22 Nov 2013

(by Morten Hviid) Last week, Asda’s marketing chief was reported as saying: “It is strange to me that our competitors are touting how much they match prices and are working towards parity. I find it almost anti-competitive that large competitors are talking about and proud to be matching each other’s prices. We are the only ones standing up and saying we’ll be 10 per cent cheaper” ( 15 November). I have previously pointed out the potential , so it is unsurprising that I agree with him on the first point. I will spell out one overlooked implication of the guarantee offered by Sainsbury’s and partially copied by Tesco below. However, he is a bit too quick to absolve Asda’s price guarantee from potential anti-competitive blame.

Two problems with the Sainsbury price guarantee

The Sainsbury guarantee cleverly restricts attention to branded goods where quality is directly comparable across products. Own-brands differ in quality and in particular in the consumers’ perception of quality so that while no supermarket would be happy to be known as disproportionally more expensive than others on branded goods, being known as dearer on own-brand can be coupled with a perception of superior quality. The guarantee therefore shifts competition from branded to own-brand products where Sainsbury may feel more comfortable competing. Tesco has introduced its own price guarantee, matching prices on branded (as well as other) products.

Next, consider the following thought experiment. What would be the effect on prices if all the main supermarket chains matched the price of branded goods where the suppliers typically have some degree of market power? Clearly the competition between branded goods when sold at different outlets [so called intra-brand competition] would be removed – by switching retailer you could not get a better deal. What would be the optimal response to this reduction in intra-brand competition? This question is very similar to the question raised by the Most Favoured Nation clauses introduced by a number of on-line platforms which have attracted the attention of competition authorities in the US and EU. They point to the following anti-competitive effect. The price guarantee reduces competition at the retail level, if only a little – Tesco, for example, could not gain market share by cutting price therefore has little incentive to do so. With reduced competition downstream, less of an increase in the wholesale price of a branded product would be passed on to retail prices so that an increase in the wholesale price would reduce sales by less than otherwise. Hence, with the reduction in competition it would be worth increasing the wholesale price a little. Since some of the increase would be passed on in retail prices, consumers would be harmed. How big is this effect? That depends on the degree of inter-brand competition – how loyal are we to our preferred brand of, for example, cereals? Only if customers are very willing to switch between corn flakes and shreddies would there be an effective constraint on the manufacturer’s price.

ASDA’s price guarantee can stifle price competition too

So the Sainsbury/Tesco price matching guarantees encourage prices to converge at above the competitive level. Next consider the following thought experiment: what if the optimal cartel prices for supermarkets implied that ASDA, because of its cost or demand base, should have prices on average 10% lower than its rivals? In that case, a price matching guarantee would actively hinder such a set of prices from being implemented. On the other hand, ASDA’s undercutting price promise would facilitate this hypothetical collusive outcome. This is not to suggest that Asda’s guarantee is meant to be, or even inadvertently is, anti-competitive. This is just to illustrate that the anti-competitive effect arises from linking prices at different outlets, and not the narrow details of how these linkages are specified.

To really understand both the direction and size of the effect of price promises, we need empirical evidence, which has so far proven hard to come by because the data needed are not readily available. We do know that anti-competitive effects are a theoretical possibility. It is interesting to see that at least one member of the retail sector senses this too.