Abstract
This Selected Issues paper provides background information and analysis on recent developments and critical issues for the Colombian economy. The study discusses the unemployment and stresses in the financial system and also focuses on fiscal issues. The following statistical data are presented in detail: national accounts at current prices and at constant prices, savings and investment, value of agricultural crops, mining production, structure of regular gasoline prices, indicators of construction activity, minimum wages, producer price index, interest rates, and so on.
II. Unemployment in Colombia: The 1990s 16
A. Background
17. While strong economic growth characterized Colombia in the first half of the 1990s (real GDP growth rates exceeded 5 percent between 1993 and 1995), the late 1990s witnessed a dramatic slowdown in activity. In 1999, the economy experienced the worst recession since the 1930s. On the external side, the country showed significant increases in the current account deficit and the peso appreciated in real terms before a sharp adjustment occurred in 1999/2000. One of the most salient characteristics of the Colombian economy in the 1990s, however, was the behavior of the labor market. The unemployment rate dropped to 7.5 percent in 1994, the lowest on record; in September 2000, it was above 20 percent, the highest ever registered.17
18. Although the economic downturn in the late 1990s helps explain the behavior of the unemployment rate, the drop in activity does not offer a complete explanation. First, the economic boom in the early 1990s hardly affected unemployment. While the economy expanded by almost 30 percent during the period 1990-94, unemployment dropped less than 1.5 percentage points. Moreover, with GDP growing 5.2 percent in 1995, unemployment increased by almost 1.6 percentage points.18 Second, by the time the 1998-99 recession began, unemployment had already increased from 9.8 percent in June 1994 to 15.9 percent in June 1998. Third, unemployment did not fall during the recovery in 2000. In fact, it increased from 18 percent at the end of 1999 to almost 20 percent at the end of 2000. This data suggest that structural problems also might help explain the behavior of unemployment. The objective of this paper is to offer a brief analysis of the labor market in Colombia during the 1990s.
B. Labor Markets Developments During the 1990s
19. The fluctuations in the unemployment rate in Colombia during the 1990s contrasts with the steady upward trend in real wages and, more generally, in real labor costs during the same period (Figure 1(a)). Table 1 presents some indicators of the evolution of these variables, namely, the real wage, an index of unit labor costs, and an index of labor costs. The index of labor costs captures the cost of wages and payroll taxes. The index of unit labor costs adjusts labor costs by the evolution of labor productivity.
Colombia Unemployment, Labor Costs and Growth
Citation: IMF Staff Country Reports 2001, 068; 10.5089/9781451808834.002.A002
Colombia Unemployment, Labor Costs and Growth
Citation: IMF Staff Country Reports 2001, 068; 10.5089/9781451808834.002.A002
Colombia Unemployment, Labor Costs and Growth
Citation: IMF Staff Country Reports 2001, 068; 10.5089/9781451808834.002.A002
Colombia: Labor Market Developments
(In percent)
September figures.
Deflated by the consumer price index. The figure for 2000 corresponds to the growth rate between March 1999 and March 2000.
Unit labor cost captures the cost of producing one unit of product; it is adjusted by labor productivity. The index is deflated by the evolution of the nominal exchange rate.
The index of labor costs captures the cost of wages, social services and pay-roll taxes of permanent and temporary workers. It is deflated by the evolution of the nominal exchange rate.
Deflated by the producer price index.
For 2000, date corresponds to the first semester.
Colombia: Labor Market Developments
(In percent)
| 1990/1995 | 1996/1998 | 1999 | 2000 | |
|---|---|---|---|---|
| Rates 1/ | ||||
| Participation | 59.1 | 60.0 | 63.3 | 64.8 |
| Employment | 53.9 | 52.1 | 50.6 | 51.5 |
| Unemployment | 8.9 | 13.0 | 20.1 | 20.5 |
| Annual Rates of Growth | ||||
| Population | 1.8 | 1.8 | 1.8 | 1.8 |
| Labor supply | 0.8 | 0.9 | 3.8 | 1.6 |
| Employment | 1.1 | -1.5 | -2.3 | 1.8 |
| Unemployment | -2.7 | 20.9 | 34.0 | 2.0 |
| Real wages 2/ | 2.0 | 5.6 | 1.2 | -3.2 |
| Unit labor costs - ER 3/6/ | 9.4 | 1.1 | -2.8 | -12.9 |
| Index of labor costs - ER 4/6/ | 13.1 | 5.4 | -5.3 | -7.5 |
| Index of labor costs - PPI 5/6/ | 6.3 | 5.1 | 2.9 | -3.8 |
September figures.
Deflated by the consumer price index. The figure for 2000 corresponds to the growth rate between March 1999 and March 2000.
Unit labor cost captures the cost of producing one unit of product; it is adjusted by labor productivity. The index is deflated by the evolution of the nominal exchange rate.
The index of labor costs captures the cost of wages, social services and pay-roll taxes of permanent and temporary workers. It is deflated by the evolution of the nominal exchange rate.
Deflated by the producer price index.
For 2000, date corresponds to the first semester.
Colombia: Labor Market Developments
(In percent)
| 1990/1995 | 1996/1998 | 1999 | 2000 | |
|---|---|---|---|---|
| Rates 1/ | ||||
| Participation | 59.1 | 60.0 | 63.3 | 64.8 |
| Employment | 53.9 | 52.1 | 50.6 | 51.5 |
| Unemployment | 8.9 | 13.0 | 20.1 | 20.5 |
| Annual Rates of Growth | ||||
| Population | 1.8 | 1.8 | 1.8 | 1.8 |
| Labor supply | 0.8 | 0.9 | 3.8 | 1.6 |
| Employment | 1.1 | -1.5 | -2.3 | 1.8 |
| Unemployment | -2.7 | 20.9 | 34.0 | 2.0 |
| Real wages 2/ | 2.0 | 5.6 | 1.2 | -3.2 |
| Unit labor costs - ER 3/6/ | 9.4 | 1.1 | -2.8 | -12.9 |
| Index of labor costs - ER 4/6/ | 13.1 | 5.4 | -5.3 | -7.5 |
| Index of labor costs - PPI 5/6/ | 6.3 | 5.1 | 2.9 | -3.8 |
September figures.
Deflated by the consumer price index. The figure for 2000 corresponds to the growth rate between March 1999 and March 2000.
Unit labor cost captures the cost of producing one unit of product; it is adjusted by labor productivity. The index is deflated by the evolution of the nominal exchange rate.
The index of labor costs captures the cost of wages, social services and pay-roll taxes of permanent and temporary workers. It is deflated by the evolution of the nominal exchange rate.
Deflated by the producer price index.
For 2000, date corresponds to the first semester.
20. As Table 1 shows, the real wage has grown during most of the decade. As long as one looks at real wages, labor costs do not react to the downturn in economic activity until the year 2000. Adjusting wages by the evolution of labor productivity or deflating them by the exchange rate does not seem to offer a notably different picture. In spite of the correction in 1999/2000, labor costs have increased during the decade both in terms of foreign and domestic currency. Part of reason for this may be the lack of credibility of the disinflation policy during most of the decade, which tended to keep the economy indexed to past inflation.
21. Table 1 also allows a first assessment of labor market dynamics. The evolution of employment suggests that the economy had reduced job creation by the second half of the 1990s. The rate of employment in 2000 fell by 2.4 percentage points in relation to the average rate during the period 1990-95. The most significant push to unemployment, though, came from the supply side. The total participation rate (the ratio between people who actively look for employment to the working age population) in 2000 was almost 6 percentage points higher than in 1990-95. This is the main reason why the economic recovery in 2000, which actually raised employment,19 did not lower the unemployment rate.
22. While the figures discussed above suggest some general trends in the labor market, the data in Table 2 provide more detailed characteristics of unemployment in Colombia during the 1990s. Overall, unemployment was concentrated among young people (between 15 and 30 years old) with low levels of education. In other words, it seems that there is a mismatch between the skills demanded and those offered by the working age population. This mismatch and the fact that unemployment is higher for the youngest might suggest that more time and resources should be allocated to education and training.
Colombia: Unemployment Classification
(In percent)
Average figures.
As a percentage of total unemployment.
March figure.
Colombia: Unemployment Classification
(In percent)
| Annual Average | ||||
|---|---|---|---|---|
| 1990/1995 | 1996/1998 | 1999 | 2000 3/ | |
| Rates 1/ | ||||
| By family status | ||||
| Head of family | 3.4 | 5.8 | 9.8 | 10.1 |
| Spouse | 11.3 | 14.1 | 20.6 | 21.4 |
| Rest of family | 12.1 | 15.7 | 22.9 | 24.3 |
| By education | ||||
| No education | 7.4 | 10.3 | 17.0 | 14.9 |
| Primary education | 7.7 | 10.9 | 16.6 | 17.0 |
| High school | 10.7 | 14.9 | 21.9 | 24.6 |
| Superior education | 4.4 | 6.2 | 10.5 | 12.0 |
| By age | ||||
| 15-19 | 24.8 | 33.9 | 44.1 | 47.0 |
| 20-24 | 16.6 | 22.0 | 31.9 | 33.4 |
| 25-29 | 10.0 | 13.8 | 20.4 | 21.5 |
| 30-34 | 7.1 | 10.2 | 16.2 | 15.8 |
| 35-39 | 5.6 | 8.5 | 13.76 | 15.2 |
| 40-44 | 4.67 | 7.21 | 12.65 | 11.48 |
| 45-49 | 4.1 | 7.1 | 11.53 | 14.6 |
| More than 49 | 3.8 | 6.0 | 10.12 | 12.16 |
| By searching period 2/ | ||||
| Up to 4 weeks | 29.8 | 22.7 | 15.5 | 12.9 |
| Between 5 and 8 weeks | 11.9 | 11.4 | 9.5 | 8.0 |
| Between 9 and 24 weeks | 15.6 | 16.5 | 16.4 | 13.4 |
| More than 24 weeks | 42.5 | 45.7 | 58.6 | 65.7 |
Average figures.
As a percentage of total unemployment.
March figure.
Colombia: Unemployment Classification
(In percent)
| Annual Average | ||||
|---|---|---|---|---|
| 1990/1995 | 1996/1998 | 1999 | 2000 3/ | |
| Rates 1/ | ||||
| By family status | ||||
| Head of family | 3.4 | 5.8 | 9.8 | 10.1 |
| Spouse | 11.3 | 14.1 | 20.6 | 21.4 |
| Rest of family | 12.1 | 15.7 | 22.9 | 24.3 |
| By education | ||||
| No education | 7.4 | 10.3 | 17.0 | 14.9 |
| Primary education | 7.7 | 10.9 | 16.6 | 17.0 |
| High school | 10.7 | 14.9 | 21.9 | 24.6 |
| Superior education | 4.4 | 6.2 | 10.5 | 12.0 |
| By age | ||||
| 15-19 | 24.8 | 33.9 | 44.1 | 47.0 |
| 20-24 | 16.6 | 22.0 | 31.9 | 33.4 |
| 25-29 | 10.0 | 13.8 | 20.4 | 21.5 |
| 30-34 | 7.1 | 10.2 | 16.2 | 15.8 |
| 35-39 | 5.6 | 8.5 | 13.76 | 15.2 |
| 40-44 | 4.67 | 7.21 | 12.65 | 11.48 |
| 45-49 | 4.1 | 7.1 | 11.53 | 14.6 |
| More than 49 | 3.8 | 6.0 | 10.12 | 12.16 |
| By searching period 2/ | ||||
| Up to 4 weeks | 29.8 | 22.7 | 15.5 | 12.9 |
| Between 5 and 8 weeks | 11.9 | 11.4 | 9.5 | 8.0 |
| Between 9 and 24 weeks | 15.6 | 16.5 | 16.4 | 13.4 |
| More than 24 weeks | 42.5 | 45.7 | 58.6 | 65.7 |
Average figures.
As a percentage of total unemployment.
March figure.
23. Table 2 also shows that the unemployment rate of people with higher education increased the most in relative terms during the 1990s, from 4.4 percent during 1990-95 to 12 percent in 2000. In fact, the ratio between the unemployment rate of highly educated people and those with no education increased by the end of the decade, particularly in 2000, and it remains to be seen whether this is a new trend or a specific phenomenon of the recent recession. In any event, although unemployment is still concentrated among less educated people, it is increasing among highly educated people.
24. The classification of unemployment by the length of time the unemployed have been searching for a job shows that long-term unemployment increased during the 1990s. The proportion of people who have been unemployed for up to four weeks declined over the period, while people searching for work for more than two years increased to represent more than 65 percent of the unemployed in 2000. While the first category can be identified with the existence of frictions, the second could be related to structural reasons. One possible explanation for the latter might be the difficulties of large numbers of people to provide the skills required as the economy experienced significant structural changes in the 1990s.20
C. Labor Supply and Demand
25. The third remarkable development during the 1990s is the increase in the unemployment rate in the category denominated “head of family” (HoF), i.e., persons who contribute the largest share of family income and should be in their most productive years. This means that in order to replace the income loss of a HoF, families may need the contribution of more than one of their other members. Unemployment within the HoF group increased, proportionally, more than any other group during the 1990s (Table 2, classification by family status). Figure 1(b) shows that while the unemployment rate among HoF almost tripled during the 1990s, the participation rate of other family members (OFM) increased from 54.1 percent to 64.1 percent between 1990 and 2000.21 On the assumption that families seek to maintain their income, the employment loss of the main income earner tends to push the unemployment rate upward by raising the participation of other family members in the labor market.
Colombia: Unemployment of Head of Family and Participation Rate of Other family Members
Citation: IMF Staff Country Reports 2001, 068; 10.5089/9781451808834.002.A002
Colombia: Unemployment of Head of Family and Participation Rate of Other family Members
Citation: IMF Staff Country Reports 2001, 068; 10.5089/9781451808834.002.A002
Colombia: Unemployment of Head of Family and Participation Rate of Other family Members
Citation: IMF Staff Country Reports 2001, 068; 10.5089/9781451808834.002.A002
26. Figure 1(c) presents a different rationale for the increase in labor supply during the 1990s. It shows the relationship between real wages and the participation rates of OFM, and might suggest that a reason behind the increase in labor supply is the better income prospects. There are two reasons why this view does not seem plausible. First, the figure shows that the increase in the participation rate accelerates at a time when the performance of real wages declines in 1999-2000. Second, while the aggregate participation rate increased during the 1990s, the participation rate of the HoF fell more than 3 percentage points. A priori, there is no reason to believe that people characterized as HoF perceived the increase in real wages differently than other people.22
Colombia: Real Wage and Participation Rate of Other Family Members
Citation: IMF Staff Country Reports 2001, 068; 10.5089/9781451808834.002.A002
Colombia: Real Wage and Participation Rate of Other Family Members
Citation: IMF Staff Country Reports 2001, 068; 10.5089/9781451808834.002.A002
Colombia: Real Wage and Participation Rate of Other Family Members
Citation: IMF Staff Country Reports 2001, 068; 10.5089/9781451808834.002.A002
27. Figure 2 shows the evolution of the unemployment rate and the price of labor in terms of capital during the 1990s. The evidence shows that the relative price of labor has more than doubled since the beginning of the 1990s.23 This is consistent with the investment boom that took place during the first five years of the decade, which made output more capital intensive. The factor substitution that this entailed softened labor demand in the 1990s. The strong economic expansion that characterized the early 1990s may have masked the disproportionately low growth in the demand for labor, but as economic activity slowed down after 1995, the unemployment rate rose reflecting the changing relative prices.
Colombia: Unemployment and the Labor/Capital Price Ratio
Citation: IMF Staff Country Reports 2001, 068; 10.5089/9781451808834.002.A002
Sources: Dirección Nacional de Programación; and Dirección Nacional de Planeación.Colombia: Unemployment and the Labor/Capital Price Ratio
Citation: IMF Staff Country Reports 2001, 068; 10.5089/9781451808834.002.A002
Sources: Dirección Nacional de Programación; and Dirección Nacional de Planeación.Colombia: Unemployment and the Labor/Capital Price Ratio
Citation: IMF Staff Country Reports 2001, 068; 10.5089/9781451808834.002.A002
Sources: Dirección Nacional de Programación; and Dirección Nacional de Planeación.D. Findings and Policy Issues
28. During the 1990s, unemployment in Colombia moved from a historical low in 1994 to the highest rate on record in 2000. Although the business cycle partially explains this dynamic, the case of Colombia suggests that it also may be related to other causes. One of these is the evolution of the relative price of production factors. The evidence shows that the price of labor services more than doubled in relation to capital over the decade. This relative price change provided incentives to reduce the use of labor relative to capital in an economy that lost much of its growth momentum from the middle of the 1990s and went into a recession toward the end of the decade. A second cause is related to the increase in the unemployment rate of the highest income earners in a family. This tends to force into the labor market one or more other family members in an effort to replace the income loss, and may help explain the extraordinary increase in labor supply that the Colombian economy experienced during the last few years of the decade. A third issue is the mismatch between skills demanded and those offered by the working age population. Unemployment was concentrated among young people with low education levels, suggesting that some people should extend their training period before entering the labor market. Overall, the evidence suggests that a good part of the increase in unemployment may be related to structural reasons. Should that be the case, the recovery from the recent recession may not be sufficient to resolve the unemployment problem.24
29. The authorities have designed three employment programs to help deal with the slack in the labor market: (i) a program to create employment for lower income groups; (ii) a training program for young unemployed (between 18 and 25 years); and (iii) a program of subsidies for families with school-aged children. The first program seeks to increase labor demand within the lower income groups. The other two programs seek to accomplish two goals: first, they aim to reduce participation rates in order to cut down short-term unemployment, and second, they are intended to reduce the mismatch of skills previously mentioned by extending the formal training period. As noted before, one of the key issues in the last decade was the increase in the unemployment rate of individuals classified as HoF. If this is behind much of the increase in the participation rate of other family members as suggested above, a reduction in unemployment of HoFs could reduce the pressure brought about by the increase in labor supply and allow other family members to resume their schooling process. This suggests that it would be important to understand what drives the increase in unemployment of HoFs and put in place mechanisms to bring that group back to work.
Prepared by Esteban Vesperoni.
This high rate of unemployment should be seen in the context of the general lack of unemployment benefits.
It could be argued that unemployment was already at its natural rate in the early 1990s, so that it could not drop with the economic boom. In any case, the behavior in 1995 cannot be explained by this argument.
The employment rate increased by almost 1 percentage point from 1999 to 2000.
The lack of skills mentioned here need not be just related to the unemployment of young people that we mentioned above, but also to middle age people who have previously been employed.
The participation rate of OFM is an average of the participation rates of the following categories: spouse, single son/daughter, and married son/daughter living in the same household.
Instead, the behavior of the participation rate of HoF may be related to people who get “discouraged” and stop looking for a job. This is consistent with the evidence suggesting that long-term unemployment has increased during the decade. It also suggests that, should not these people abandon active job searching, the unemployment rate for HoF would have been even higher.
This is consistent with the experience of other Latin American countries that have liberalized trade and capital movements and experienced a real exchange rate appreciation during the decade.
For instance, if there is a connection between the increase in unemployment of HoFs and the increase in long-term unemployment, and the HoFs are mainly middle-aged population, changes in production practices (resulting from the new relative price between labor and capital) may be affecting their ability to adapt to new employment conditions. In a case like this, a recovery would not change the situation in the labor market materially.
