Abstract
This Selected Issues paper analyzes the global crisis and potential growth in Mexico. The paper uses two methodologies to assess to what extent the global crisis is likely to weigh on Mexico’s growth potential. The first approach is sectoral, examining the historical relationship between financial stress and growth in manufacturing industries. The second approach uses a growth-accounting framework to take a closer look at likely developments in the factors that drive potential growth. The paper also examines spending patterns and risks of expenditures volatility from unanticipated shocks in macroeconomic variables under two alternative fiscal rules.
II. Labor Market Informality and Macroeconomic Performance1
1. More than half of the Mexican labor force is employed in the unregulated informal sector as workers or owners of predominantly small firms. While this share of informal employment is in the middle of the Latin American distribution, it is a multiple of the average among Mexico’s OECD peers (OECD, 2008, Perry et al, 2007). Since a common view of informal employment in Mexico as well as in other emerging markets is that of a disadvantaged labor market segment arising in response to rigidities in the formal sector, a high prevalence of informality could have important implications for macroeconomic performance. In a recession, the informal sector can then be thought to bear the brunt of the adjustment needed if inflexible labor contracts in the formal sector hinder lay-offs and downward pressure on wages. Over longer time horizons, high and persistent informality has also often been pointed to as an important factor behind the disappointing productivity growth in Mexico and other Latin American countries as informality for example severely limits access to finance and adoption of new technologies for firms, and adequate training for employees (Perry et al, 2007).
2. Labor market surveys in Mexico enable interesting comparisons between the formal and informal sectors of the labor market. The quarterly INEGI employment surveys are available for the period 1987-2009, thus covering the 1994 peso crisis, the years of trade liberalization following NAFTA and the current global crisis. They have an extensive coverage of informal and formal labor markets in terms of wages, hours worked etc, are representative of the city and state level and used by the government to generate unemployment statistics. In accordance with previous literature on informality in Mexico, for what follows an individual is considered to be in the informal sector if he/she is not registered at a social security institution or lacks medical insurance through the job.2 Originally comprising only urban areas, rural municipalities were added to the surveys during the last decade. We stick to the original urban sample throughout to ensure consistency, although results hold up if the whole sample is used.3
3. Like other emerging markets, the Mexican labor market is both similar and different from mature economies. As in industrial countries Mexican unemployment increases when output falls in economic downturns; this countercyclicality is further shown by a negative contemporaneous correlation between the HP-filtered cyclical components of unemployment and GDP.4 Labor force participation increased in Mexico during the period under study, largely through women entering the workforce. Different from advanced countries but similar to other emerging markets, labor force participation is weakly countercyclical over the whole period of study meaning that households typically increase their labor supply in economic downturns. This aggregate view of the labor market however masks some important differences between the formal and informal sector.
4. When Mexico faces economic downturns, the informal sector tends to buffer the blow to the formal sector. As seen from the left-hand panel of Figure 1, the informal sector in Mexico typically increases its share of employment in recessions, something that is further confirmed by a negative correlation between the high-frequency variations in informal sector employment and GDP. Although the relative shares of formal and informal employment have been quite stable over the last two decades, the spike in informality induced by the 1994-95 recession took over three years to subside. It is noteworthy that informal employment and unemployment have increased less during the current crisis compared to then despite similar declines in output. Possible explanations for this positive outcome include the corporate sector’s stronger resilience to shocks, the authorities’ policy response and to a lesser extent declines as opposed to increases in labor force participation.5
Informal Sector Employment and Wages over the Business Cycle
Left panel: Cyclical component of GDP (left axis) and informal employment (right axis), as measured by the HP-filter. Right panel: Formal-informal sector wage differentials (left axis) and unemployment (right axis), in percentage points. Series in the right-hand chart have been smoothed using a three-quarter moving average.
Citation: IMF Staff Country Reports 2010, 070; 10.5089/9781451981834.002.A002
Source: INEGI, IFS and staff calculations.Informal Sector Employment and Wages over the Business Cycle
Left panel: Cyclical component of GDP (left axis) and informal employment (right axis), as measured by the HP-filter. Right panel: Formal-informal sector wage differentials (left axis) and unemployment (right axis), in percentage points. Series in the right-hand chart have been smoothed using a three-quarter moving average.
Citation: IMF Staff Country Reports 2010, 070; 10.5089/9781451981834.002.A002
Source: INEGI, IFS and staff calculations.Informal Sector Employment and Wages over the Business Cycle
Left panel: Cyclical component of GDP (left axis) and informal employment (right axis), as measured by the HP-filter. Right panel: Formal-informal sector wage differentials (left axis) and unemployment (right axis), in percentage points. Series in the right-hand chart have been smoothed using a three-quarter moving average.
Citation: IMF Staff Country Reports 2010, 070; 10.5089/9781451981834.002.A002
Source: INEGI, IFS and staff calculations.5. Increases in informality are to a large extent driven by severe recessions rather than moderate cyclical swings. Although we find informal employment to be countercyclical with respect to GDP over the whole period under study, this finding is mainly driven by the large increases in informality following the peso crisis in 1994 and to a lesser extent the 2008-2009 recession. For shocks to GDP that are more like the “normal” business cycles of the period, we find the response of the informal sector employment share to be of the expected sign but mostly insignificant in a standard unrestricted VAR framework.6
6. Differences in hours worked and wages between formal and informal jobs are important mechanisms for the adjustment to shocks. While formal and informal sector average hours worked per week are very similar in terms of levels and cyclical behavior, informal hours worked have more volatile short-run fluctuations. This suggests that the adjustment of hours worked in the informal sector over the business cycle is a response to the rigidities in the formal sector pointed out by Chiquiar and Ramos-Francia (2009) and others.
7. Wage setting in the informal sector is in some ways similar to the formal sector… There is a long-standing debate regarding the impact of minimum wage increases on the wage distribution in Mexico (Fairris et al, 2006). While only a small fraction of formal sector workers earned wages smaller than the minimum during the period under study, we find that up to a third of formal sector employees had wages set as quite exact multiples of the prevailing minimum wages.7 To a somewhat lower extent this also held for informal sector wages, which is interesting given that minimum wage norms are only binding for the formal sector. We also found that minimum wage increases typically feed through to informal sector wages with a one-quarter lag. That is, while changes in minimum wages were implemented immediately for formal sector workers they are typically delayed for informal workers.
8. …but informal sector wages are more responsive to macroeconomic shocks. Informal wages fall relatively more in recessions but catch up in booms when the formal labor market recovers, as evidenced by a negative correlation between the formal-informal sector wage gap and the cyclical component of GDP. In line with Alcaraz (2009), we also find a negative correlation between the cyclical components of unemployment and informal sectors wages while formal sector wages do not seem to respond to deteriorating labor market conditions. When conducting a simple unrestricted VAR exercise, we find that a shock to unemployment has a significant negative effect on informal sector wages reaching its maximum impact after two quarters.8
9. Neither formal nor informal sector wages are likely to recover quickly after the current crisis. After the 1994-95 crisis, the reduction in the formal-informal sector wage gap was slow and gradual, as seen from the right-hand panel of Figure 1. From the figure, it is evident that while unemployment and the formal-informal sector wage gap have tended to move together, during the recovery from the 1994-95 crisis unemployment fell faster than relative wage differentials as activity rebounded. This time around, the output gap is projected to close only gradually over the medium term, and although there are signs that labor demand is beginning to resume it will take time before the formal sector fully recovers which in turn can allow informal sector wages to begin picking up in relative terms.9
10. The informal sector not only works as a shock absorber but also affects Mexico’s long-run economic performance. There are strong indications from labor market surveys that firms with informal employees are much smaller than those with formal jobs. Moreover, many self-employed in the informal sector are older workers who return to this sector to start their own companies after a career in the formal sector (Duval-Hernandez and Orroca Romano, 2009). It is unfortunate that these workers seem to perceive formal-sector barriers to the start-up of companies as too high to operate there when they decide to leave formal salaried work for self-employment. Hence both for existing firms and start-ups, the combination of small-scale activities and informality reinforce each other in hindering access to external finance. This prevents firm growth and investment in new technologies, which have important links to improved productivity.
11. Reducing informality would likely lead to productivity gains and boost output… Mexico’s average productivity growth was about 1 percent per annum during the 2000s, far below many OECD peers.10 It is conceivable that part of this weak performance is due to the prevalence of a large informal sector; some tentative evidence is provided through the negative correlation between the growth rates of the informal sector employment share and manufacturing productivity (if we use the latter as a proxy for total factor productivity). The labor market surveys we have used in this paper do not provide information about output or sales and hence separate productivity series for the formal and informal sector cannot be constructed. The best available estimate of formal-informal sector productivity gaps is thus the wage differential between the sectors, if we assume that wages reflect workers’ efficiency. For all sectors, a wage gap of around 15 percent remains between formal and informal employment after controlling for observable and non-observable worker characteristics such as age and educational level.11 Using this wage gap as a proxy for productivity and the historical average size of the informal sector, the current cost of informality is likely at most 9 percent of GDP.12 The gains from reducing informality would however be larger than that as many of the possible measures to create incentives to move from informal to formal employment would also improve the productivity of already existing formal sector jobs.
12…. but poses considerable challenges in terms of identifying and implementing the key economic reforms. Both in Mexico and other countries barriers to the adjustment of formal sector employment such as restrictions on short-term contracts and high dismissal costs were designed to protect the welfare of workers but can also make it difficult to maintain formal sector employment, and especially so in times of rapid economic change. While the currently proposed labor market reforms include welcome changes in the aforementioned areas, their scope is relatively modest and thus unfortunately not likely to have major effects on the level of informal employment. An important remaining challenge is also to carefully design social programs, including the qualifications for benefits, to provide the proper incentives for the beneficiary to go from transfers to salaried work (Levy, 2008).
13. The productivity loss stemming from informality also depends on the type of economic activity. Informal employment makes up over a fourth of higher-productivity sectors such as manufacturing, and dominates e.g., services and agriculture which are often considered less efficient sectors. Percentage differences in median wages between formal and informal employment differ widely depending on the nature of the activity and ranges from less than 2 percent in agriculture to over 35 percent in the commercial sector. As seen from the figure to the right, the share of formal employment is positively related to the wage gap, apart from the outlier public administration. The highest wage gaps are hence found in the more productive sectors such as manufacturing, mineral extraction and commercial activity and the lowest in less efficient sectors such as agriculture and services. Since the former also contribute over four times more to GDP than the latter, reducing the informal employment share in e.g., manufacturing and mining by one percentage point would probably give more leverage in terms of productivity gains compared to a corresponding change of household services and agriculture jobs.
Appendix 1. Summary Statistics and Correlation Coefficients
As measured by the correlations between the cyclical components of the HP-filter.
This refers to the whole INEGI sample starting in 2002.
| Period Average | Standard Deviation | Cyclical correlation with output/1 | |||
|---|---|---|---|---|---|
| Labor force participation | 54.5 | 3.18 | -0.14 | ||
| percent | |||||
| Unemployment, percent | 6.1 | 1.83 | -0.39 | ||
| Urban informal sector | 38.9 | 3.2 | -0.35 | ||
| employment share, | |||||
| percent | |||||
| Total informal sector | 58.3 | 1.8 | -0.16 | ||
| employment share, | |||||
| percent /2 | |||||
| Formal sector hours | 43.97 | 1.28 | 0.20 | ||
| worked, per week | |||||
| Informal sector hours | 42.97 | 1.57 | 0.17 | ||
| worked, per week | |||||
| Formal-informal sector | 22.4 | 9.0 | -0.17 | ||
| wage gap, percent |
As measured by the correlations between the cyclical components of the HP-filter.
This refers to the whole INEGI sample starting in 2002.
| Period Average | Standard Deviation | Cyclical correlation with output/1 | |||
|---|---|---|---|---|---|
| Labor force participation | 54.5 | 3.18 | -0.14 | ||
| percent | |||||
| Unemployment, percent | 6.1 | 1.83 | -0.39 | ||
| Urban informal sector | 38.9 | 3.2 | -0.35 | ||
| employment share, | |||||
| percent | |||||
| Total informal sector | 58.3 | 1.8 | -0.16 | ||
| employment share, | |||||
| percent /2 | |||||
| Formal sector hours | 43.97 | 1.28 | 0.20 | ||
| worked, per week | |||||
| Informal sector hours | 42.97 | 1.57 | 0.17 | ||
| worked, per week | |||||
| Formal-informal sector | 22.4 | 9.0 | -0.17 | ||
| wage gap, percent |
As measured by the correlations between the cyclical components of the HP-filter.
This refers to the whole INEGI sample starting in 2002.
Appendix 2. Results from VAR Estimations
1. For the output that follows, the source is INEGI, IFS and staff calculations. Quarterly data for 1987:Q1-2009:Q3. Original series were transformed with logs and first differences to make data stationary before estimation. Lag length in the VARs were determined by the Akaike criterion.
Response of the informal sector employment share to a positive one-standard deviation shock to GDP
Citation: IMF Staff Country Reports 2010, 070; 10.5089/9781451981834.002.A002
Response of the informal sector employment share to a positive one-standard deviation shock to GDP
Citation: IMF Staff Country Reports 2010, 070; 10.5089/9781451981834.002.A002
Response of the informal sector employment share to a positive one-standard deviation shock to GDP
Citation: IMF Staff Country Reports 2010, 070; 10.5089/9781451981834.002.A002
Response of informal sector wages to a positive one-standard deviation shock to unemployment
Citation: IMF Staff Country Reports 2010, 070; 10.5089/9781451981834.002.A002
Response of informal sector wages to a positive one-standard deviation shock to unemployment
Citation: IMF Staff Country Reports 2010, 070; 10.5089/9781451981834.002.A002
Response of informal sector wages to a positive one-standard deviation shock to unemployment
Citation: IMF Staff Country Reports 2010, 070; 10.5089/9781451981834.002.A002
Response of formal sector wages to a positive onestandard deviation shock to unemployment
Citation: IMF Staff Country Reports 2010, 070; 10.5089/9781451981834.002.A002
Response of formal sector wages to a positive onestandard deviation shock to unemployment
Citation: IMF Staff Country Reports 2010, 070; 10.5089/9781451981834.002.A002
Response of formal sector wages to a positive onestandard deviation shock to unemployment
Citation: IMF Staff Country Reports 2010, 070; 10.5089/9781451981834.002.A002
References
Alcaraz, C., 2009, “Informal and Formal Labour Flexibility in Mexico”, Desarollo y Sociedad, 2009:1, pp. 115–143
Alcaraz, C., Chiquiar, D., Ramos-Francia, M., 2008, “Diferenciales Salariales Intersectoriales y el Cambio en la Composición del Empleo de la Economía Mexicana” Banco de México WP 2008-20
Chiquiar, D., Ramos-Francia, M., 2009, “Competitiveness and Growth of the Mexican Economy”, Banco de México WP 2009-11
Duval-Hernandez, R., Orroca Hermano, P., 2009, “A Cohort Analysis of Labor Participation in Mexico, 1987-2009”, IZA WP No. 4371
Fairris, D., Popli, G., Zepeda, E., 2006, “Minimum wages and wages structure in Mexico”, Review of Social Economy 66 (2), pp. 181–208
Levy, S., 2008, Good Intentions, Bad Outcomes. Social Policy, Informality, and Economic Growth in Mexico. Washington, D.C. Brookings Institution Press
OECD Employment Outlook, 2008, Organisation for Economic Co-Operation and Development. Paris.
Perry, G.E., Maloney, W.F., Arias, O., Fajnzylber, P., Mason, A.D., Saavedra-Chanduvi, J., 2007, “Informality: Exit and Exclusion”. World Bank Latin American and Caribbean Studies. Washington D.C.
Prepared by Kristin Magnusson.
Self-employed are not required to register with one of the above institutions. Following Alcaraz (2009), we hence classify self-employed as informal if they are not registered with the local or federal government to make sure that our measure of informality does not include self-employed who comply with existing regulations. The measurement of informality, including in public administration, likely reflects the employment of some part-time workers.
The main difference when rural areas are included are discrete upward jumps in the size of the informal sector and the wage gap due to the high prevalence of informality in the agricultural sector, which dominates employment in rural municipalities and pays low wages.
For further information about these correlations and some summary statistics, see Appendix 1.
During the 1994-1995 crisis the unemployment rate increased by about 6 percentage points as opposed to by 2 percentage points so far in the current crisis. Controlling for the extensive margin of employment, by assuming that all individual s who joined the labor force then would have remained outside the labor market and those who left this time around would have stayed and been registered as unemployed, would reduce this 4 percentage point difference in outcomes by about half a percentage point.
For VAR results and details on the estimation, see Appendix 2.
To arrive at this share we repeated the exercise in Fairris et al (2006) who studies wage setting in Mexico during the period 1984–1992, for the last five years of our data set. We look at the distribution of daily wages and calculate the proportion that earns a multiple with increment 0.25 of the national minimum wage +/- 5 percent.
For VAR results and details on the estimation, see Appendix 2.
Unemployment peaked at 6.1 percent in June 2009 and formal employment continued to increase until December when more than 185,000 jobs were lost. This drop was however largely due to seasonal effects as employers traditionally adjust their payrolls and reduce the work force in the last month of the year. Indeed, December 2009 saw the smallest drop for that month and variable over the last ten years.
Using productivity data from the manufacturing sector.
To arrive at the average wage gap for the period, we calculate the ratio between the estimate in Alcaraz, Chiquiar and Ramos-Francia (2008) who control for observable worker characteristics for the years 2001-2004, and the crude wage gap we find for the same years, We then extrapolate this relation for the whole period 1987—2009 to arrive at the 15 percent wage gap.
Since INEGI started surveying both urban and rural areas in 2005, the informal sector employment share has been quite stable at around 60 percent. It is arguably a very strong assumption that all informal workers will become as efficient as formal ones if they change status, but a back-of-the-envelope calculation of the upper bound of foregone output from informality is given by the product of the average wage gap and the share of informal employment.
