Restaurant Prices and the Mexican Peso.

• Author: Fullerton Jr., Thomas M.

Thomas M. Fullerton Jr. [*]

Roberto Coronado [+]

Of prime interest to border economies is exchange rate performance and currency valuation. Commonly used tools for this task include purchasing power parity (PPP) nominal benchmarks, and inflation-adjusted trade-weighted indices. The latter have the advantage of relying on commonly available international macroeconomic data but overlook microeconomic information that may offer additional insight to issues surrounding exchange rate policy debates. Other efforts have utilized small samples of international product price comparisons to shed light on currency valuation questions. This paper develops one such tool by repeated sampling of prices charged for identical menu items sold at restaurant franchises in El Paso, Texas, and Ciudad Juarez, Chihuahua. A battery of statistical tests indicate that the international currency value of the peso consistently differed from the exchange rate implied by the border region restaurant price ratios in 1997, 1998, and 1999.

1. Introduction

   Balance-of-payment concerns facing Latin American economies have caused international currency valuations to receive ongoing scrutiny throughout the past decade and a half. Given its importance to regional economic performance, border areas between countries such as the United States and Mexico, Canada and the United States, Colombia and Venezuela, and Ecuador and Colombia also monitor exchange rate developments very closely. Typically, one of two types of monitoring tools have been utilized for formal econometric work in this branch of international economics: nominal purchasing power parity (PPP) indices and trade-weighted real exchange rates. Both of the latter rely on easily obtained macroeconomic data.

   A widely cited exception to those macroeconomic data approaches is published quarterly by The Economist magazine utilizing hamburger prices from a well-known restaurant corporation with worldwide franchises. The research at hand develops a similar tool utilizing a variety of international restaurant franchises with operations in El Paso, Texas, and Ciudad Juarez, Chihuahua, on the border between the United States and Mexico. More than 70 signature menu item prices are sampled every month in order to develop an exchange rate measure based on the so-called law of one price hypothesis. Parametric and nonparametric statistical testing is then conducted to examine whether the border restaurant price ratio bears any relationship to the exchange rate value of the peso over the 36-month sample period for which data are available (July 1997-June 2000).

   Section 2 of the paper includes a literature review. Data and methodology are described in section 3. Section 4 provides an empirical analysis of the results. Concluding remarks and suggestions for future research are included in section 5.

2. Literature Review

   Exchange rate monitoring has become a staple of international corporate and public policy analysis in the two-plus-decade period subsequent to the demise of the Bretton Woods fixed exchange rate system. A common approach is the measurement of trade-weighted real exchange rates designed for country risk analysis and balance-of-payments pressure gauging. Such indexes will generally indicate overvaluation with respect to a preselected base with values in excess of 100 and undervaluation with values below 100 (Batten and Belongia 1986; Fullerton 1989). Another practice is the calculation of theoretical nominal exchange rates under a PPP approach (Beltran del Rio 1999). The latter are often popular among financial market analysts because they are defined in the actual units in which trades are conducted.

   International restaurant franchise menu pricing has given rise to a new empirical variant on the law of one price in recent years. The Economist magazine runs a periodic comparison of hamburger prices to nominal exchange rates for a group of 20-plus countries (The Economist 1998). Hamburger prices in local currency units are divided by a corresponding dollar price. Whenever the resulting price ratio is higher than the exchange rate, it implies that the local currency is overvalued. When the ratio is below the prevailing exchange rate, it implies the opposite. The popularity of this readily identifiable index has spawned several subsequent studies that study both geographic and time-series characteristics and implications of the approach. While Cumby (1996) and Ong (1997) find fairly good evidence in favor of the hamburger index, Pakko and Pollard (1996) conclude that nontradability and institutional factors cause it to perform unreliably.

   A growing number of studies examine the PPP principle using data that involve more than one product and take into account geographic factors such as distance (Engel and Rogers 1996; Parsley and Wei 2000). In one such effort, James Gerber sampled individual menu items from seven restaurants in San Diego and Tijuana (Regents of the University of California 1997). Results reported therein indicate that the implied valuation for an exchange rate will vary significantly depending on which menu item is selected for comparison. Jenkins (1997) utilizes price data for 22 separate goods and services consumer price indices for six sets of cities in the United States and Canada. Evidence is uncovered that indicates that deviations between implied and actual exchange rates may arise because of menu costs, taxes, and nontradability. Similar results have also been reported for a cross section of products in the United Kingdom (Fraser, Taylor, and Webster 1991). Geographic commercial factors may also cause at least temporar y deviations between regional price comparisons and national exchange rates (Clark, Sawyer, and Sprinkle 1997, 1999, 2000). Differences in industrial composition and concomitant effects on business cycle phase responses generally contribute to such divergences.

   There are other reasons why regional and national price ratios may frequently deviate from prevailing exchange rates. Currency market overshooting can result from interest rate disparities combined with asset market and goods market adjustment differences (Dornbusch 1976). Sluggish price movements, menu costs, and nontradability may also prevent cross-border price comparisons from matching exchange rates in a statistical sense (Rogers and Jenkins 1995). Given these considerations, it would not be surprising to discover that the sample data collected in El Paso and Ciudad Juarez do not generate overwhelming support for the PPP hypothesis. The latter possibility is especially relevant due to the fact that three years of data represent a fairly short sample for examining what is generally regarded as a long-run characteristic of frictionless currency markets.

   Income level disparities may also contribute to divergences between restaurant price ratio data and currency values. Comparative price levels for identical products generally vary with national income levels (Summers and Heston 1991; Heston and Summers 1996). This may result in part from an interaction between capital-labor ratios and service prices (Bhagwati 1984). Despite...