Price convergence in US cities: a cointegration approach with two structural breaks

Abstract

This article investigates how the price indices of major cities of the US respond to the shock from a city and from monetary policy. We find that the crisis of Bretton Woods system in 1968 and the oil crisis in 1974 should be incorporated as structural breaks in monetary policy variables and price indices. Using cointegration technique with structural break in our aggregated data, we find that the average half-life is 1.75 years, which is closer to what some of others found in disaggregated data, and that the interest rate is an effective tool for controlling cities’ price in short run.

Keywords:

• price convergence
• half-life
• cointegration
• structural breaks
• US cities

JEL Classification::

• C32
• E52
• F42
• E31

Acknowledgements

Sung K. Ahn's research was supported by the Korea Research Foundation Grant (KRF-2005-070-C00022) funded by the Korean Government (MOEHRD).

Notes

¹ For middle income countries, Carrion-i-Silvestre et al. (2004) examined the price convergence in Spanish cities, while Vargas-Tellez (2008) and Sonora (2005) studied the price convergence in Mexican cities. In the case of developing countries, Morshed et al. (2006) investigated the consumer price behavior of 25 major cities in India, while Rangkakulnuwat and Ahn (2006) examined the consumer price index convergence of five regions in Thailand.

² The half-life is defined by the period in which the marginal change in the stationary component of the impulse response becomes half of the initial response.

³ For earlier contributions, please see Camarero et al. (2000), Trivez (2001), Roger (2001), Holmes (2002), Sosvilla-Rivero and Gil-Pareja (2004), Friberg and Matha (2004) and Jenkins (2004),

⁴ See http://epp.eurostat.ec.europa.eu/portal/page?_pageid=2714,1,2714_61582043&_dad=portal&_schema=PORTAL.

⁵ Bernard and Durlauf (1995) examined output convergence while we examine price convergence.

⁶ There exists an extensive literature on the testing of stationarity with structural breaks. For earlier contribution, see Perron (1989), Zivot and Andrews (1992) and Perron (1997), and for more recent studies, see Lee and Strazicich (2001), Lumsdaine and Papell (1997), Lee and Strazicich (2003), Bai and Perron (2003).

⁷ The method to compute estimated parameters of β, δ, and the breaks dates (T ₁, … , T_m ) are explained in detail in Bai and Perron (2003).

⁸ This measure is not without bias and criticism (see, e.g. Seong et al., 2006).

⁹ We ignore the deterministic terms here, since the converting does not change the parameter of deterministic terms.