STRATEGIC BIDDING AND CONTRACT RENEGOTIATION

• Published date: 01 May 2019
• Author: Hojin Jung,Carlos Lamarche,Richard Sicotte,Georgia Kosmopoulou
• DOI: http://doi.org/10.1111/iere.12368
• Date: 01 May 2019

INTERNATIONAL ECONOMIC REVIEW

Vol. 60, No. 2, May 2019 DOI: 10.1111/iere.12368

STRATEGIC BIDDING AND CONTRACT RENEGOTIATION∗

BYHOJIN JUNG,GEORGIA KOSMOPOULOU,CARLOS LAMARCHE,AND RICHARD SICOTTE1

Chonbuk National University, Korea; University of Oklahoma, U.S.A.; University of Kentucky,

U.S.A.; University of Vermont,U.S.A.

When firms bid in procurement auctions, they take into account the likelihood of future contract renegotia-

tions. If they anticipate that certain input quantities will change ex post, they have an incentive to strategically

skew their itemized bids, thereby increasing profits for themselves and costs for the procuring agency. We

develop and estimate a structural model of strategic bidding using a data set of road construction projects in

Vermont. We find that bidding strategies lead to increased markups for renegotiated items and reduced markups

for nonrenegotiated items, results consistent with bid-skewing.

1. INTRODUCTION

Most, if not all, procurement contracts are later renegotiated. These renegotiations are often

precipitated by engineers determining, after the contract is awarded, that the actual require-

ments of the project differ from what was originally anticipated. In many contexts, the rules for

amending the contract ex post are carefully delineated. For example, the U.S. government’s

procurement regulations prohibit changing contractually agreed prices unless an item is added

in the field or there is a relevant price adjustment clause (Kosmopoulou and Zhou, 2014; Kos-

mopoulou et al., 2016). However, adjustments to the quantities of labor and materials are not

only permitted but are commonplace.

Procuring agencies and contractors reasonably expect—are even certain—that renegotiations

will ensue, but they are not certain of the precise nature of the changes. Moreover, it is widely

acknowledged that firms modify their bidding strategies based upon their own expectations of

subsequent alterations to the contract. These perceptions by industry participants are backed

by past scholarly work. For example, Athey and Levin (2001) found that forward-looking firms

skew their bids in U.S. Forest Service timber auctions by submitting high (low) unit prices on

types of timber in anticipation that their actual proportion will decrease (increase) from original

Forest Service estimates. Bajari et al. (2014), in their study of the California highway construc-

tion industry, argue that renegotiations are costly for firms, because they need to alter previous

plans and reallocate firm resources. They estimate that when contract quantities are changed

ex post, firms bear adaptation costs of $2.20 for each dollar of additional work. Accordingly,

firms bid less aggressively on projects in which they anticipate contract renegotiations. Our

study examines how the prospect of ex post renegotiation in road construction projects in Ver-

mont affects contractors’ bids and profits. We investigate bid-skewing, a situation when a firm

∗Manuscript received October 2016; revised July 2018.

1The authors would like to thank Javier Donna, seminar participants at the Athens University of Economics

and Business, Colorado School of Mines, Northeastern University, University of California Merced, University of

Virginia, Vanderbilt University, Virginia Tech, and the University of Piraeus, and conference participants at the

2013 International Industrial Organization Conference, the 2013 Conference on Research on Economic Theory and

Econometrics, and the 2014 Workshop on Public Procurement and Concession Design for comments related to this

work.We are indebtedto staff at the Vermont Agency of Transportation (VTrans) for providing useful information. This

research has been supported by VTrans Grant No. SPR722. Please address correspondence to: Georgia Kosmopoulou,

Department of Economics, University of Oklahoma, 308 Cate Center Drive, CCD1, Norman, OK 73019. E-mail:

georgiak@ou.edu.

801

C

(2018) by the Economics Department of the University of Pennsylvania and the Osaka University Institute of Social

and Economic Research Association

802 JUNG ET AL.

anticipates a positive quantity adjustment in one item and strategically bids higher on that item

while decreasing its bid on other items in order to maintain a competitive aggregate bid. We

find strong evidence that firms engage in strategic bid-skewing, raising profit margins.

Our article differs from existing work in two important ways. First, we adapt the analysis

of contract renegotiations of Bajari et al. (2014) to itemized bid-level data, which allows us

to investigate skewed bidding as in Athey and Levin (2001). A thorough exploration of bid-

manipulation in complex contracts requires careful examination of item-level bids. The analysis

of bid aggregates at the project level, while informative, does not reveal the extent to which firms

strategically manipulate their bids. We employ itemized bid information to construct estimates

of the markup of bids above costs, and we compare how they vary across items with and without

positive quantity renegotiation. The variation in markups across items provides evidence on

how firms’ anticipation of change orders affects their bidding behavior. Second, we conduct

this work on a new data set of all construction projects undertaken by the Vermont Agency of

Transportation (VTrans) over a five-year period. One of the novel features of these data is that

they include firm-level financial information, normally not available because of their proprietary

character. Another novelty is that it permits us to control for price adjustment clauses, which

have been omitted in previous studies.

In the next section, we explain the renegotiation, or “change order” process for VTrans con-

struction contracts. We introduce our data and conduct a reduced-form analysis that illustrates

the sensitivity of standard bid regressions to different types of contract revisions. We also de-

velop a structural model of strategic itemized bidding where firms expect adjustments to the

quantities of work and materials in the contract, and we estimate costs and markups, exploring

patterns of bid-skewing at the item level. A study of the size of adjustments due to renegotiation

at the project level can be used to assess the overall impact of uncertainty and firm heterogeneity

on markups, but the test may confound such effects with influences from a number of sources,

including adaptation costs. We circumvent this problem by focusing our analysis on a subsample

of projects that have a similar set of tasks and whose characteristics closely fit the Independent

Private Value (IPV) model. We use nonparametric estimation methods similar to the ones

developed by Guerre et al. (2000) and Bajari et al. (2014) to estimate the distribution of latent

costs after controlling for project heterogeneity.

We first present reduced-form evidence of strategic behavior, including results that show bid-

skewing within projects. The structural analysis reveals that unit costs and markups are increased

among items that were renegotiated after a project was awarded. Our results also show that

while bidders increase their markups on items that have positive quantity adjustments by 10%–

17% at the median level, they lower their bids and markups on items that are not renegotiated

to maximize their potential surplus ex post while maintaining the likelihood to win at a high

level. The behavior leads to a significant increase in the cost of contracting to the state and the

public, higher than that reported by studies considering all forms of renegotiation.

2. DATA AND REDUCED-FORM ESTIMATES

2.1. Overview of Change Orders on VermontTransportation Contracts. Our data set consists

of the complete bidding and payment records of all construction projects auctioned off between

May 2004 and December 2009 by the VTrans. There are 846 bids on 312 individual projects. At

the item level, there are 50,465 bids on 712 items. We classify auctions by project type: asphalt

projects, bridge projects, and miscellaneous projects.2The agency awards contracts to the lowest

bidders in sealed bid auctions held monthly. When advertising a project to the public, VTrans

provides detailed engineer’s plans and information on the work site, the required completion

date, and a brief description of the project.3The engineer’s plans provide a list of quantities for

2Miscellaneous projects include traffic signaling and lighting, grading and draining, parking lots, and landscaping.

3Prequalification status is achieved by the successful completion of two procedures: (i) annual prequalification: the

prequalification committee at VTrans annually assigns each firm certain limitations as to the value of projects and