Imagine you run a toll road where cars under a certain weight cross free but heavier ones pay a fee. You'd expect a suspicious number of trucks to show up at 9,999 pounds. Now imagine the weight limit changes every January and is different for every road — if the trucks keep clustering just under each road's own limit, you know it's not coincidence. That's what this paper does with Colombian public procurement. The committed claim: Colombian public buyers systematically manipulate estimated contract values to stay below legal thresholds that trigger competitive bidding, and this manipulation has real costs — fewer bidders, smaller discounts, and more post-award contract enlargements. The authors find 49,631 excess procurement processes bunched just below the small-purchase threshold and 3,157 below the tender threshold, across 8.6 million SECOP II records from 2011–2026. What makes this paper genuinely sharp is its identification strategy. Colombia's thresholds are entity-specific (five budget bands set by law) and shift every January with the minimum wage, while round numbers stay fixed. This separates threshold-gaming from round-number bunching — a confound that plagues most bunching studies. The spike tracks the threshold, not the round number. It appears only at each entity's own band. And it's an order of magnitude smaller for procedures that carry no value cap. That's a clean triple-difference. Territorial governments run by elected officials bunch more than appointed national agencies, suggesting political incentives drive discretion-seeking. The cost measurement is where the paper earns its keep. Contracts bunched below the tender threshold attract half as many bidders as those just above, obtain smaller price discounts, and are more frequently enlarged after award — a classic pattern of strategic underestimation followed by change orders. This isn't just gaming for convenience; it degrades procurement outcomes in measurable ways. A second natural experiment strengthens the story. Colombia bans direct contracting for four months before presidential elections. During each of four electoral cycles, direct contracting falls 95% during the ban — but surges 209% in the two weeks before it takes effect, with partial substitution toward below-threshold procedures. Buyers respond to the timing constraint the same way they respond to the value constraint: by gaming the margin. The external validation from Chile is a nice touch. Chile's fixed UTM thresholds show the same bunching pattern, and a threshold abolished in 2024 shows no bunching afterward — the clearest possible placebo test. The replication code and data are included as ancillary files. The paper doesn't estimate the total fiscal cost of threshold manipulation, and it doesn't run a structural model of optimal threshold design — both obvious next steps. But as a measurement exercise, this is unusually clean. The triple identification (entity-specific bands × minimum-wage indexing × round-number separation) is stronger than most bunching papers in the literature. The practical implication is straightforward: if you design procurement rules with sharp value thresholds, buyers will manipulate values to stay below them, and the cost shows up in reduced competition and inflated post-award contracts.
