Skip to content

Open auctions increase bidder participation by ~10–22%

Intuition (plain-language)

Drop the SME-only fence and more firms show up to bid — about 22% more in the first six months, settling toward 10% over a year and a half. More bidders is the proximate mechanism behind the price drop: a thicker pool means the order statistic that sets the price is drawn from a tougher field. Part of the fade over the longer window reflects the control groups themselves being phased into SME-only, not the entry effect dying out.

Reduced-form motivation layer

The headline number on this page comes from the v1–v4 reduced-form DiDiR pipeline. The v8 manuscript carries this as motivation in §1 but does not headline it; the v8 canonical claim is the structural decomposition — see Exclusion dominates the price decomposition and Static welfare cost ~28.9%.

🟡 Opening switched group 65 to non-SME bidders increased the average number of participating firms by ~22% in the 6-month window, attenuating to ~10% in the 18-month window, identified by DiDiR around March 2018 (AN-002, p<0.01 across all specifications). The bid-count margin moves in the same direction.

The attenuation across windows reflects the control groups' gradual SME restriction rollout — the placebo on firm counts is significant in the pre-period (AN-004), but the placebo on the price margin is null. The reading is therefore that the firm-count DiDiR conflates the group-65 policy switch with a secular trend in the control groups, while the price DiDiR isolates the policy effect itself.

Event study log firms

Event study (figure A.3 / fig_03_numfirms_es): semester-by-semester group-65 vs control gap in log number of bidder firms. Pre-period gap narrows already under the secular trend; the post-period jump is the mechanical result of removing non-SME bidders from group 65.

Caveat. The reduced-form firm-count effect combines two margins: the removal of non-SMEs and the entry response of additional SMEs. The DiDiR cannot separate them. The structural decomposition does — see The protected SME pool responds but does not replace lost discipline — and finds that the SME pool roughly doubles its participation, but not enough to recreate the lost discipline.

Sources.

  • Own analysis: AN-002 (DiDiR firm-count tables), AN-004 (firm-count placebo and its interpretation), AN-010 (structural decomposition of the bidder-count response).
  • Reports: PGE-SP opinion (March 2018) anchors the institutional break.
  • News anchors: none direct.
  • Cross-refs: H:entry-thickens-pool; docs/results.md.
  • Validation: scripts/02_analysis.R → output/tables/tab_participants.tex and tab_validbids.tex.