Improving the Efficiency of Public Procurement : Empirical evidence using micro-level contract data

Sammanfattning: Among OECD countries, close to one-third of government expenditure is allocated using public procurement. This implies that the skill of public authorities in handling the procurement format has a significant impact on how much the public sector accomplish in relation to their goals and responsibilities. The aim of this thesis is to provide new insights into how public procurers can achieve more for the money by improving the design of their auctions and contracts. Specifically, the thesis examines how the efficiency can be influenced by the following three aspects: the method of bid evaluation (the award mechanism), whether items are procured separately or grouped into larger contracts, and variations among procurers in the activities and quantities they specify in a given case.The thesis includes four essays that analyze detailed contract information from Swedish procurements of road reinvestments and public bus services. For a comprehensive sample of the examined contracts, information has been collected from procurement documents and compiled into rich microdata sets. The use of microdata has enabled detailed reviews of how the procurements have been designed as well as statistical analysis of associations between the procurer’s choices and the outcome. While the essays primarily focus on results and implications regarding their respective research questions, the kappa provides a more general discussion about quantitative microanalysis in the field of procurement. This reflects a complementary aim of the thesis, which is to contribute to well-informed decisions regarding whether and how contracting authorities should facilitate analyses of this kind.Two of the essays investigate to what extent and how the Swedish national infrastructure manager Trafikverket can improve its efficiency in procurements of highway pavement replacement. Trafikverket spends approximately EUR 200 million on these works annually. Hence, if better procurement methods could lower the cost by 5% without reducing the quality, about EUR 10 million would be freed up annually for additional road maintenance or other uses. Both essays indicate that even greater efficiency gains can be achieved through improved dissemination of best practices. The results support Trafikverket's policy to bundle similar and adjacent road work but imply that the contracts are generally not large enough to fully benefit from the economies of scale. A conclusion is that it is motivated for Trafikverket to consider the scope for efficient bundling when deciding on the timing of these pavement replacement measures.The other two essays examine how the Swedish regional Public Transport Authorities (PTAs) have implemented the scoring rule award mechanism and its efficiency. This alternative to the lowest price and quality-only mechanisms is promoted within the EU procurement directive and is dominant in many countries. However, the literature on how well public agencies implement this policy is scarce and fragmented. One of the essays shows that scoring rules are common in Swedish procurements of public bus services, but also that there are significant differences between the PTAs in whether and how they have applied this award mechanism. Several implementations have unsound and likely unintended properties, in the light of economic theory. The findings suggest that best practices are not spread effectively between the organizations. The second essay on this theme exploits the property that one PTA alternated between lowest price and a performance focused variant of scoring rules and, equally important, could provide data on a relevant and reliable outcome measure. Based on 30 months of observations on monitored punctuality, performance is not found to be better in contracts awarded using scoring rules. The results suggest that the promotion of scoring rules is not accompanied with adequate guidance on when and how to implement it efficiently.The kappa highlights how microdata at the contract level enable research that considers both the ‘uniqueness’ of each procurement and random variation. In the analyses of road reinvestments, the detailed data allows us to consider several important characteristics of the contracts and the treated road segments. Similarly, the analysis of bus punctuality considers several aspects of the traffic assignment. In this way, quantitative microanalysis can provide reliable quantifications of how costs and quality are affected by both the procurer's actions, while considering the external conditions. Many of these associations are too complex to be reliably assessed by an individual civil servant, even if it is highly knowledgeable and experienced. However, the thesis emphasizes that even quantitative analysis can yield misleading results if the data material is not good enough, for example, if the sample is too small or relevant aspects are not captured.A central argument in the thesis is that the standardized and rigid procurement process creates particularly excellent conditions for collecting useful procurement data. However, the potential for this has largely remained untapped by the public agencies themselves. Instead, these data have primarily been compiled within individual research projects, where even the collection of tendering documents has been a notorious obstacle. In Sweden, Trafikverket's information management and cost control have repeatedly been criticized. However, Trafikverket is far from the ‘worst in class’ in these matters, and the criticism mainly reflects the agency's position as one of Sweden's largest procurers. An overall conclusion of the thesis is that public agencies in general should take a more active role in improving how procurement data are collected and managed. Of particular importance is the establishment of well-structured archiving processes for procurement documents, along with systems to track changes in both costs and content after a contract is awarded. For smaller authorities, it is reasonable that this development is coordinated or driven by a centralized function, whereas Trafikverket is well positioned to take a leading role in finding and disseminating best practices. Ultimately, the responsibility for instructing and overseeing improvements in this area lies with the governing bodies of the public agencies, given that the agencies do not set their own agendas.