Mirror, Mirror

One of the hopes for the single European market was that it would benefit consumers by enforcing the ‘law of one price’, where identical traded goods sell for the same price everywhere. New research on the comparative prices of a well-known furniture retailer reveals that there is still a long way to go.

International price comparisons have moved from the preserve of statisticians to headline news. Newspapers send correspondents around the world to discover that for anything from CDs to restaurant meals, the United States is a buyer’s paradise. Prices in the European Union, above all in the UK, always seem to be much higher. Indeed, it is rumoured that the UK is known as ‘Treasure Island’ by those who sell such goods as motor cars and food there.

These large price gaps have become a matter of growing public concern. The UK government, for example, has responded by commissioning a comparative price study. But international price differences matter not just because of what happens when the people of one country feel they are being swindled. As the Japanese know all too well, high prices reduce purchasing power, thereby reducing consumption – and imports – and so annoying a country’s trading partners.

In truly competitive market places, price differences create arbitrage opportunities that limit divergences – gold prices in London and New York, for example, differ only marginally. So persistent price differences indicate a lack of effective competition, casting doubt on the much discussed globalization of product markets.

But making these price comparisons properly is a tricky business. You need a product that passes three key tests: first, the product has to be exactly the same everywhere; second, it must be tradable; and third, since discounts and special offers mean that prices vary within a country, it needs to be sold at the same price in all shops in each country. Restaurant meals and most services fail these tests, as do cars (due to emission and other standards), television sets (because of different standards) and a number of other manufactured goods. The Economist’s famous Big Mac index suffers from the fact that hamburgers differ, even if only slightly, across national boundaries. They also require inputs whose prices differ from country to country.

Is it possible to find goods that are identical, tradable and priced the same within a country? In a recent CEPR Discussion Paper, Jonathan Haskel and Holger Wolf suggest a simple answer: IKEA. The Swedish household furniture retailer now has stores in 25 countries. The traded goods it sells are identical and come from the same manufacturer. In Europe, the 1998 catalogues are exactly the same in each country aside from the language and local prices. And the goods are sold within each country at a single catalogue price.

Haskel and Wolf obtained an IKEA catalogue from a store in each country where the firm operates. They tabulated the prices of 119 basic non-designer goods, from chairs to kitchenware and converted those prices at contemporaneous exchange rates into US dollars. The table shows their findings on the price of mirrors for a selection of the countries. The Guldros, for example, a simple wood-framed round mirror, which sells for $99 in the United States, is most expensive in Denmark, where it is priced at $119. The UK too is relatively expensive, with the same mirror selling for $115.

Are these price differences proof of a lack of competition? Not necessarily. Retail prices reflect not only the traded good itself, but local rents, wages, taxes, tariffs and other costs that competitors would have to bear. So local price differences can occur even in competitive markets. Herein lies the basic problem in using any international price comparisons for testing for market power.

Haskel and Wolf found a way of using their IKEA data to get round this difficulty. If high local prices were due only to high local costs, then the local prices of all goods sold by IKEA should be relatively higher. So the test for market power is to compare the relative price of Guldros mirrors with the relative price of other mirrors sold by IKEA – the ‘double relative price’.

The second column of the table shows the prices of IKEA’s Alg square mirror: Denmark, where the Guldros was most expensive, is now the least expensive; and in the UK, where the Guldros was more expensive than in the United States, the Alg is cheaper. This suggests that price differences are about more than just local costs; they also reflect local market power that enables IKEA to set different prices.

What determines the extent of these price differences? It is of course unlikely that consumers will travel to stores in other countries to find cheaper IKEA products. But inter-country price differences do signal the opportunity for other firms to compete. One would expect these pressures to increase the closer that countries are together, if they are members of the European Union, if they share a common language, and so on.

The dispersion of relative prices for the 119 goods Haskel and Wolf have analysed shows just this pattern. Price differences are smaller between adjacent countries and between countries with larger markets, where fixed costs of entry may be spread more easily.

These price differences are a snapshot, taken at one moment in time. Do they tend to persist? By looking at catalogues over a period of years, an interesting pattern emerges. When international prices are not too different, they converge but only very slowly. But when relative prices get wildly out of line – by about 150% – prices converge very rapidly. So market power does enable price differences to be maintained. Yet when they get too large, the discipline of competition undoes them.

Haskel and Wolf’s study sends a message about the value of international price comparisons. First, because of variations in local costs, simple relative price differences do not say very much about competition or the lack of it. Studies should look at double relative prices, that is to say, relative prices for similar goods. Second, static studies can be misleading since they ignore the process of relative price adjustment. Competition tends to work its magic but can do it only slowly.

This article discusses research reported in ‘Why Does the Law of One Price Fail? A Case Study’ by Jonathan Haskel and Holger Wolf, CEPR Discussion Paper No. 2187 (July 1999). Haskel is at Queen Mary and Westfield College, London and a Research Fellow in CEPR’s Labour Economics programme; Wolf is at Georgetown University, Washington.

Prices of Round and Square Mirrors at IKEA in 1998 US dollars

The Guldros is a round mirror, 59 x 78cm, with bevelled glass; 
the Alg is a two-pack of square mirror tiles, 45 x 60cm. Source: IKEA catalogues.