Statistical screening for cartel-like pricing · generated 2026-06-19
How often rivals selling the same product post the very same price.
Spread of prices across retailers per product. Healthy competition shows dispersion; a cartel compresses it (bars piled on the left = suspicious).
Share of shared products where the two chains post an identical price.
| Retailer A | Retailer B | Shared | Identical | Ratio CV |
|---|---|---|---|---|
| kritikos | masoutis | 1117 | 78.3% | 0.1825 |
| masoutis | synka | 3482 | 45.8% | 0.0905 |
| kritikos | synka | 1032 | 44.1% | 0.1038 |
| masoutis | sklavenitis | 1821 | 38.4% | 0.2045 |
| colruyt | delhaize | 42 | 33.3% | 0.1359 |
| kritikos | sklavenitis | 1186 | 31.6% | 0.256 |
| galaxias | masoutis | 3278 | 26.8% | 0.2191 |
| lidl | sklavenitis | 490 | 25.1% | 0.2085 |
| galaxias | kritikos | 1109 | 23.8% | 0.2615 |
| mymarket | sklavenitis | 1927 | 21.5% | 0.1665 |
Clustering at psychological endings (.99/.95/.00) hints at shared recommended pricing.
Where to investigate — not findings of wrongdoing.
Same product on offer at multiple chains at once.
Built from a single price snapshot via the public PosoKanei API. The screens follow the empirical "collusion markers" competition authorities use to scope investigations: price identity, low dispersion, pairwise matching, and focal-point clustering. A near-constant price ratio between two chains (low ratio CV) is consistent with resale-price maintenance / MSRP rather than a horizontal cartel.
None of these metrics prove collusion. Rule out common-cost shocks (wholesale, energy, FX, VAT) and common ownership before drawing conclusions, and corroborate with time-series analysis of synchronized price changes.