روش های ارزیابی مناقصه و انتخاب تامین کنندگان در تأمین تجهیزات عمومی

The EU procurement directives stipulate that public contracts be awarded to the lowest bidder or to the bidder with the economically most advantageous tender; the latter requiring that a scoring rule be specified. We provide a simple theoretical framework, based on standard microeconomic theory, for tender evaluation (scoring and weighing) and discuss the pros and cons of methods such as highest quality (beauty contest), lowest price and price-and-quality-based evaluations. We argue that the most common method, price-to-quality scoring, is inappropriate for several reasons. It is non-transparent, making accurate representation of the procurer's preferences difficult. It is often open to strategic manipulation, due to dependence on irrelevant alternatives, and it tends to impose particular and unjustified non-linearity in bid prices. The alternative quality-to-price scoring method, where money values are assigned to different quality levels, is a better alternative. However, when the cost of quality is relatively well-known and several providers can offer optimal quality, lowest price is the preferable supplier selection method, while beauty contests may be preferred when purchasing budgets are inflexible.

Public procurement is an important government activity, with an estimated value corresponding to 15% of world GDP.1 In the US, public procurement has traditionally been strictly regulated and procurement contracts are usually awarded to the lowest qualified bidder, although other methods have recently attracted interest (Potoski, 2008 and Bajari and Lewis, 2011). In the EU, lowest price is used less frequently and instead supplier selection methods that combine price and quality into a total score are used more often.2 Drawing on standard assumptions in economic theory and some practical considerations we discuss benefits and drawbacks of using lowest price or alternative methods for supplier selection – the economically most advantageous tender, EMAT, in EU terminology – as well as the appropriate design of scoring rules. The use of EMAT normally requires the specification of a scoring rule. It has been argued that the strict regulation of public procurement in the US and the so-called buy-low-bid rule created an atmosphere in which suppliers cut costs by offering the bare minimum acceptable quality—sometimes acceptable only in a formal sense. The government (notably, the US Department of Defense) responded by issuing increasingly detailed and complex product specifications; a response which tended to increase transaction costs and to reduce competition. In response to these perceived problems, a US National Commission – the 1993 “Winter Commission” – proposed fundamental reforms of the procurement practices. The commission's aim was to give managers the tools to pursue public value, whether through purchasing or other means. Hence, the reform of government procurement practices can be seen as part of the wider New Public Management trend (see e.g. Kelman, 2002). In the EU, procurement policy has moved in the opposite direction.3 The 2004 procurement directives, Directives 2004/17/EC and 2004/18/EC, articles 55 and 53, require that the call for tender specifies how bids will be evaluated in terms of a supplier selection method, giving less room for discretion while reducing the risk for discrimination.4 An evolution in this direction has been endorsed by the academic community (e.g., Chen, 2008, Mateus et al., 2010 and Telgen and Schotanus, 2010). The 2011 proposals for a revision of the EU Directives continue to move in the same direction by proposing that the use of EMAT should be mandatory. An ongoing trend has also been the transformation of EU procurement rules from framework rules that gave the member states significant discretion to codify the national procurement rules that actually applies to firms—into a system of common rules (Arrowsmith, 2006). I.e., also the member states’ discretion has diminished. As an example, the previous procurement rules made it possible to arrange a tender and only then, after the bids had been received, choose between giving the contract to the bidder offering the lowest price or to the economically most advantageous tender. The old directives were also in other respects less strict in requiring that the supplier selection methods and, when applied, the scoring rule to be precisely specified in advance.5 Using both price and quality in supplier selection can enhance the efficiency of public procurement, although it adds complexity to the procedure. However, the supplier selection methodologies and more specifically the scoring rules that are used in practice are often poorly designed. We argue that the application of simple economic principles can enhance the effectiveness of supplier selection. The economically most advantageous tender can be the bid with the highest quality for a given price, in so-called beauty contests. It can also be the bid that achieves the highest combined price and quality score. If this approach is used, price(s) and one or more quality measures will have to be combined into a single measure. Either quality (differences) must be evaluated in monetary terms or price must be transformed into a score that is commensurate with the quality score. Using the first method, quality value in excess of the minimum requirement can be subtracted from the price bid or, alternatively, the value of the quality gap relative to the maximum quality level can be added to the price bid. I.e., the supplier selection method will be quality-adjusted lowest price. Here the expression quality-to-price scoring will be used to describe this method. If the second method is used, price must be transformed into a score that is added to the quality score, making the tender a price-adjusted highest-quality tender. We will refer to this as price-to-quality scoring. Lowest-price tender evaluation is in principle straight-forward, while in practice it may be challenging to define effective and appropriate minimum quality requirements, as well as to weigh multiple prices into a single cost measure. Quality-only tender evaluations (beauty contests) can be more complicated. If quality is measured in more than one dimension, the quality measures will have to be combined (weighed) into a single overall score, but, unlike prices, they cannot simply be added together. The main focus of the present study is the related problem of combining (weighing) quality and price into a single overall score, using either quality-to-price or price-to-quality scoring. As will become evident, the scoring rule we propose will also be useful when quality measures in different dimensions are combined into an overall quality score. Overall, public procurement is a phenomenon that has attracted little attention from academics—much less so than private procurement (Dini et al., 2006 and Rendon and Snider, 2010). Public procurement is different from private procurement as covered extensively in Telgen and Schotanus (2007) in its stronger emphasis on rules and predictability. Since the public procurer has less discretion to select any other bidder than the one that was awarded the highest score it is more critical to use a well-designed model for tender evaluation. Further, there appears to be a paucity of research that bridges the gaps between theoretical analyses of abstract scoring rules and the practical application of scoring rules that can be used in real procurement. One of the few exceptions is Dini et al. (2006), which offer guidance also for practitioners.6 For example, they emphasize the favourable properties of scoring rules that are linear in prices. Asker and Cantillon, 2008 and Asker and Cantillon, 2010 demonstrate that scoring rules dominate beauty contests, price-only auctions and menu auctions and that, when bidders have private information of a multidimensional nature and quality is contractible, scoring rules outperform bargaining models. Further, scoring rules should be designed to correspond to the utility function of the procuring entity.7 In practice, however, it is not apparent how to design a good scoring rule. Public procurement can be seen as a process in several steps, from the identification of needs via design of the tender process, choice of supplier selection methods and a scoring rule for evaluating bids and contract design, to ex-post control and contract enforcement mechanisms. This paper will focus on how already identified needs, in a setting with verifiable quality, can and should be expressed in a model for evaluating bids: a scoring rule. We will, independently of the legal framework, discuss supplier selection methods and scoring rules from a theoretical perspective. To facilitate the analysis, we introduce a graphical analysis of procurement when quality matters. Furthermore, we compare different procurement schemes under uncertainty, drawing on environmental economics theory. We argue that quality-to-price scoring outperforms price-to-quality scoring. An appealing characteristic of the quality-to-price models is that they do not require explicit weights; neither weights for price and quality, nor weights for different quality criteria. The value of the different quality criteria can simply be added together and then subtracted from or added to the bid price, as appropriate. We emphasize that the supplier selection method and scoring rule should reflect the preferences of the procuring authority or its principal, the society.8 However, when the cost of obtaining the desired quality is known and when several firms can deliver the stipulated quality, transaction costs will be lower if lowest price is used to select supplier. In contrast, quality-only scoring may be the best method when funds are earmarked for a particular project. We compare our theoretical results with actual procurements, as reflected in a sample of 189 Swedish public procurements of four services: elderly care, waste transport, food wholesale services and cleaning services. We find that lowest price is used in more than one-third of the procurements while supplier selection based on scoring rules that include both price and quality measures are used in more than half of the procurements.9 Within this latter category, the large majority relies on scoring rules that, from an economic perspective, have one or more undesirable properties, indicating that a proper understanding of the complexities of scoring and evaluation in two (or more) dimensions is essential for public procurement. This paper is organized as follows. A theoretical framework for procurement when quality matters is outlined in Section 2. The choice between lowest price, quality-only scoring and scoring rules that combines price and quality is discussed in Section 3. Section 4 compares price-to-quality and quality-to-price scoring rules and weighing is discussion in Section 5 while Section 6 concludes.

In this paper we discuss the design of supplier selection methods for public procurement. We compare supplier selection based on lowest price to scoring rules that can be used when supplier selection is based on the economically most advantageous tender. The theoretical framework is relative simple. Most of our analysis is based on the assumption that the utility of the procuring authority is linear in prices and non-linear in quality. However, we briefly discuss the possibility that the authority's budget contains funds that are earmarked for particular purposes; a setting which tends to make utility non-linear in prices. We also maintain the assumption that quality is verifiable, so that the supplier can be held accountable if quality is degraded. If quality is, in fact, not verifiable, ex-post quality degradation (moral hazard) and non-verifiable quality variations (adverse selection) is likely to occur and it appears reasonable to put a greater weight on quality, possibly going all the way to pure quality competition, since it will be easier to give the contract to a firm with good reputation if price is given a low – or zero – weight in the bid evaluation. Bluntly put, it is easier to bend the rules to account for non-verifiable quality when setting the quality scores than when comparing prices. Doing so will allow the authority to maintain a high-quality equilibrium with the same kind of mechanism that is used in markets with non-government buyers. Returning to a setting where quality is verifiable, pure price competition in combination with appropriate compulsory quality conditions (lowest price) is appropriate when there is little uncertainty concerning production costs for different product specifications (and when several providers can offer optimal or close to optimal quality) or when it is vital to reach a minimum quality threshold. EMAT with quality-to-price scoring is more appropriate when it is not critical to achieve a particular quality level, but when there are reasons to believe that excessive quality will be very expensive. If the total score depends both on quality and and price, then the procuring authority should assign monetary values to quality – values that can be added to or subtracted from the actual bids – rather than calculate a price score that can be added to a quality score. That is, quality-to-price scoring is preferable to price-to-quality scoring. One reason is that we are all used to making assessments in monetary terms when making choices between products that differ in terms of quality. Another reason is that valuation is typically close to linear in bid prices, something that automatically follows from using our preferred method but, in most cases, not from price-to-quality scoring. Quality valuation as a scoring rule for supplier selection also has the advantage that the ranking of the bids will not depend on irrelevant alternatives (non-competitive bids) due to endogenously determined reference prices, although dependence on irrelevant alternatives is not an unavoidable consequence of price-to-quality scoring. When calculating the overall price, the expected purchase volumes for each component product should be used for weights. Adjustment for quality value should be done before the weighted price (or total cost) is calculated. Therefore, when using this scoring rule, there is no role for weights for price and quality, respectively. (If, for legal reasons, weights are required, equal weights for price and quality can nominally be assumed.) We can speculate that the main attraction of price-to-quality scoring to practitioners, paradoxically, has been that it is not transparent. Since it is difficult to understand the implications of a particular scoring rule, the procuring authorities have not been impelled to consider their preferences in detail and it has been difficult, for them and for other concerned parties, to object against a proposed formula. The opaqueness of the rules has also hidden the fact that many versions of price-to-quality scoring do not treat bidders equally. In particular, non-linear functions for price scores implies that different bidders will have to reduce their prices by different amounts to achieve a given increase of the overall score. We argue that price-to-quality scoring, as applied in practice, often violates both the transparency principle and the principle of equal treatment enshrined in EU's procurement rules. To summarize, when quality is verifiable, policy recommendations are as following: (i) if bids are to be evaluated according to a combination of price and quality, this should be done by assigning monetary values to quality characteristics (quality-to-price scoring), rather than by transforming bid prices into scores (price-to-quality scoring). (ii) If price-to-quality scoring should at all be used, it should be based on pre-specified reference prices. (iii) Whether quality-to-price scoring is based on quality discounts or on quality surcharges is inconsequential, but absolute discounts or surcharges are to be preferred over relative ones. (iv) Normally quality values per unit or per characteristic should be determined first and only then aggregated, rather than the converse. (v) In non-complex situations, when the costs of producing to different quality levels are well known and several firms have the ability to provide (close to) optimal quality, lowest price, being a simple and robust supplier selection method, is to be preferred. In other situations, quality-to-price scoring is likely to be the best response to cost uncertainty. We assume that preferences are normally close to linear in bid prices. However, in particular situations this may not be an accurate representation of the procuring authority, which leads us to our final recommendation. (vi) If the procuring authority has earmarked funds, or if the procurement is large enough to use up a substantial part of the authority's whole budget, quality-only with fixed price may be the best way to select supplier.