Academic journal article Economia

Downward Wage Rigidities in the Mexican Labor Market: 1996–2011

Academic journal article Economia

Downward Wage Rigidities in the Mexican Labor Market: 1996–2011

Article excerpt

In this paper, we provide evidence of the existence and evolution of downward real and nominal wage rigidities (DRWR and DNWR, respectively) in Mexico in the period 1996–2011. DNWR constrain nominal wage changes to be greater or equal to zero, whereas DRWR constrain nominal wage changes to be greater or equal to a reference point, which could possibly be different from zero. This point is commonly referred to as the focal point for wage negotiations or the wage indexation point. The existence of either type of wage rigidity is typically associated with the labor market institutions in a given country. For instance, legal provisions against reducing nominal wages could result in DNWR. On the other hand, the inclusion of indexation clauses in labor contracts, which explicitly tie wage changes to the inflation rate or any other focal point, would lead to DRWR. Gauging the relative prevalence of the two types of wage rigidity, if they exist, is relevant since rigidities could prevent the adjustment of the labor market to shocks, thus amplifying the effects of such shocks on real output.

Mexico is an interesting case in which to study downward wage rigidities in the period 1996–2011 for several reasons. First, in this period fiscal and monetary policy were successful in stabilizing the Mexican economy after the 1994 crisis. In particular, the Bank of Mexico, the Mexican central bank, adopted a strategy of rendering monetary policy implementation more transparent, keeping a restrictive bias and responding adequately to inflationary shocks.1 This strategy allowed the Mexican central bank to gradually move toward an inflation-targeting regime, which was officially adopted in 2001. As a result, annual inflation decreased from 27.7 percent in December of 1996 to 3.98 percent in December of 2003, and it remained at single-digit levels for the rest of the period. Second, the Mexican labor law, which dates from 1970 and remained virtually unchanged until 2012, constrains wage adjustment through several provisions, for instance by penalizing employers for reducing wages. Third, the collective wage bargaining system in Mexico, which is argued to be relatively centralized, together with the importance of the minimum wage as a reference point for wage negotiations, could contribute to the existence of DRWR.2 In sum, the interaction of all these elements could potentially generate an environment in which DNWR might become more prevalent over time, but DRWR are still pervasive.

Previous studies, which focus mainly on developed countries, use several methods to document the existence and extent of downward wage rigidities: (i) graphical detection of asymmetries and bunching of observations in the histogram of nominal wage changes; (ii) regression-based methods to formally test whether the excess mass at zero or any other point is significant; or (iii) maximum likelihood estimation of a particular type of censored regression model to obtain the parameters of DNWR and DRWR, such as the probability of being in a given wage regime and the focal point of wage negotiations.3 Despite the variety in the methods used, previous studies share some common features. They use microdata from firms, household surveys, or administrative records, and the key variable of interest is the annual change in log nominal wages. The majority focus on job stayers, that is, workers who stayed in the same job from one year to another, to keep the job characteristics constant.

In this paper, we follow the studies that estimate wage rigidity parameters by maximum likelihood. We use a modified version of the Altonji-Devereux model, which was extended by Goette, Sunde, and Bauer to estimate the prevalence of both types of wage in the presence of measurement error in wage changes.4 This model distinguishes between the notional wage change, which is a latent variable capturing the wage change that would be implemented in the absence of any restrictions, and the actual wage change, which might be censored at zero if wages are subject to DNWR or at a non-zero value (that is, the wage indexation point) if wages are subject to DRWR. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.