Single bidding. A binary variable that focuses on the number of participants per contract. It is re-coded into a binary to denote whether there was only one bidder (1) or more than one bidder (0). Having received only one bid in a competitive market indicates that there is restricted competition.
Number of bidders (quartiles). The variable is defined based on the number of bids received, and is recoded to indicate 1, 2, 3, and 4 or more bids. The missing values are re-coded into a separate category.
(Share of new) Market entrants. The indicator is operationalised based on the share of new companies being awarded a contract in a specific year, and not being awarded a contract in the preceding year (t-1), based on CPV-3 classification for markets. For instance, if a company is awarded a contract in 2018, but did not have a contract in 2017, then for 2018 it will be coded as 1, i.e. as a new market entrant.
Market concentration. The indicator is operationalised by calculating the share of the supplier total market value per year. It aims to capture the size of the supplier relative to the market (at CPV level 4).
(Rate of) Non-local suppliers. The rate of non-local suppliers indicator is operationalised based on the registered location of the buyer and awarded bidder. If the buyer and bidder are from the same settlement, it is assigned as local bidders. If the buyer and bidder location is not from the same settlement, it is assigned as non-local bidders.
Improving Competitive Practices in Hungary’s Public Procurement
Annex B. Precise definitions of variables used in the analysis
Copy link to Annex B. Precise definitions of variables used in the analysisCompetition indicators
Copy link to Competition indicatorsEconomic fundamentals
Copy link to Economic fundamentalsYear. Control variables that account for the year of the tender. It makes sure that all variation is observed within the same year to avoid potential confounding effects on yearly variation.
Month. Categorical variable that overviews the number of tenders per month. The variable is measured by looking at the month of the first call for tender publication. With missing values from the call for tender publication, the month is imputed using the median difference between contract award date and first call for tender publication within sectors (using CPV level 4, as a more conservative imputation). End of the year spikes in procurement might occur due to unsuitable budget planning or inefficient organisation of tenders throughout the year and the association of between the increased rate of single bidding in the latter months of the year, or the last quarter (McCue, Prier and Lofaro, 2021[1]). Therefore, the possibility of budget surplus for public authorities could indicate the need to spend the allocated funds in the last few months of the year. This can trigger shorter and less adequate planning procedures, as well as shortened advertisement periods. Academic literature finds spikes of the share of single bidding incidence in the last month of the year.
Sector. To identify sectors, the Common Procurement Vocabulary (CPV) codes are used. Based on the CPV codes, more general sectors are identified, using 2 digits, or more specific sectors, using the 3- or 4-digit CPV notation. Considering that some CPV codes can entail a wide range of products and to estimate the average single bidding instances, differences within sectors are analysed.
Region. Regional and economic differences are captured by a variable based on the buyer’s location and contract implementation region, using the NUTS2 classification as a proxy. Theoretically, as competition improves, an increase in the number of distinct companies that have won a contract in a particular market can be noticed. Subsequently, an increase in the number of potential suppliers should be associated with lower rates of single bidding.
Market size. The variable is calculated by grouping markets (CPV level 2 and 3) and calculating the total value of contracts within these groups. CPV level 2 markets are divided into quintiles (5 groups), while the 3-digit markets are divided into deciles (10 groups). Smaller markets, i.e. lowest quintiles and deciles groups, have larger share of single bidding incidence.
Number of potential suppliers on the market. The variable is calculated by grouping markets (CPV level 4) and counting the distinct number of awarded suppliers. Markets, such as petroleum, coal and oil products, mining and basic metals, or clinical forensics equipment have the lowest number of unique suppliers. Finally, the variable is categorised by creating quintiles (5 groups) for the CPV level 4 markets.
Organisational capacities
Copy link to Organisational capacitiesBuyer type. Control variable that takes into account whether the buyer is a national or regional authority or a national or regional agency. The rationale behind the control is to consider the potential differences of organisational capabilities of different agencies or authorities. Such structural indicators are related to the buyer’s capacity and resources, considering that national authorities or national agencies could have at their disposal larger personnel or more qualified employees.
Buyer’s average decision period length per bid. The indicator is the average time difference between the bid submission date and the announcement of the contract award per buyer and year. The indicator captures the capacity of agencies and availability of resources they have to process submitted bids. Theoretically, quicker periods for making decisions should be associated with more effectively organised agencies, and, consequently, can stimulate competition. Single bidding is lowest with the most efficient agencies whose buyer’s average decision period length per bid is less than 22 days. Too long periods can signal inefficient agency or low capacities that can discourage potential suppliers from making bids.
Tender and contract design
Copy link to Tender and contract designProcedure type. Non-open procedures lead to uncompetitive tenders. It can create an environment of unfair competition by limiting the number of competing suppliers. Examples of such risky procedure types include negotiated procedures without prior announcement, or contracts directly awarded to suppliers. The variable is based on the procedure chosen by the contracting authority during the tendering process. The wide range of original procedure types are then mapped into fewer, standardised categories. Procedure type is an important determinant bearing in mind that buyers can misuse the application of certain procedure types for the purpose of favouring specific bidders. Research has shown that specific firms can enjoy extra returns and more favourable conditions when negotiated or exceptional procedures are more frequently used. (Auriol, Straub and Flochel, 2016[2]) Additionally, discretion afforded to buyers to use negotiated procedures has a tendency to create favouritism. (Chong, Klien and Saussier, 2015[3])
Tender description. Length of procurement description (number of characters). To calculate measurable categories, first, the deviation from the mean based on 3-digit CPV codes is calculated. Second, it is split into categories. Tenders that have the fewest number of characters (largest deviation) until the 1 quarter are classified as category 4. Tenders between the mean and the 1 quarter are classified as category 3. Tenders between the mean and the 3 quarter are the classified as category 2. Last, tenders that have the longest description are classified as category 1.
Contract size (deciles). Categorical variable that denotes the value of the tender split into 10 deciles.
Number of lots. The variable measures the number of lots for each tender. To calculate measurable categories, the variable was split into deciles that represent specific groups. For instance, the first decile represents one-lot tenders, the third group three-lot tenders, and so on. The last group is the most diverse as it entails tenders whose lot number ranges from 6 to 695.
Ex-post modifications of contracts. Number of contract amendments.
Transparency, open data and e-procurement
Copy link to Transparency, open data and e-procurement(No) call for tender publication. Call for tender publication measures whether the contracting authority has issued a notice for the tendering procedure. Lack of such an announcement in an official journal leads to limited competition as fewer potential suppliers are informed about the tender. Furthermore, such limitation can also indicate that the potential supplier is selected based on favouritism. These tenders are not conducive to the principles of open and fair competition.
Length of advertisement period. The indicator measures the difference between first contract notice publication date and the deadline that suppliers have to submit their bids (in days). Theoretically, a short period between the two dates is associated with unfair competition as there is less time for adequate preparation. Similarly, too lengthy periods can also signal potential tender modifications that can cause favouritism towards some bidders.
References
[2] Auriol, E., S. Straub and T. Flochel (2016), “Public procurement and rent-seeking: the case of Paraguay”, World Development 77, pp. 395-407.
[3] Chong, E., M. Klien and S. Saussier (2015), “The Quality of Governance and the Use of Negotiated Procurement Procedures: Evidence from the European Union”, Institut d’Administration des Entreprises 2015-3.
[1] McCue, C., E. Prier and R. Lofaro (2021), “Examining year-end spending spikes in the European Economic Area: a comparative study of procurement contracts”, Journal of Public Budgeting, Accounting & Financial Management, Vol. 33/5, pp. 513-532.