Skip to content

A 10% SME price preference delivers redistribution at near-zero static price cost

Intuition (plain-language)

There is a cheaper way to favor SMEs than locking everyone else out. Give SMEs a 10% price handicap but keep non-SMEs in the room: the disciplining bidders stay, the government still pays the real winning bid, and SMEs win more often. The static price cost relative to open bidding is essentially zero. It buys less redistribution than the full set-aside — but the real policy choice is not "support SMEs or not," it is "exclude rivals" versus "tilt the allocation while keeping the price-forming pool intact."

🟡 A static 10% SME price preference simulated under the recovered cost primitives keeps non-SME bidders eligible and disciplining the price-forming order statistic. The resulting price shift relative to the open benchmark \(S_1\) is essentially zero in standardized non-pharma, while the SME win-rate rises by a meaningful margin (AN-012). The preference is not a free version of the set-aside — it delivers less redistribution (the SME win-rate rise is smaller than under the full SME-only regime) — but it shifts the static welfare frontier in a way that the exclusionary regime cannot match: the price-forming pool remains intact, the government pays the actual winning bid, and the implicit transfer to SMEs is delivered through allocation tilt rather than through bidder elimination.

The headline reading is that the relevant policy frontier is not SME support versus no support — it runs between exclusionary redistribution and support that preserves the price-forming bidder pool. The 10% preference is a static design benchmark of the second branch, not a forecast of a legal preference regime.

Caveat. The preference benchmark holds entry fixed at the pre-policy level by construction. If non-SMEs would systematically exit when preferred against (because they cannot expect to win), the preference behaves like a soft set-aside and the static-cost result collapses. The simulation does not test entry elasticity; an entry-elastic version would be a separate exercise. The 10% margin is the cleanest benchmark; the diminishing-returns curve and the preference level at which non-SMEs start dropping out is in the optimal-preference search (v7-jpube-tight/scripts/61_optimal_preference.R, not yet a standalone AN).

Sources.

  • Own analysis: AN-012 (preference benchmark simulation); AN-011 (welfare arithmetic it is compared against).
  • Reports: none direct.
  • News anchors: none direct.
  • Cross-refs: H:preference-near-zero-cost; H:static-welfare-loss-large; paper §5 (Set-aside versus preference).
  • Validation: v7-jpube-tight/scripts/61_optimal_preference.R and 60_distributional_incidence.R for the optimal-preference search and the within-SME distribution of the win-rate gain.