Design of public procurement auctions: evidence from cleaning contracts

• Published date: 01 June 2018
• Author: Otto Toivanen,Ari Hyytinen,Sofia Lundberg
• Date: 01 June 2018
• DOI: http://doi.org/10.1111/1756-2171.12232

RAND Journal of Economics

Vol.49, No. 2, Summer 2018

pp. 398–426

Design of public procurement auctions:

evidence from cleaning contracts

Ari Hyytinen∗

Sofia Lundberg∗∗

and

Otto Toivanen∗∗∗

We analyze a regime change from beauty contests to first-price sealed-bid and scoring auctions,

using Swedish data on public procurementof cleaning services. In beauty contests, the lowest bid

often lost, leaving substantial money on the table. The procurement costs were similar before and

after the regime change: (i) Entry strongly decreases the procurement cost but did not change.

Entry would have decreased had the municipalities not adjusted the objects of auctions. (ii)

Municipalities favored in-house suppliers in the old regime, leading to more aggressive bidding

by others. With favoritism reduced, these changes balanced each other out.

1. Introduction

Public procurement constitutes a large part of economic activity in developed and de-

veloping economies.1Unfortunately, inappropriate and inefficient practices are thought to be

widespread in public procurement (e.g., Bandiera, Prat, and Valletti, 2009; Cai, Zhang, and

Henderson, 2013). Recommendations for greater integrity and adoption of rule-based practices

are therefore common, and the benefits of using formal competitive auctions are often strongly

advocated (Klemperer, 2002; Tadelis, 2012; Bajari, Houghton, and Tadelis, 2014). Though

∗University of Jyv¨

askyl¨

a; ari.hyytinen@jyu.fi.

∗∗Ume ˚

a University; sofia.lundberg@umu.se.

∗∗∗Aalto University and KU Leuven; Otto.Toivanen@aalto.fi.

This article supersedes the earlier procurement analyses of ours that circulated under various titles, such as “Favoritism

in public procurement: evidence from Sweden”and “Politics and procurement: evidence from cleaning contracts.” Those

earlier versions had a different focus and used part of the same contract data used in this article. Wewould like to thank a

number of colleagues and seminar participants for useful comments on those earlier versions, and Marc Rysman (Editor),

three anonymous referees, Rob Porter, Tuomas Takalo, Janne Tukiainen, Roope Uusitalo, and seminar participants at

EARIE 2016 (Lisbon) and VATT (Helsinki 2017) for comments and discussions on the current one. The usual caveat

applies.

1Public procurement amounted to about 12% of the Gross Domestic Product (GDP) among the Organization for

Economics Cooperation and Development (OECD) member countries in 2013 (OECD,2015). The EU Commission has

estimated that each year, different levels of government spend about 20% of EU’s GDP to procure goods, works and

services (European Commission, 2012).

398 C2018, The RAND Corporation.

HYYTINEN, LUNDBERG AND TOIVANEN / 399

increasing buyers’ discretion may be beneficial when quality is only partly contractible (e.g.,

Coviello, Guglielmo, and Spagnolo, 2017), rule-based competitive auctions are, in general, be-

lieved to be less prone to inappropriate and inefficient practices. One reason for this belief is that

in such auctions, there are explicit criteria for selecting the winner.2This article provides an anal-

ysis of the consequences of a regime change that induced a shift away from discretionary beauty

contests to a more rule-based procurement environment, in which only first-price sealed-bid or

scoring auctions were allowed.

Our data come from Swedish public procurement auctions of a clearly defined low-tech

product, internal cleaning service contracts. The data cover two regimes that differ in how the

law regulated public procurement. In the old regime, the Swedish procurement law allowed the

municipalities exceptionally high degrees of freedom in choosing how to procure the services

and in how to pick the winner.3The municipalities ended up using beauty contest auctions. In

such an auction, the auctioneer does not commit to an award (allocation) rule (Klemperer, 2002;

Yoganarasimhan, 2016). During the new regime, a much stricter procurement law was in place

and the municipalities had to commit to a first-price sealed-bid auction or to a scoring auction

(Che, 1993; Asker and Cantillon, 2008, 2010) with an explicit award rule.

The stricter procurement law disciplined the Swedish procurement bureaucrats by forcing

them to abandon the practice of using flexible beauty contests. A flip side of this change was that

rule-based procurement regulation and more formal procurement processes, augmented by more

detailed product requirements and bidder qualifications, may make procurement burdensome for

potential bidders and thereby increase entry costs (Bandiera, Prat, and Valletti,2009). Our primary

goal is to understand what the regime change meant for municipalities’ procurement costs by

analyzing the determinants of winning bid and entry (participation) as well as the behavior of

procurement bureaucrats, including potential favoritism.

The cleaning contracts procured by Swedish municipalities provide a good testing ground

for us, because cleaning services have a very simple production process, are simple to contract

on, and, as wewill demonstrate, do not allow significant quality differentiation ex ante or ex post.

The product we study is thus much simpler and more homogeneous than most of the products

that have been considered in prior work. This has three implications for applicable theory: first,

it is hard to believe that the Swedish municipalities could haveobtained significant welfare gains

by ex ante tilting the award rule in some nonprice dimension via a scoring rule. This is in contrast

to, for example, the US highway procurements studied by Lewis and Bajari (2011). Second,

ex post negotiations of significant adaptations to the scope of work or of cost over-runs are

unlikely and indeed were rare, unlike what has been found for some of the more complicated

contracts and products (Bajari, Houghton, and Tadelis, 2014; Decarolis, 2013, 2014, Decarolis

and Palumbo, 2015). Neither is there much scope for provision of noncontractible quality (e.g.,

Cameron, 2000; Spagnolo, 2012; Tadelis, 2012; Coviello, Guglielmo, and Spagnolo, 2017).

Finally, there is less uncertainty in the costs of performing the cleaning job over the duration of

the contract than there is, say, in completing a major construction work (Spulber, 1990).

These considerations suggest that there should be relatively few reasons to depart from the

policy of granting the contract to the lowest bidder (for a similar argument in a different context,

see Di Tellaand Schargrodsky, 2003; Bandiera, Prat, and Valletti,2009). This implies that holding

other things constant, the procurement costs ought to be lower in the new regime, which only

allowed the use of more rule-based and arguablymore competitive first-price or scoring auctions.

Despite the simplicity of the procured service, our findings portray a complex picture of

the mechanisms at work. In the beauty contests of the old regime, the lowest bidder won rarely,

2Using internal records of a bribe-paying firm from an Asian developing country,Tran (2008) finds that mandating

the use of auctions curbs corruption only if the procurement auctions are open and if buyers’ discretion to pick the winner

is limited.

3Sweden applied the EU procurement law already in the early 1990s and also subsequently. The Swedish rules

were relatively lax in the 1990s, because the European rules of that time allowedhigh degrees of freedom in organizing

procurements.

C

The RAND Corporation 2018.