Interesting Procurement Paper (Li & Xu, 2016): A Blueprint for Variable Remuneration of Public Procurement Officers? A Warning against some types of Centralised Procurement?

I have just read the paper by D Z Li and M Xu, 'Competition in Procurement Auctions with Corruption' (February 2, 2016), which assesses an interesting scenario of competition in public procurement tenders where the person in charge of running the procedure (the procurement officer, or 'bureaucrat', in their terminology) can require bribes from winning bidders, and where those bribes can be proportionate to the final value of the contract awarded.

Their paper is interesting because it fleshes out the incentives that a bureaucrat that expects to obtain a rent at the end of the procedure has, both in terms of affecting the number of bidders (to reduce it), and the level of disclosure of information (to conceal information in order to cover the corrupt practice). My personal intuition is that their insights should be useful to consider non-corrupt scenarios involving buyer rents other than bribes and, in particular, the introduction of bonuses or other variable retribution mechanisms for public buyers, which could well create the same incentives (as discussed below). Moreover, I find the paper thought-provoking because (legitimate) kick-backs are used to finance the activities of central purchasing bodies, which raises issues of their impact on social welfare if they behave like individual bureaucrats would (as also discussed below).

The paper and its model

As they explain in their abstract:

We study the effects of corruption on equilibrium competition and social welfare in a public procurement auction. A bureaucrat runs the auction on behalf of the government. He invites firms into the auction at positive costs, and may request a bribe from the winning firm afterward[s]. We first show that, in the absence of corruption, the bureaucrat invites more firms than social optimum under quite standard assumptions. Secondly, the effects of corruption on competition and social welfare vary across different forms of bribery. In the case of fixed bribe, corruption has no effect on equilibrium competition, yet [it does] induce social welfare loss due to the distortion cost of increased public spending. In the case of proportional bribe, the corrupt bureaucrat will invite less firms into the auction, which may result in Pareto-improving allocation in equilibrium. Finally, we also show that information disclosure may consistently induce more firms to be invited, if compared with the case of no information disclosure, no matter [whether] there is corruption or not.

I find some of the assumptions and insights of their paper particularly thought-provoking. They (implicitly) base their model on the existence of an agency relationship between the bureaucrat and the government, as well as between the government and society at large [for discussion, see here and here and, in Spanish, here]. This makes the model interesting from the perspective of the social externalities that improperly designed public procurement models can create, particularly if they allow public buyers to pursue (self-serving) goals that do not align with promotion of social welfare.

In their paper, Li & Xu explain that 'the government is modelled as a government division ... who cares about its own procurement pay-off rather than the overall social welfare' (p. 2). This can lead to designing the procurement process in a manner that invites too many interested bidders because 'the optimal number of firms that maximizes the government's pay-off is larger than the efficient number of firms that maximizes social welfare' (ibid), and due to the fact that 'the government prefers [a] higher level of competition in the procurement process' than would be socially efficient (p. 3). The undesirability of the excessive number of bidders is mainly derived from the costs they incur in order to participate in the tender, which are wasted for all those that did not stand a real chance of winning the contract (or, indeed, for all except the winning bidder).

The main insight of their paper is that, while the existence of a fixed bribe hurts both the government and society at large due to the higher cost of procurement, the existence of a proportionate bribe may 'increase social welfare [by inducing an efficient number of firms, or just one firm, to be invited], yet it hurts the government, as the government prefers higher level of competition in the procurement process' because that reduces its (private) procurement cost and imposes the externality derived from excessive tendering costs assumed by the disappointed private bidders [for discussion on the absence of consideration of these costs in economic surveys supporting recent public procurement law reforms in the EU, see here].

Their insight is based, among other elements, on the 'standard assumption for procurement auctions that firms' cost distribution is of decreasing reversed hazard rate (DRHR)' (p. 1, for an explanation of the reversed hazard rate and how it operates, see here). As Li & Xu explain, 'The intuition behind this ... is that increasing competition will gradually squeeze out the expected rent of the winning firm. Furthermore, the expected rent converges to zero when the number of firms approaches to infinity'. Or, in very simple words, that the lower the number of bidders, the higher the expected rent by those that participate. That is what would allow bidders to tender less competitive prices when competitive pressure is reduced (ie less bidders are invited), which would also be in the interest of the bureaucrat expecting to receive a proportional bribe (a higher rent for the winning bidder carries a higher rent for the bureaucrat as well).

They also stress in clear terms that 'information disclosure will increase both the efficiency and the optimal number of firms in the procurement auction. The intuition is that, under information disclosure, firms' cost estimates become more heterogen[e]ous, and therefore, for [a] given number of firms, the auction becomes less competitive than before' (p. 13). Furthermore, 'under information disclosure, firms become more heterogen[e]ous in their cost estimates, and the winning rent, which [in their model] is the difference between the lowest and second lowest costs, may also get larger as well'. However, t'when a corrupt bureaucrat can control information release, it would be more difficult to detect corruption. As we know, information disclosure implies more firms to be invited into the auction, and corruption under the proportion[ate] bribe implies less firms to be invited. The combined effects of these two are mixed'. Overall, then, the implications of their findings seems to be that a corrupt bureaucrat will have mixed incentives on whether to reduce the volume of information disclosed in the tender process because more information may increase its own proportionate rent, but it will also trigger both more interest in the tender and more risk of detection of the corruption.

a blueprint for variable remuneration of procurement officers?

As mentioned, my intuition is that these insights can be useful to consider non-corrupt scenarios involving 'bureaucrat' rents other than bribes and, in particular, the introduction of bonuses or other variable retribution mechanisms for public buyers, which could well create the same incentives. My intuition is that, should the bureaucrat have a legal financial incentive to obtain a rent a the end of the tender, and should the existence of this rent not need to be hidden, it would have an incentive to pursue strategies that maximize social value (even if not necessarily government pay-offs) by disclosing information that reduces the number of potentially interested bidders for which the tender is not actually competitive. Moreover, the financial incentive could include an element of reverse proportionality, so that the bonus would be larger when the government pay-off is increased (ie when the total cost of the procurement is reduced as much as possible within the framework of the competition between the efficient number of bidders). If this is true, then, one of the main aspects that Member States should consider going forward would not only be linked to decisions on how to transpose and develop the rules for restricted procedures and for procurement procedures involving negotiations, but also linked to the establishment of appropriate systems of incentives for procurement officers (bureaucrats) to make the right choice of procedure and to conduct the tender in a way that is aligned with social welfare and with (intra)governmental pay-offs.

what implications for kick-back based central purchasing financing?

Central purchasing bodies (CPBs) can be financed in many ways, but a popular model is for them to receive kick-backs (in the form of rappels of fees) from suppliers included in the framework agreements and other contracts that CPBs manage. Those kick-backs are generally proportionate the value of the call-offs that end-user contracting authorities place with each supplier. In that case, the CPB is not in a different economic position than a procurement officer (bureaucrat) expecting to receive a proportionate rent (or bribe) at the end of the procurement process it runs. Therefore, it seems to me that one of the transferable insights of Li & Xu's paper is that CPBs will be structurally in a situation where they might as well aim to achieve the highest rent, which would require for them to reduce the number of bidders and (possibly, but not necessarily) the information disclosed at the outset of the procedure, so as to reduce the number of competitors and increase their expected rents--thus triggering higher kick-backs for the CPB. This would match well with the intuition that CPBs can become self-interested organisations in the way they run their framework agreements, and not pay excessive attention to the real interest of their principal (end-user contracting authorities) or society at large, particularly if the use of their services is mandatory (ie if they do not need to justify net advantages, at least for the end-user contracting authorities, in order to attract volume of orders).

If this intuition holds true, it would be interesting to look at the impact of the financing of CPBs through kick-backs in more detail, in order to assess whether this system of financial incentives and rewards fosters social welfare overall, or is only beneficial for the CPB (and/or, the government) at the expense of broader social interests. This would be particularly relevant if, as anecdotal evidence indicates, access to centralised procurement is difficult for most firms (and, in particular, SMEs), so that CPBs structurally reduce the number of bidders for their (large) contracts, which the model in the paper would suggest increases the rents for both the CPBs and the included suppliers, but imposes both direct costs on government (through higher procurement costs that could be achievable in alternative settings of increased competition within CPB procurement) and indirect social costs via externalities [for discussion of some of these economic issues, see here].