The effect of information feedback in construction bidding
Alexander Soo and Bee Lan Oo (School of Civil Engineering, University of Sydney)
Abstract
With the goal to achieve efficiency in bidding competitions, many codes of bidding procedure recommend clients provide contractors with bidding feedback information. Contractors strive to bid competitively via learning based on their experiences in past bidding attempts. The level of bidding feedback information, however, varies across clients. In many cases, clients do not provide feedback or provide insufficient feedback to contractors. Focussing on two information feedback conditions (full and partial), we
examine: (i) the changes in bidding trend over time, and (ii) the effects of bidding feedback inforyation on xidders?? competitiveness in bidding. Data were gathered using a bidding experiment that involved student (inexperienced) bidders with a construction project management background. The results show that the variations in bids over time for full information feedback condition are statistically significant, but not for bids from bidders with partial bidding feedback information. Bidders with full bidding feedback information are more competitive than those with partial bidding feedback information. The findings add to both our theoretical and empirical understanding of construction bidding: an
understanding of the process of changes in the price of building work, and how the process can be manipulated through the release of bidding feedback information.
Keywords: Bidding, Construction, Experiment, Feedback.
Introduction
Constr{ction clients?? ox??ective of a|arding a contract is to agree {zon a contract tuze and price that produces reasonable risk and maximises the incentive for efficient and economic performance of a contractor (Kerzner 2006). They will naturally aim to strike the best bargain by introducing some kind of competition, by some form of negotiation or by a mixture of the two. Descending first-price sealed-bid auctions are most commonly used in the industry where competing contractors submit independent bids (i.e., offers to sell construction services) and the lowest bidder wins at the lowest bid price. The competitive bidding is the driving force for contractors providing lower bid prices to suit the clients'
construction and financial needs (Drew and Skitmore 1992). Nonetheless, Runeson (2000) noted that a bid offer must be equal to or above the minimum price at which a contractor is prepared to undertake also considering acceptable probability of profit without an unacceptable risk of loss.
Contractors adopt various strategies to enhance their chances of winning projects. Their
experiences in past bidding competitions play a role in offering competitive bid prices. Using bidding data of building projects, Fu et al. (2004) found that experienced bidders who bid frequently are more competitive than bidders who bid occasionally, where experiential learning in recurrent bidding plays a key role.In this, bidding feedback information is necessaru to facilitate contractors?? learning (??agel and Levin 2002). Varying information feedback conditions have been shown to affect a bidder's competitiveness to different degrees in sealed-bid auctions, whereby affecting the revenues for those accepting bids to
Australasian Journal of Construction Economics and Building
buy or accepting offers to sell (e.g. Issac and Walker 1985; Dufwenberg and Gneezy 2002; Engelbrecht-Wiggans and Katok 2008).
Many codes of bidding procedure recommend clients provide contractors with bidding feedback information (e.g. New South Wales Government 2005; Ministry of Finance 2005). The level of bidding feedback information, however, varies across clients. For example, while the Singapore Government releases the full list of bidders and their bids on the GeBIZ government website (www.gebiz.gov.sg), the New South Wales Government in Australia releases the details of successful bid only
(www.tenders.nsw.gov.au). In many cases, clients do not provide feedback or provide insufficient feedback to contractors (Drew and Fellows 1996). Although their survey respondents indicated that they obtain bidding data from a variety of sources, including: competitors, subcontractors, friendly acquaintances, suppliers and newspapers, little is known about the effects of varying levels of bidding feedback information on construction bid prices. This study aims to examine the effects of two information feedback conditions in construction bidding. Using an experimental approach, two groups of inexperienced bidders were randomly allocated to full and partial information feedback conditions,
respectively. With regard to the two information feedback conditions, the specific objectives are to examine: (i) the changes in bidding trend over time, and (ii) the effects of bidding feedxac} inforyation on xidders?? competitiveness in bidding. The study provides an insight into changes in the price of building work associated with the release of bidding feedback information.
Effects of Information in Construction Bidding
Information related to construction bidding can be separated into two categories, namely public information and feedback information. Public information, as the name implies, refers to information that is publicly available such as project type and size, project location and client identity. Feedback information refers to the information provided at the end of bidding coyzetitions~ ??ezending on clients?? zroc{reyent procedure, the levels of bidding feedback information given to contractors varies. Drew and Fellows (1996) identified that contractors use bidding feedback information for four different purposes: (i) for deciding on whether or not to bid for future projects, (ii) for determining mark-up for future projects, (iii) for analysing their bidding performance; and (iv) for analysing bidding performance of their competitors.
Public Information
There have been many studies on the effects of public information in competitive bidding. The leading article referenced by many authors appears to be that written by Milgrom and Weber (1982). They define public information being either a cost estimate provided by the client or project information such as geological data or proprietary information (private surveys, etc). In ascending first-price sealed-bid auctions (i.e., highest bidder wins at the highest bid price), they found that bidders bid more
aggressively when public information is released, hence raising the seller's profit. In another study by de Silva et al. (2008), which examined the impact of public information on highway procurement where lowest bidder wins the job, they found that the release of a cost estimate on bridge procurement projects led to a sharp decrease in the bids received. The adjustment in bid prices from what is believed to be a more accurate cost estimate from the procurement agency suggests that bidders have a strong intent on winning the bid, and lowering the bid increases their chances of winning. These studies suggest greater revenues for those agents accepting bids or accepting offers. However, a recent study by de Silva et al. (2009) detected that bidders typically bid higher on procurement contracts from the Oklahoma Department of Transportation, following the release of cost estimate from the procurement agency which allowed them to readjust their bids in order to gain a higher profit margin.
Feedback Information
??he effects of varuing inforyation feedxac} conditions on xidders?? xidding xehavio{r have been identified via experimental setting in conventional economics literature. These
Soo, A and Oo, B L (2010) ‘The effect of information feedback in construction bidding’, Australasian Journal of Construction Economics and Building, 10 (1/2) 65‐75
66
Australasian Journal of Construction Economics and Building
experiments were designed in two main settings, namely ascending first-price sealed-bid auction and descending first-price sealed-bid auction. In ascending first-price sealed-bid auctions,
Engelbrecht-Wiggans and Katok (2008) conducted an experiment that involved student subjects to bid for a fictitious asset in 100 auctions. They found that revealing the winning bids led to a higher average bid, and that the critical feedback information is the winning bids, as other information feedback condition (full information about past bids) appears to exhibit variability (either an increasing or decreasing trend) in the student s{x??ects?? xids~ Siyilarlu, Ne{gexa{er and Perote (??????8v detected that the release of winning bid triggers bidders in their experiment to base their bids on the winning bid, as opposed to a no information feedback condition where they must rely on their experiences through recurrent bidding. They found that the release of feedback information led to higher overall bids. Neugebauer and Selten (2006) have also detected overbidding in partial information feedback condition that lead to generally higher bids among subjects in their experiment that aimed to examine the effects of partial and no inforyation feedxac} conditions on xidders?? behaviour. Another experiment by Issac and Walker
(1985), however, shows that full information feedback condition produces lower bid prices in comparing the effects of full (all past bid) and partial (winning bid) information feedback conditions among na?ve
bidders in an ascending sealed-bid auction. In another study by Ockenfels and Selten (2005), they found that a no information feedback condition generally leads to higher bids in an ascending sealed-bid auction. Their research involved observing the upward and downward impulses of bidding
subjects. The weighted impulse is defined in their research as a mix of winning utility and risk aversion strategies. They noted that bidding subjects with no feedback information were unable to consider the relative sizes of the downward and upward impulses and systematically underestimate the downward impulses leading to higher bids.
In a descending first-price sealed-bid auction, Dufwenberg and Gneezy (2002) studied the
effects of vario{s inforyation feedxac} conditions on xidders?? xehavio{r (i~e~, revealing all bids, all winning bids or no bids) for ten rounds. They found that the condition where full bidding feedback information was provided led to much higher bid prices over the limited and no information feedback conditions, thereby raising the seller??s zrofit~ ??heir res{lts also sho| that information released about a bidder's own
performance generally leads to a lower bid, which is in line with Engelbrecht-wiggans and ??ato}??s (??????8v finding that the winning price is the most critical feedback information. Moving towards a modelling attempt on feedback information in descending first-price auctions, Esponda (2008) suggests that the release of feedback information leads to an incorrect evaluation by bidders due to overestimation of the expected profit from the past bids. His empirical analysis shows that the feedback information does have a correlation to bid prices.
Altho{gh the effects of different inforyation feedxac} conditions on xidders?? xehavio{r seey to be mixed in the above experimental studies, it is clear that the experiment subjects had used the feedback information,to varying degree, for analysing the bidding performance of competing bidders. This would allow for insight into coyzetitors?? xidding trends or movements (Friedman 1956). The information gained could be used to estimate the probable range of competitors' bids (Milgrom 1989), and to
differentiate the less serious competitors from the more serious (McCaffer 1976). Kortanek et al. (1973) noted that a xidder??s xidding strategu |hich reflects its xidding behaviour at any time is a direct zrod{ct of learning, governing the xidder??s coyzetitiveness~
Bidding trends over time
Much has been reported on variations in bids over time, suggesting that price differences originate in systematic variations rather than random variations (e.g. de Neufville et al. 1977, Flanagan and Norman 1985, Chan et al. 1996). Runeson and Skitmore (1999) argued that variations in bids over time can be explained by changes in demand, firm capacity level and competitor behaviour. Highly correlated with changes in deyand is the xidders?? need for work, which tends to be high in recession time as demand decreases. As demonstrated by
Soo, A and Oo, B L (2010) ‘The effect of information feedback in construction bidding’, Australasian Journal of Construction Economics and Building, 10 (1/2) 65‐75
67
Australasian Journal of Construction Economics and Building
McCaffer and Pettitt??s (197??v c{s{y c{rve, xidders tuzicallu red{ce their xids grad{allu, relative to their competitors, until they win a contract. The attainment of the contract leads to a sharp increase in the subsequent bids by the bidders as they were not so eager to win a job. In their analysis using around 600 contracts involving about 400 contractors, it appears that there is a continuous cycle of upward and
downward bidding trends over time. However, in a further study by Skitmore and Runeson (2006) that aimed at testing the statistical significance of the bidding trends detected in McCaffer and Pettitt (1976), they found that: (i) winning bids are not in general preceded by increasingly more competitive bids, and (ii) the trend of high and low bids over a period of time is most likely due to the presence of highly
uncompetitive bids or outliers, for their dataset comprised bids from eight bidders. In yet another paper, Oo and Lo (2010) have identified the specific types of bidding trends before and after a winning bid using the longest series of bids from 67 bidders. Their model parameter estimates support the upward and downward trends before and after a winning xid in McCaffer and Pettitt??s c{s{y c{rve (197??v~
Here, |e are concerned |ith changes in xidders?? zricing xehavio{r over tiye in reszonse to two
information feedback conditions. As Runeson and Skitmore (1999) conjecture, the assumption that competing bidders do not modify their behaviour at any time (i.e., no allowance for continuity) is unlikely to be valid. The strategic motivation for changes in pricing behaviour is the long-term survival of a firm (Skitmore and Smyth 2007). Many construction organizations feel they have to fight for survival (Skitmore et al. 2006), especially when demand levels lead to overcapacity, intensified competition or changing client needs (Skitmore and Smyth 2007). With this as a backdrop, contractors have to rely on effective pricing methods in order to translate potential business into reality. This inevitably involves effective utilisation of bidding feedback information towards winning jobs with high profit potential. This study follows the notion that information never has a negative value to the decision-maker (Milgrom and Weber 1982). At worst, irrelevant bidding feedback information can be ignored by the bidders. Based on the research aim and objectives, there are two hypotheses that form the foundation of empirical investigation in this paper:
H1: Inexperienced bidders do change their pricing behaviour in response to a given set of bidding feedback information.
H2: Inexperienced bidders with full bidding feedback information are more competitive than inexperienced bidders with limited bidding feedback information in their bidding attempts.
Research Method
An experimental approach was chosen as the most suitable method to test the effects of information
feedback conditions in construction bidding. Given that there are so many possible factors affecting contractors?? decision ya}ing in zricing - only an experimental research design would allow for manipulation of variable(s), something that would not have been possible using field data. The
complete design characteristics that apply to the experiment presented next were detailed in Oo and Soo (2010).
The experiment involved final year undergraduate students with a construction project management background and were enrolled in a project procurement course. They were randomly split into two groups, one with full bidding feedback information (list of all bidders and their submitted bids), the other with partial bidding feedback information (winning bid and identity of winning bidder). The two primary groups were further split into five subgroups (4 students in each subgroup) to emulate a bidding
competition of five competing bidders. It should be noted that the experiment was conducted in a controlled environment with an experiment coordinator ensuring that strictly no communication was allowed between the groups to prevent bidder collusion.
Soo, A and Oo, B L (2010) ‘The effect of information feedback in construction bidding’, Australasian Journal of Construction Economics and Building, 10 (1/2) 65‐75
68
Australasian Journal of Construction Economics and Building
The experiment was carried out for eight rounds (one round per week) where seven hypothetical projects were generated in each bidding round. All bidding feedback information (either partial or full) was provided at the beginning of each bidding round. Within a timeframe of one hour, the subjects were required to decide which project to bid for, and the bid price if deciding to bid. The general instruction to the participants was that their ultimate aim was to survive and prosper, in a competition in which the lowest bidder wins the job, but how this can be achieved was left to them. This reflects the
strategic nature of the construction pricing problem. The hypothetical projects were constructed using information from past real contracts obtained from the NSW e-tendering website. The projects selected were conventional buildings, such as schools and institutional buildings that involve usual design, and do not require any unusual construction technologies. This was done to control the effect of project type on xidders?? xidding decisions~ Azart froy the zro??ect inforyation (location, duration, client and contract type), the subjects were also given an unbiased cost estimate for each hypothetical project, which is the net project construction cost that includes the site overheads and project preliminaries (i.e., total of
direct cost estimate + site overheads). Here, identical hypothetical projects were given to both the partial and full information feedback groups to enable direct comparison. In an attempt to make the exzeriyent yore realistic and to yaintain s{x??ects?? interests over eight ro{nds, zrofit/loss was generated for each hypothetical project by deducting a randomly assigned final cost from the winning bid. The subject who generated the biggest profit at the end of eight rounds was declared the winner and received a mystery prize.
The main limitations of this experiment are that: (i) it is utilising students as experimental bidding subjects, and (ii) it does not consider other possible information feedback conditions. However, it is believed that the student subjects in the experiment have responded seriously since they asked the relevant questions after reading the instructions. In addition, the mystery prize for winning bidder was introduced as an incentive for the subjects to perform seriously in the experiment. In fact, the use of students in bidding
experiments is a common azzroach to exayine the effects of vario{s xidding variaxles on xidders?? xidding behavour (e.g. Dufwenberg and Gneezy 2002; Issac and Walker 1985; Neugebauer and Perote 2008). A study by Dyer et al. (1989) compared na?ve (student) bidders against experienced (company executives) bidders in an experimental construction bidding environment. They found that there is no significant difference between the na?ve and experienced bidders in terms of bidding performance. They didhowever note that na?ve and experienced bidders differ in terms of their risk attitudes with na?ve bidders being more risk aversive as opposed to risk neutral. With regards to the other limitation of other possible information feedback conditions, it is feasible to replicate the experiment with (i) no information feedback condition, and (ii) information feedback on winning bid only.
The data collected from this experiment was analysed using the Statistical Package for the Social Sciences (SPSS). The measure of competitiveness between bids is to express each bid as a ratio above the unbiased cost estimate, i.e., mark-up competitiveness ratio, MUCR = xidder??s xid/{nxiased cost estiyate~ ??he {nxiased cost estiyate zrovides a coyyon baseline for comparison between the partial and full information feedback groups. A MUCR of 1 indicates that a bid is at the unbiased cost estimate (i.e., zero mark-up), and below 1 indicates a bidder has submitted a bid lower than the unbiased cost estimate that leads to a lower bid. Lower MUCR values indicate greater competitiveness since the lowest bidder wins at the lowest bid price. Consistency in bidding was then gauged from the resultant standard deviation.
Results and Discussion
The results are presented in three parts. The first part presents an exploratory analysis of bids obtained from both the partial and full information feedback groups. The competitiveness ratio (MUCR) was plotted against the corresponding bidding rounds to explore the differences in bidding trend between the two groups. The second part discusses
Soo, A and Oo, B L (2010) ‘The effect of information feedback in construction bidding’, Australasian Journal of Construction Economics and Building, 10 (1/2) 65‐75
69