Most studies compare markets of type A to markets of type B or markets of type C. We instead look at different market structures (A, B, C) to disentangle market power effects from efficiency gains (see table above).
In type A markets, both the acquirer and the target (R1 and R2) are present, The market is likely to experience price raises from market power and price declines from efficiency gains. In type B markets, only one of the two (acquirer R1 or target R2) are present and we expect efficiency gains but no changes in local market power. Type C markets are unaffected by the merger because neither the acquirer nor the target operate in these local markets.
Market power effects are identified by comparing type A markets to type B markets. Post-merger price increases in regions where acquirer and target were active compared to regions where only one of the two operated (see graph above for our results).
Efficiency gains are identified by a comparison of type B markets with type C markets. Thus, we compare post-merger price declines in regions where acquirer or target was active compared to regions where none of the two operated. (see graph above).
Left: Mechanism of "Invest-and-Harvest". Attract new consumers with free offers and raise prices when they are locked-in due to switching costs. This mechanism is present in the diaper market, where consumers are inert and firms can distinguish between consumers that are new (babies) and mature (toddlers)
Middle: The paper shows high repeat purchase probabilities of 60-80% for the brand (increasing over the consumer life-cycle) and 30% for the no-name (stable). Over the consumer life-cycle, the prices for the brand increase much more than the prices of the no-name.
Right: Reduced-form evidence of the existence of dynamic ricing incentives. This graph shows that firms react in prices when there is a sudden event in the market, such as brand exit or consumer tests for diaper quality. When consumers learn that brand and no-name are of the same quality, then prices converge due to more switching of consumers that have been locked-in before.
Left: Higher cartel overcharges and longer cartel duration when upstream cartel uses RPM
Right: RPM (solid black) plotted against actual retail price (horizontal dashes) that decreases at high demand holidays (vertical lines)
SSNIP market definition test at upstream level under two types of vertical contracts, IJIO 2020 [LINK]
Left: Assuming full pass-through from wholesale to retail price leads to large market definition when doing the SSNIP at upstream level
Right: Assuming incomplete pass-through from wholesale to retail price leads to smaller markets when doing the SSNIP at upstream level