Nordic Food Markets - Appendix 2. What explains price variations
Face - Contents - Bottom - Previous/Next "Nordic Food Markets" Appendix 2. What explains price variationsModelling the price decision A number of empirical studies have focused on the price decision of grocery stores in order to explain the competitiveness of the stores. According to these studies one way to employ a price function takes the following form:
The price function consists of the marginal cost, where cost function (TC) is a function of the quantity sold by the store (Q), the number of stores belonging to the retail group (#stores), and the interaction terms between quantity sold, number of stores and time. The variables potentially affecting the mark-up of a store are divided into three groups: store characteristics, structure of the market area and demand conditions in the market. The variables describing store characteristics are the quantity sold, rental costs of the store (proxy for good location), cooperative indicator, and a service level indicator. The intuitive reason for these variables is that stores do not offer physical products only but distribution services as well. The variable describing the market structure are the Herfindahl index among stores, the Herfindahl index among groups and the capacity share of the retail group of stores (share). The parameter estimates of all these variables are usually either positive or zero. The variable describing the demand conditions in the market are the income and the time trend. These variables do not explain the competitiveness of the stores, but they are needed in the pricing equation to ensure that the parameters of the variables explaining competitiveness are unbiased. Factors which affects the relative price level When looking at the price variation among countries it depends on a number of macroeconomic factors and other factors. The price of the currency naturally affects the relative price level for private consumption. The reason is that national prices fluctuate much more slowly than exchange rates. It is obvious that the short-term variations in exchange rates may have a large effect. The long-term structural price level differences, however, hinges upon other factors. Exchange rates, which basically are the price of a currency, affect directly the results of a price level comparison between countries. For various reasons, exchange rates can stay for prolonged periods on levels that are not in equilibrium. In Sweden for example, the krona has for a considerable time been considered undervalued by the Riksbank. This means that, ceteris paribus, should the price of the krona increase to its “true” level, the price level in Sweden compared to other countries would be even higher. The Swedish exchange rate does fluctuate compared to other exchange rates, like the Danish. Since the beginning of 2005 the exchange rate has declined by 3-4 pct point. There are many reasons for deviations from equilibrium exchange rates, including trade balance and interest rates differentials, as well as expectations of future changes in fundamentals. Fluctuating exchange rates and different exchange rates across countries are connected with high costs. The introduction of a join exchange rate (like the ) would save the cost of exchange and guarantees for the firms. This might stimulate the competition, because the international trade will be less complicated. By increased competition the prices will decline in the long run. Finnish investigations and investigations from EU can confirm this development. The investigations show that the immediate reaction on the consumer prices from introducing the euro in Finland was a rise in the prices, but after a year the prices fell down again. Today, the Finnish prices have stabilised on the normal CPI from before Finland joined the EU, the food prices are even a little below the normal CPI. Richer countries usually have higher prices. This relationship is consistent with economic theory and is confirmed in Figure A.2.1. As is evident in the figure, the Nordic countries have GDP per capita rates close to or above the EU average and price levels which are even higher. Norway and Denmark have higher gross domestic product growth rates than Finland and Sweden. Compared with EU15, Finland and Sweden are in relative terms mean-income and high-price countries. A number of countries including Germany, Netherlands, Austria, Ireland, Italy, and Belgium have lower prices but higher gross domestic product compared to Finland and Sweden. No country within the EU exhibits higher prices and lower GDP per capita than Sweden and Finland for this year. Figure A.2.1. Price level and real GDP per capita (EU15=100), 2001
Source: SCB and OECD (2002a). The economic rationale for a positive relation as confirmed in Figure A.2.1 is simply that richer countries exhibits higher productivity in the tradable sector, due to higher levels of education and R&D, which increases labour productivity and hence wages. The wage-effect in the tradable sector spills over into non-tradable sector, putting an upward pressure on wages there as well. Danish surveys show that wage differences overall may account for up to 3.5 % of the price difference between Denmark and EU9 . However, several service sectors, which are not subject to competition from abroad, weigh heavily in this comparison. In chapter 4.4 it is argued that when it comes to the food sector the impact on prices from wage differences is equalled out by the differences in productivity and the different retail structure between countries. Does lack of competition explain the price variation? The debate on prices often becomes confusing since a judgement must be made as to whether such factors explain the full price difference between one country and its neighbours or just a part of it. A report by the Swedish Competition Authority (2001) attempts to address this question by analysing the relative consumer price levels of OECD members during the 1990s using panel regression techniques. The price indices were modelled in terms of variables chosen with inspiration from the literature on purchasing power parities, including gross domestic product, the level of taxes, labour costs, changes in private consumption and exchange rates and also population density (to capture variations in transportation costs). The results indicate that about half of the Swedish price difference, which amounts to approximately 20 percent as an average for the 1990s, can be explained by these variables. The remaining half constitutes a “fixed effect” and is not due to these factors. The open question is: to what extent does lack of competition in Swedish markets explains the residual? Unfortunately, no variable describing the efficiency of competition was available for inclusion in the model, which would have enabled us to test this factor directly. However, a somewhat rudimentary variable of industry concentration was derived for a number of sectors in the EU for a few years in the 1990s, which shows that Sweden exhibits comparably high levels of concentration in most cases. The variable was included in the analysis of a restricted sample and the results indicate that it is strongly significant as a determinant for price levels in Europe. These findings, together with general experiences gained during the last ten years, led the Authority to conclude that weak competition in Sweden represents up to half the price difference between Sweden and the EU. This conclusion led to an intensive debate during the 2001 and 2002 in Sweden. Lately, criticism has been aired that the impact of competition was exaggerated (Bergman 2005). A number of empirical studies, most of them applied to US data, have confirmed a positive relationship with concentration and prices using conventional multivariate regression techniques. As a high concentration rate is the most commonly adopted indicator of competition, it is concluded that competition matters – poor competition leads to higher prices. The price-concentration methodology has been employed by the Swedish Competition Authority (2002), revealing substantial regional price differences for the food retail sector. A basket of 1000 food items costs 7 percent less in West Sweden compared to the county of Stockholm. The estimations reveal that competition clearly affects price formation. Physical distance has an influence - prices become lower the smaller the distance to the nearest competitor. However, it is complicated to establish robust models, which compare food prices across countries and reveals the competition condition. Even though it is possible to identify parameters and variable which are similar in all countries there will be some differences which should be modelling individual for each countries. Thus, it is individual markets which are working on each individual premises. In general, we can conclude that the intensity of competition does play a role in explaining food price differences, along with a number of other factors. Version 1.0 December 2005 • © Danish Competition Authority. |
