Three Essays on Procurement Auctions
Bibliographic record
Abstract
This thesis is an empirical study of procurement auctions. First, I explore how varying the timing of a sequence of auctions affects both bidder behaviour and the welfare of procurers and bidders. We develop a structural auction model with endogenous participation in which bidding may be simultaneous or sequential. Bidders perceive auctioned objects as either complements or substitutes. We apply this model to auctions for roof-maintenance projects in Montreal. We show that complementarities account for as much as 17\% of the total size of contract combinations. We develop an algorithm to search a schedule of auctions and show that the total cost of projects can be reduced by over 8\%. I also study the effect of corruption on outcomes of procurement auctions. I define corruption as any illegal behaviour used by firms to lower their costs. In Quebec, firms that are found to be involved in corruption are not allowed to bid in procurement auctions for a period of five years and added to a list. I consider the outcome of auctions in markets where firms on that list were active. Bids by corrupt firms are lower than bids by clean firms and corrupt firms are more likely to win auctions in which they participate. The behaviour of clean firms adjust as they interact with corrupt firms. Costs of corrupt firms recovered through structural estimation are shifted down by around 4.1\% relative to clean firms. Only half of this advantage is reflected in their bids. Finally, I run a case study of corruption in procurement auctions. The English Montreal School Board (EMSB) was placed under trusteeship by the government of Quebec in November 2019, following underreporting of contracts that the EMSB was legally required to report. Using public data, I confirm underreporting in the period that precedes the trusteeship and I find that the value of contracts reported by the EMSB after the trusteeship is in line with that of other school boards. Using a Difference-in-difference approach, I confirm that the underreporting disappears as the trusteeship begins. Simulations show damages of around \$16 millions between 2009 and 2015.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.029 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.005 |
| Science and technology studies | 0.003 | 0.006 |
| Scholarly communication | 0.006 | 0.007 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.003 | 0.004 |
| Insufficient payload (model declined to judge) | 0.033 | 0.006 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".