 0.009 0.125 0.051 0.101 (0.077) (0.073) (0.076) (0.075) Sector fixed e↵ects Yes Yes Yes Yes Controls Yes Yes Yes Yes Observations 251 264 255 254 R2 0.023 0.076 0.034 0.039 Adjusted R2 0.001 0.055 0.011 0.015 Source: The World Bank, Enterprise Survey. Note: * p < 0.1; ** p < 0.05; *** p < 0.01. Ordinary Least Square. Dependent variables: To what degree are each of the following an obstacle to attempting to secure a government contract, on a scale between 0 (no obstacle) and 4 (very severe obstacle)? – (1) Administrative procedures before contract signature (q3a); (2) Lack of trust in the process (q3b); (3) Geographical barriers (q3c); (4) Expected challenges during contract execution (q3d). Each dependent variable tables value 0 if the answer is “no obstacle”, and 1 if the answer is between “Minor obstacle” to “Very severe obstacle”’. Controls: firm age (log), number of employees (log), a dummy variable if the firm is run by a woman, country fixed effects and sector fixed effects (two-digits ISIC Rev. 3.1 codes). Results are weighted so that the sample is representative at national level. The sample refers to only Romanian firms. Robust standard errors clustered at firm size and sector level. > > > T A B L E D . 1 2 - Perception of Obstacles and Oarticipation Rates – Croatia (Robustness Check) Dependent variable: To what degree are each of the following an obstacle to attempting to secure a government contract? (Croatia) Administrative Trust Geography Contract Execution (1) (2) (3) (4) Participated in Public Procurement 0.024 0.056 0.048 0.066⇤ (0.048) (0.035) (0.032) (0.036) Sector fixed e↵ects Yes Yes Yes Yes Controls Yes Yes Yes Yes Observations 324 324 324 324 R2 0.012 0.016 0.023 0.023 Adjusted R2 0.006 0.003 0.004 0.005 Source: The World Bank, Enterprise Survey. Note: * p < 0.1; ** p < 0.05; *** p < 0.01. Ordinary Least Square. Dependent variables: To what degree are each of the following an obstacle to attempting to secure a government contract, on a scale between 0 (no obstacle) and 4 (very severe obstacle)? – (1) Administrative procedures before contract signature (q3a); (2) Lack of trust in the process (q3b); (3) Geographical barriers (q3c); (4) Expected challenges during contract execution (q3d). Each dependent variable tables value 0 if the answer is “no obstacle”, and 1 if the answer is between “Major obstacle” to “Very severe obstacle”’. Controls: firm age (log), number of employees (log), a dummy variable if the firm is run by a woman, country fixed effects and sector fixed effects (two-digits ISIC Rev. 3.1 codes). Results are weighted so that the sample is representative at national level. The sample refers to only Croatian firms. Robust standard errors clustered at firm size and sector level. > > > T A B L E D . 1 3 - Perception of Obstacles and Participation Rates – Poland (Robustness Check) Dependent variable: To what degree are each of the following an obstacle to attempting to secure a government contract? (Poland) Administrative Trust Geography Contract Execution (1) (2) (3) (4) ⇤⇤⇤ ⇤⇤ Participated in Public Procurement 0.015 0.167 0.070 0.071⇤ (0.041) (0.039) (0.035) (0.039) Sector fixed e↵ects Yes Yes Yes Yes Controls Yes Yes Yes Yes Observations 563 567 564 563 R2 0.060 0.044 0.052 0.027 Adjusted R2 0.050 0.033 0.042 0.016 Source: The World Bank, Enterprise Survey. Note: * p < 0.1; ** p < 0.05; *** p < 0.01. Ordinary Least Square. Dependent variables: To what degree are each of the following an obstacle to attempting to secure a government contract, on a scale between 0 (no obstacle) and 4 (very severe obstacle)? – (1) Administrative procedures before contract signature (q3a); (2) Lack of trust in the process (q3b); (3) Geographical barriers (q3c); (4) Expected challenges during contract execution (q3d). Each dependent variable tables value 0 if the answer is “no obstacle”, and 1 if the answer is between “Major obstacle” to “Very severe obstacle”’. Controls: firm age (log), number of employees (log), a dummy variable if the firm is run by a woman, country fixed effects and sector fixed effects (two-digits ISIC Rev. 3.1 codes). Results are weighted so that the sample is representative at national level. The sample refers to only Polish firms. Robust standard errors clustered at firm size and sector level. > > > T A B L E D . 1 4 - Perception of Obstacles and Participation Rates – Romania (Robustness Check) Dependent variable: To what degree are each of the following an obstacle to attempting to secure a government contract? (Romania) Administrative Trust Geography Contract Execution (1) (2) (3) (4) ⇤ Participated in Public Procurement 0.048 0.052 0.067 0.012 (0.040) (0.035) (0.036) (0.035) Sector fixed e↵ects Yes Yes Yes Yes Controls Yes Yes Yes Yes Observations 251 264 255 254 R2 0.238 0.090 0.165 0.135 Adjusted R2 0.219 0.069 0.144 0.114 Source: The World Bank, Enterprise Survey. Note: * p < 0.1; ** p < 0.05; *** p < 0.01. Ordinary Least Square. Dependent variables: To what degree are each of the following an obstacle to attempting to secure a government contract, on a scale between 0 (no obstacle) and 4 (very severe obstacle)? – (1) Administrative procedures before contract signature (q3a); (2) Lack of trust in the process (q3b); (3) Geographical barriers (q3c); (4) Expected challenges during contract execution (q3d). Each dependent variable tables value 0 if the answer is “no obstacle”, and 1 if the answer is between “Major obstacle” to “Very severe obstacle”’. Controls: firm age (log), number of employees (log), a dummy variable if the firm is run by a woman, country fixed effects and sector fixed effects (two-digits ISIC Rev. 3.1 codes). Results are weighted so that the sample is representative at national level. The sample refers to only Romanian firms. Robust standard errors clustered at firm size and sector level. >>> Annex E – Sample Description > > > E . 1 Global Sample > > > T A B L E E . 1 - Standard WBES, Latest Available Wave for Each Country Country Wave Region N Country Wave Region N Afghanistan 2014 SA 410 Latvia 2019 ECA 359 Albania 2019 ECA 377 Lebanon 2019 MENA 532 Angola 2010 SSA 360 Lesotho 2016 SSA 150 Antigua and Bar- 2010 LAC 151 Liberia 2017 SSA 151 buda Argentina 2017 LAC 991 Lithuania 2019 ECA 358 Armenia 2020 ECA 546 Luxembourg 2020 ECA 170 Austria 2021 ECA 600 Madagascar 2013 SSA 532 Azerbaijan 2019 ECA 225 Malawi 2014 SSA 523 Bahamas 2010 LAC 150 Malaysia 2015 EAP 1,000 Bangladesh 2013 SA 1,442 Mali 2016 SSA 185 Barbados 2010 LAC 150 Malta 2019 MENA 242 Belarus 2018 ECA 600 Mauritania 2014 SSA 150 Belgium 2020 ECA 614 Mauritius 2009 SSA 398 Belize 2010 LAC 150 Mexico 2010 LAC 1,480 Benin 2016 SSA 150 Micronesia 2009 EAP 68 Bhutan 2015 SA 253 Moldova 2019 ECA 360 Bolivia 2017 LAC 364 Mongolia 2019 EAP 360 Bosnia and Herze- 2019 ECA 362 Montenegro 2019 ECA 150 govina Botswana 2010 SSA 268 Morocco 2019 MENA 1,096 Brazil 2009 LAC 1,802 Mozambique 2018 SSA 601 Bulgaria 2019 ECA 772 Myanmar 2016 EAP 607 Burkina Faso 2009 SSA 394 Namibia 2014 SSA 580 Burundi 2014 SSA 157 Nepal 2013 SA 482 Cambodia 2016 EAP 373 Netherlands 2020 ECA 808 Cameroon 2016 SSA 361 Nicaragua 2016 LAC 333 Cape Verde 2009 SSA 156 Niger 2017 SSA 151 Central African 2011 SSA 150 Nigeria 2014 SSA 2,676 Republic Chad 2018 SSA 153 North Macedonia 2019 ECA 360 Chile 2010 LAC 1,033 Pakistan 2013 SA 1,247 China 2012 EAP 2,700 Panama 2010 LAC 365 Papua New Gui- Colombia 2017 LAC 993 2015 EAP 65 nea Congo 2009 SSA 151 Paraguay 2017 LAC 364 Costa Rica 2010 LAC 538 Peru 2017 LAC 1,003 Côte d'Ivoire 2016 SSA 361 Philippines 2015 EAP 1,335 Croatia 2019 ECA 404 Poland 2019 ECA 1,369 Cyprus 2019 ECA 240 Portugal 2019 ECA 1,062 Czech Republic 2019 ECA 502 Romania 2019 ECA 814 Russian Federa- Denmark 2020 ECA 995 2019 ECA 1,323 tion Country Wave Region N Country Wave Region N Djibouti 2013 MENA 266 Rwanda 2019 SSA 360 Dominica 2010 LAC 150 Samoa 2009 EAP 109 Dominican Republic 2016 LAC 359 Senegal 2014 SSA 601 DRC 2013 SSA 529 Serbia 2019 ECA 361 Ecuador 2017 LAC 361 Sierra Leone 2017 SSA 152 Egypt 2020 MENA 3,075 Slovak Republic 2019 ECA 429 El Salvador 2016 LAC 719 Slovenia 2019 ECA 409 Eritrea 2009 SSA 179 Solomon Islands 2015 EAP 151 Estonia 2019 ECA 360 South Africa 2020 SSA 1,097 Eswatini 2016 SSA 150 South Sudan 2014 SSA 738 Ethiopia 2015 SSA 848 Sri Lanka 2011 SA 610 Fiji 2009 EAP 164 St Kitts and Nevis 2010 LAC 150 Finland 2020 ECA 759 St Lucia 2010 LAC 150 St Vincent and Gabon 2009 SSA 179 2010 LAC 154 Grenadines Gambia 2018 SSA 151 Sudan 2014 SSA 662 Georgia 2019 ECA 581 Suriname 2018 LAC 233 Ghana 2013 SSA 720 Sweden 2020 ECA 591 Greece 2018 ECA 600 Tajikistan 2019 ECA 352 Grenada 2010 LAC 153 Tanzania 2013 SSA 813 Guatemala 2017 LAC 345 Thailand 2016 EAP 1,000 Guinea 2016 SSA 150 Timor-Leste 2015 EAP 126 Guyana 2010 LAC 165 Togo 2016 SSA 150 Honduras 2016 LAC 332 Tonga 2009 EAP 150 Trinidad and To- Hungary 2019 ECA 805 2010 LAC 370 bago India 2014 SA 9,281 Tunisia 2020 MENA 615 Indonesia 2015 EAP 1,320 Turkey 2019 ECA 1,663 Iraq 2011 MENA 756 Uganda 2013 SSA 762 Ireland 2020 ECA 606 Ukraine 2019 ECA 1,337 Israel 2013 MENA 483 Uruguay 2017 LAC 347 Italy 2019 ECA 760 Uzbekistan 2019 ECA 1,239 Jamaica 2010 LAC 376 Vanuatu 2009 EAP 128 Jordan 2019 MENA 601 Venezuela 2010 LAC 320 Kazakhstan 2019 ECA 1,446 Vietnam 2015 EAP 996 West Bank and Kenya 2018 SSA 1,001 2019 MENA 365 Gaza Kosovo 2019 ECA 271 Yemen 2013 MENA 353 Kyrgyz Republic 2019 ECA 360 Zambia 2019 SSA 601 Lao PDR 2018 EAP 332 Zimbabwe 2016 SSA 600 Source: The World Bank, Enterprise Survey. > > > T A B L E E . 2 - Firm Characteristics from Standard WBES, Latest Available Wave for Each Country Variable Observations Mean FIRM AGE 16 EAST ASIA AND PACIFIC FEMALE-RUN 10,984 33% EMPLOYEES 60 FIRM AGE 20 EUROPE AND CENTRAL ASIA FEMALE-RUN 26,299 18% EMPLOYEES 73 FIRM AGE 22 LATIN AMERICA AND CARRIBEAN FEMALE-RUN 14,551 21% EMPLOYEES 48 FIRM AGE 21 MIDDLE EAST AND NORTH AFRICA FEMALE-RUN 8,384 7% EMPLOYEES 33 FIRM AGE 17 SOUTH ASIA FEMALE-RUN 13,725 11% EMPLOYEES 80 FIRM AGE 15 SUB-SAHARAN AFRICA FEMALE-RUN 19,351 16% EMPLOYEES 39 Source: The World Bank, Enterprise Survey. Note: Firm age is the time between the date of the survey and the date of firms’ establishment; employees is the total number of full-time equivalent workers in the establishment recorded on the date of the survey; and, the establishment is considered female-run if the respondent declares that the top-manager is female. > > > E . 2 Follow Up Sample > > > T A B L E E . 3 - WBES COVID Follow-Up,Ccountries with Procurement Module Country Wave Observations Croatia 2021 324 Poland 2021 983 Romania 2021 500 Source: The World Bank, Enterprise Survey Covid follow-up. > > > T A B L E E . 4 - Firm Characteristics from WBES 2019 and WBES COVID Follow-Up 2021, Countries with Procurement Module ES 2019 SAMPLE ES 2021 SAMPLE Observations Mean Observations Mean FIRM AGE 21 324 24 Croatia FEMALE-RUN 404 21% 97 26% EMPLOYEES 39 324 26 FIRM AGE 18 955 20 Poland FEMALE-RUN 1,362 27% 361 29 % EMPLOYEES 14 983 13 FIRM AGE 18 500 19 Romania FEMALE-RUN 814 17% 154 16% EMPLOYEES 26 500 21 Source: The World Bank, Enterprise Survey Covid follow-up. Note: Firm age is the time between the date of the survey and the date of firms’ establishment; employees is the total number of full-time equivalent workers in the establishment recorded on the date of the survey; and, the establishment is considered female-run if the respondent declares that the top-manager is female. >>> Annex F – Institutional Background for Croatia, Poland, and Romania As Croatia, Poland and Romania are members of the 2015 and May 2021 followed a closed procedure, and an open European Union (EU), their public procurement legislations procedure was used for the vast majority of public procurement. the values, principles and procedures provided by EU Despite these features, there are low levels of competition for Directives, namely - Directive 2014/24/EU, Directive 2014/25/ many categories of goods, services and works, and relatively EU, Directive 2014/23/EU. The next sections provide a short low levels of SME participation in public procurement overall. summary of the institutional context for public procurement in > > > these countries. > > > F . 2 Poland57 F . 1 Croatia66 The legal framework for public procurement in Poland was adopted following entry to the EU in 2004, and changes In Croatia, the public procurement regulation and institutional introduced in 2008 sought to increase the efficiency of setting include foundational elements for a transparency and public procurement. The Public Procurement Law also accountable public procurement system. The Procurement implements relevant EU directives (Classic Directive 2014/24/ Policy Directorate has overseen very substantial reforms EU, Sectoral Directive 2014/25/EU, Defense and Security of public procurement in Croatia in recent years. Public Directive 2009/81/EC and Remedies Directives) and is also procurement is now regulated by the Public Procurement Act bound by other acts of European law. In April 2008, incentives (the PP Act), ZJN 2016 (NN 120/2016), which was adopted for participation in public procurement of private sector were in December 2016 and came into effect on 1 January 2017. introduced: Using procedures other than open tenders, where The PP Act is fully aligned with the EU acquis communautaire justified; application of non-price criteria, where desirable; including Directive 2014/24/EU on public procurement, Directive application of green procurement; application of social 2014/25/EU on procurement by entities operating in the water, clauses; promotion of innovations; and support of the SMEs for energy, transport and postal services sectors. Invitations to bid electronic procurement, though the impact of these incentives must be published through the eProcurement system, which on procurement practices is assessed to have been limited. is considered one of the foundations of open and transparent To increase efficiency, additional adjustments were introduced procurement. Mandatory publication of procurement plans to the legal framework, including making mandatory the use was required under the Public Procurement Act 2011 and it of non-price criteria (since 2013); applying of certain social remained a requirement under the PP Act 2016, and this is clauses (e.g., obligation to hire workers on an employment an element contributing to ensure effective competition and contract); examining abnormally low prices; partitioning the transparency of public procurement. Article 28 of the PP procurement; electronic communication (the implementation Act requires contracting authorities/entities to maintain and of the 2018 Directives); and direct payments to subcontractors publish a register of public procurement contracts, which must in the event of unlawful non-payment by the general contractor. include information on subcontractors, the date of conclusion A new Public Procurement Law (approved in September 2019) of the contract, the total amount paid to the supplier under came into force on January 1, 2021, replacing the previous the contract and the reason for total payment being higher act that had been in force since 2004. The law implements than the original contract value where this occurred. There is thresholds and other regulations of the latest EU Directive a well-established complaint process for public procurement 2014/25/EU, among other changes. The legal framework in Croatia which allows bidders to raise concerns about fair establishes one of the most open and transparent public and equal treatment. The State Commission for Supervision procurement systems in Europe. The key body responsible of Public Procurement reviews and decides on complaints, for regulating and overseeing public procurement as well and publishes information on the outcomes of cases. as providing public procurement data is the Polish Public Procurement Office (PPO). Public procurement represents a substantial part of the Croatian economy and it is demonstrably open. Public procurement The new Public Procurement Law sets a requirement for the represents a substantial part of the Croatian economy. The preparation of four-year State Purchasing Strategy (SPS). The total value of contracts and framework agreements published first SPS, which will be implemented over the 2022-25 period, in 2019 was more than HRK 43 billion excluding VAT, or more intends to make a priority of ensuring that public procurement than 13 percent of Croatian GDP. Between 2015 and May is instrumental in achieving key strategic policy objectives. 2021, only 0.5 percent of procurement processes between 66. This section is based on a report on the performance of public procurement in Croatia (World Bank, 2022d) and the Croatia Country Economic Memorandum 2022 (World Bank, 2023b). 67. This section is based on the Concept Note document for the 2022 World Bank project “Poland RAS: Supporting Poland in implementation of its State Purchasing Policy” (P179835). These objectives include support to sustainable and innovative monitoring system in the public procurement area. procurement, improving access of micro enterprises and SMEs to the public procurement market, professionalization In accordance with the legal provisions in force, contracting of public procurement and enhancement of health aspects authorities have the obligation to send for publication a in public procurement including during COVID-19 pandemic. contract notice when initiating a public procurement procedure, The Ministry of Development and Technology is the custodian concession, launching a dynamic procurement system, or of the SPS. organizing a design contest. Contract notices shall be sent for publication, by electronic means, in the Electronic Public > > > Procurement System (SEAP), at national level, and in addition F . 3 Romania68 in the Official Journal of the European Union (TED), in the case The national strategy on public procurement, adopted in of public procurement whose estimated value, without VAT is 2015, envisaged measures aimed to improve the quality of above the thresholds provided for in Directives 2014/23/EU, the legislative framework, to ensure overall coherence and 2014/24 /EU and 2014/25 / EU. efficiency of the institutional system, to enhance the regularity and quality of the public procurement process and at the The Romanian Government has made progress in same time to raise capacity of the contracting authorities, implementing many aspects of the National Strategy on with emphasis on professionalization and integrity issues. A Public Procurement in the four years since the adoption of the new legislative package for public procurement was adopted Government Decision no. 901 in 2015. Some of the remaining in 2016 which transposes the EC Directives in the area of challenges consist in the need to ensure the predictability of public procurement. Modifications of the public procurement the procurement legislation and the efficiency of the public legislation were adopted in December 2017 and May 2018 procurement system. Romania has many small contracting and a National Office for Centralized Procurement was set up authorities and competition is generally low, as shown for in May 2018. example by the high proportion of contracts with a single bidder, the high proportion of procurement procedures that The National Agency for Public Procurement (ANAP) operates were negotiated without any call for bids and the low value of both at central and regional level. It is responsible for strategy, procurement advertised on Tenders Electronic Daily (TED) as regulation, establishing and implementing the control and a proportion of GDP. 68. This section is based on the Concept Note document for the 2022 World Bank project “Assessment of the Public Procurement System and Further Support to the Im- plementation of the Public Procurement Strategy” (P169141), the 2021 EU “ROMANIA Public Procurement Monitoring Report”, and the 2021 World Bank “Report on the overall assessment of the public procurement system” (P169141).