A review of international experience

number of beneficiaries
(thousands)
subsidy to hiring  subsidy to contribution 
average subsidy to hiring 
average subsidy to contribution
subsidy ($) 
0
5
10
15
20
25
SOURCE: Authors, based on data from Superintendencia de Pensiones.

9.  COMPLEMENTING CHILE’S PENSIONS WITH SUBSIDIZED YOUTH EMPLOYMENT AND CONTRIBUTIONS 
187
young workers. The SEJ subsidy is similar to the SPTJ and has similar objectives, but its 
introduction was motivated more by high unemployment among the target group than 
by concerns about the pension system. The SEJ, launched in July 2009, was aimed at pro-
moting formal employment of young low-income workers.
To qualify for the subsidy, a dependent worker must fulfill three requirements: be 
age 18–25, belong to a household that is among the 40 percent poorest in Chile, and earn 
gross monthly wages of less than $740 and total yearly gross wages of less than $8,880. To 
qualify for the subsidy, the employer has to pay social security contributions on time for 
all its workers, not just the ones for which it is receiving the subsidy. Self-employed work-
ers have access to the subsidy under the same conditions. They must provide evidence that 
they have paid their social security and health insurance contributions on time.
Two-thirds of the subsidy goes to the worker and one-third to the employer. The 
amounts start at 30 percent of the wage on which social security and health contributions 
have been paid for wages below $328 (figure 9.8). The subsidy is flat, at $98.40 for wages 
up to $410. For higher wages, the subsidy falls to zero for monthly wages of $740.
FIGURE 9.8  Monthly SEJ subsidy in Chile, by level of income
income: $328;
subsidy: $98
$738 
0  
20  
40  
60  
80  
100  
120  
0  100 200 300 400 500 600 700 800 
monthly income ($) 
subsidy ($) 
income: $410;
subsidy: $98 
SOURCE: Authors, based on legislation.
Workers with more than one employer receive only one subsidy, the amount of 
which is calculated based on total monthly gross wages. For workers, the monthly pay-
ment of the subsidy is an advance of 75 percent of the subsidy. Once a year, the subsidy is 
recalculated based on gross annual wages. Workers can receive the subsidy after age 25 for 
each year they were enrolled in higher education and for time on maternity leave after age 
18. In 2010, on average 67,048 young workers received the subsidy each month, and the 
average subsidy was $33 (figure 9.9). 
Public spending on this subsidy has been growing. In 2011, it reached about 
$88 million (table 9.2).

188 
MATCHING CONTRIBUTIONS FOR PENSIONS: A REVIEW OF INTERNATIONAL EXPERIENCE
Given the importance of the SEJ subsidy in terms of both fiscal spending and public 
policy, the Budget Office asked for an impact evaluation in 2011 (see Centro de Micro-
datos 2011). Because of the short period of time that has passed since the subsidy scheme 
was launched, the terms of reference for the evaluation did not include an assessment of 
the impact on future pensions or potential savings from social pensions. The focus was on 
the labor impact of the subsidy. The evaluation used administrative data from unemploy-
ment insurance, the Ficha de Protección Social (Social Protection Index—FPS) targeting 
mechanism, and the subsidy itself. Using the regression discontinuity method, beneficia-
ries at the margin of the age limit were compared with people over the age limit; people at 
the margin of the cut-off score of the FPS were compared with people who did not qualify.
The evaluation shows that coverage of the SEJ subsidy was low, at about 4.6 per-
cent of the eligible population in 2009 and 4.8 percent in 2010. The share of employed 
people in the eligible population that received the subsidy was 21.2 percent in 2009 and 
21.3 percent in 2010. The subsidy benefited mostly young workers with more stable jobs, 
raising questions about how many of these workers would have had formal employment 
in any case and about the deadweight loss of the subsidy.
Use of the subsidy by enterprises was very low (3.75 percent): subsidy of the demand 
for labor had less coverage than subsidy of the supply of labor. Enterprises asked for the 
subsidy for only 10 percent of the total number of workers who received it in 2009 and 
for just 11 percent in 2010. Among companies that used the subsidy, 44 percent were 
microenterprises with fewer than 10 workers.
The results of the evaluation using the age factor show no significant impact of the 
SEJ, possibly because the SPTJ covers workers up to age 35. The FPS, which relies heavily 
on self-reporting, is a weak targeting mechanism, because the bottom two quintiles end 
up including almost 60 percent of Chile’s population. This weakness also puts into ques-
tion the results of the impact evaluation.
FIGURE 9.9  Number and average amount of SEJ subsidy paid in Chile 
number of beneficiaries (thousands) 
90
80
70
60
50
40
30
20
10
0
Sept.
2009
Nov.
2009
Jan.
2010
March
2010
May
2011
July
2011
Sept.
2011
Nov.
2011
Sept.
2010
Nov.
2010
Jan.
2011
March
2011
May
2010
July
2010
number 
average subsidy 
subsidy ($) 
90
80
70
60
50
40
30
20
10
0
SOURCE: National Employment and Training Service.

9.  COMPLEMENTING CHILE’S PENSIONS WITH SUBSIDIZED YOUTH EMPLOYMENT AND CONTRIBUTIONS 
189
Bearing this caveat in mind, the impact evaluation of the SEJ shows a positive impact 
on the supply of labor: employment and labor force participation rates among the eligible 
population rose, particularly in 2009. The evaluation estimated an average increase in the 
probability of employment of 4.5 percentage points in 2009 and 2.1 percentage points 
in 2010. The peak in terms of increased probability of employment was in August and 
September 2009, suggesting that the initial advertising of the subsidy and identification of 
beneficiaries had an immediate impact that quickly waned.
There seems to have been a significant impact on the rate of formal employment, 
which was one of the objectives of the subsidy, measured as the proportion of time the 
worker was contributing while working. The number of months contributed increased by 
6.6 percentage points among people receiving the subsidy. In contrast, the increase among 
those not receiving the subsidy was only 4.8 percent. The labor force participation rate 
rose by an estimated 3.0 percentage points in 2009 and 2.7 percentage points in 2010. 
The evaluation concluded that, in times of high unemployment, the SEJ could be a 
useful tool for promoting employment of low-income workers but that the reduction in 
the impact in 2010 leads to questions about the need to retain the subsidy in times of high 
employment creation and falling unemployment. An unanswered question is whether the 
larger impact of the SEJ in 2009 was caused by the instrument itself or a strong public 
effort to place young workers using the subsidy during a year marked by economic crisis 
and presidential and congressional elections. The evaluation concludes that the SEJ is not 
a suitable instrument for addressing social security and pension issues, which are longer 
term in nature.
COMPARATIVE ANALYSIS AND INTERACTION BETWEEN SUBSIDY SCHEMES
Table 9.3 summarizes the main characteristics of both youth employment subsidy 
schemes. The subsidies share the objective of promoting formal youth employment and 
in particular on-time payment of social security contributions; the expectation is that they 
could lead to a culture of contribution that will eventually result in better pensions. How-
ever, because of the long lag in results, the subsidies will not replace the minimum pension 
guarantees put in place in 2008.
As employers can pick only one of the subsidies, if they use a subsidy at all they will 
probably choose the SEJ, which provides larger subsidies and results in a larger reduction 
of the effective wage paid by the employer. Workers can access both subsidy schemes at the 
same time. There are no data on the incidence of either subsidy. The net impact on future 
pensions is thus not clear. Young low-income workers who are not in the bottom two 
income quintiles, workers age 25–35 who have not made 24 contributions, and work-
ers who are not employed under the labor code (for example, public sector workers) can 
access only the SPTJ.
Conclusion
Chile’s experience with youth employment subsidies is interesting because Chile is an 
emerging economy and its programs target the poorest 40 percent of the population. 
The results of the impact evaluation are in line with international experience. The impact 
on formal employment and labor market participation was positive, but the design of 

190 
MATCHING CONTRIBUTIONS FOR PENSIONS: A REVIEW OF INTERNATIONAL EXPERIENCE
the subsidy and its targeting (using the FPS, which relies on self-reporting) could have 
been better. The initial design of the employment subsidies considered using a targeting 
mechanism specific to the employment subsidies. In the rush to launch the employment 
subsidies before the elections of November 2009, however, the decision was made to start 
with the FPS. A new targeting mechanism has not yet been developed; in light of the 
results, it may not be developed until a complete revision of both employment subsidies 
is undertaken.
Another issue to be considered is the overlap in the objectives of the two subsidies. 
The SPTJ subsidy is much more explicit about promoting pension fund contributions, 
but the higher benefits of the SEJ and the similarity of the benefits requires an evaluation 
of the continuity of the SPTJ and the possibility of double-dipping by a significant share 
of beneficiaries. Another issue that needs to be addressed is the low utilization of the sub-
sidy by enterprises.
There are indications that the subsidies may have been used as an electoral cam-
paign mechanism. The number of allocated subsidies peaked just before the elections, 
after which the effort to promote the mechanism declined. The design of an improved 
scheme should take this factor into account.
The evaluation of the SEJ and international experience show that most of the impact 
of employment subsidies is short term. This evidence raises questions about the suitability 
of employment subsidies to promote social security among youth. Although the subsidies 
TABLE 9.3  Comparison of Chile’s youth employment subsidy programs
Feature
SPTJ
SEJ
Start date
Subsidy to employer: October 2008; 
subsidy to worker: July 2011
Subsidy to employer and worker: July 
2009
Age 
requirement
18–35 
18–25; extension for maternity and 
education (expires at age 21 if secondary 
education incomplete)
Targeting
None
Bottom two quintiles according to FPS 
vulnerability measure
Wage 
requirement
< $540 (1.5 × monthly minimum wages) 
< $740 a month, measured as average for 
the year
Length of 
benefi t
First 24 contribution payments, 
continuous or discontinuous
No limit as long as above requirements 
are fulfi lled
Benefi t
5% of the minimum wage (50% of 
pension contribution), about $19 paid 
monthly to employer and 5% of minimum 
wage into worker’s pension fund account
30% of current wage for wages up to 
$328, $99 for wages of $328–$410, $99 
less 30% of the difference between actual 
wage and $410 for wages $410–$740; 20 
percentage points of the subsidy go to the 
worker and 10 percentage points to the 
employer
Type of 
worker
Dependent
Dependent workers under labor code; 
independent workers who pay income tax 
and are current with their contributions for 
pension and health insurance

9.  COMPLEMENTING CHILE’S PENSIONS WITH SUBSIDIZED YOUTH EMPLOYMENT AND CONTRIBUTIONS 
191
may have some impact on employment and labor force participation, they are unlikely 
to provide a meaningful alternative to the solidarity pillar in addressing the problem of 
low pensions. The subsidies seem more suitable as countercyclical policy in times of high 
unemployment.
Notes
1.  The 0.77 percent management fee is the result of the last auction among asset managers for 
new workers entering formal employment. The 2008 reform introduced an auction system 
that registers all new workers with the asset manager that bid the lowest management fee for a 
period of two years. The latest bidding process resulted in a management fee of 0.77 percent 
for all new workers during 2012 and 2013. Other workers can join the asset management firm 
at the same management fee.
2.  The plan may establish a minimum period before employees vest the amounts contributed by 
the enterprise.
3.  Globally, the experience with employment subsidies has not been positive. A review by Bucheli 
(2005) of labor market programs in a variety of countries concludes that recipients of employ-
ment subsidies do not have better outcomes than people in control groups, except people 
with disadvantages (the long-term unemployed, women). Puerto (2007) provides evidence 
that the international experience with youth employment subsidies has been more positive in 
emerging economies than in industrial countries, where results have been mixed. Along the 
same line, Marx (2001) concludes from a review of the literature of empirical evidence that 
the net impact of employment subsidies is considerably smaller than most theoretical models 
would predict. The main problem is that most subsidized workers would have obtained jobs 
without the subsidy (deadweight loss). Adequate targeting of the subsidies is essential for posi-
tive results.
References
Bucheli, Marisa. 2005. “Las políticas activas de mercado del trabajo: un panorama internacional de 
experiencias y evaluaciones.” Serie de estudios y perspectivas de CEPAL, Comisión Económica 
para América Latina y el Caribe, Montevideo.
Centro de Microdatos, Departamento de Economía, Universidad de Chile. 2011. “Informe final 
de evaluación de impacto del programa de subsidio al empleo joven.”
Marx, Ive. 2001. “Jobs Subsidies and Cuts in Employers’ Social Security Contributions: The Ver-
dict of Empirical Evaluation Studies.” International Labour Review 140 (1): 69–83.
OECD (Organisation for Economic Co-operation and Development). 2010. Off to a Good Start? 
Jobs for Youth. Paris: OECD.
Puerto, Olga. 2007. “International Experience on Youth Employment Interventions: The Youth 
Employment Inventory.” Background paper for the World Bank’s 2007 Economic and Sector 
Work on Sierra Leone Youth and Employment, World Bank, Washington, DC. 
Superintendencia de Pensiones. 2011. “Panorama Previsional de Septiembre de 2011.” 
http://www.spensiones.cl/portal/informes/581/w3-article-7793.html.

192 
MATCHING CONTRIBUTIONS FOR PENSIONS: A REVIEW OF INTERNATIONAL EXPERIENCE
Annex: Composition of Chile’s Labor Force, 2010 and 2011 
Sex/age
Participation rate (labor 
force/population)
Occupation employment/
population
Unemployment rate 
(unemployed/labor force)
2010
2011
2010
2011
2010
2011
Total
 
 
 
 
 
 
15–19
19.4
20.2
14.9
15.8
23.1
21.7
20–24
57.5
57.4
47.8
48.2
16.9
16.0
25–29
77.0
78.0
68.2
70.0
11.5
10.3
30–34
80.1
81.7
73.9
75.7
7.7
7.3
35–39
79.9
80.5
74.4
76.0
6.9
5.6
40–44
78.9
80.1
74.3
76.0
5.9
5.1
45–49
76.1
78.6
72.1
75.4
5.3
4.1
50–54
73.7
75.4
69.5
72.3
5.7
4.2
55–59 67.7
68.6
64.7
65.9
4.4
3.9
60–64
52.2
54.4
50.1
52.3
4.1
3.8
65–69
34.5
37.1
33.4
35.9
3.2
3.2
70+
13.1
15.2
12.9
14.9
2.1
1.8
15-24
37.5
38.4
30.5
31.7
18.6
17.5
15-64
64.8
66.2
59.3
61.3
8.4
7.4
Men
 
 
 
 
 
 
15–19
23.4
23.8
18.8
19.3
19.6
19.0
20–24
67.1
66.2
56.7
57.1
15.5
13.7
25–29
87.2
87.9
77.9
79.6
10.7
9.5
30–34
93.5
94.8
87.3
89.3
6.6
5.7
35–39
94.4
94.5
89.0
90.4
5.7
4.4
40–44
94.5
93.7
90.1
90.0
4.6
3.9
45–49
93.6
94.2
89.6
91.1
4.2
3.3
50–54
91.8
92.3
87.3
88.9
4.9
3.7
55–59 87.5
87.7
83.5
84.5
4.6
3.6
60–64
75.6
77.2
72.4
74.1
4.2
4.0
65–69
51.3
54.4
49.5
52.5
3.5
3.5
70+
22.6
24.7
22.1
24.2
2.5
2.0
15-24
43.8
44.3
36.6
37.6
16.6
15.2
15-64
77.8
78.6
72.1
73.6
7.4
6.3
Women
 
 
 
 
 
 
15–19
14.9
16.0
10.5
11.8
29.5
26.3
20–24
47.1
48.1
38.2
38.8
19.0
19.4
25–29
66.8
68.0
58.5
60.3
12.5
11.3
30–34
66.0
68.8
59.9
62.3
9.3
9.3
35–39
66.1
67.4
60.5
62.6
8.4
7.2
40–44
64.0
66.6
59.1
62.1
7.6
6.8
45–49
59.8
63.7
55.8
60.3
6.8
5.3
50–54
57.2
59.3
53.2
56.4
6.9
4.9
55–59 48.9
51.2
46.9
48.9
4.2
4.3
60–64
31.4
33.7
30.2
32.5
3.7
3.6
65–69
18.9
21.1
18.5
20.6
2.4
2.5
70+
6.1
7.8
6.0
7.7
0.9
1.4
15–24
30.4
32.0
23.8
25.3
21.7
21.1
15–64
51.8
53.9
46.7
49.1
9.9
8.9
SOURCE: National Institute of Statistics.

193
CHAPTER 10
Matching Contributions in 
Colombia, Mexico, and Peru: 
Experiences and Prospects
Luis Carranza, Ángel Melguizo, and David Tuesta
Although Colombia, Mexico, and Peru introduced major structural reforms in the mid-
1990s, they still exhibit very limited levels of social protection coverage (28 percent of the 
labor force in Colombia, 27 percent in Mexico, and 22 percent in Peru). These figures 
reflect the limited effectiveness of basing a pension scheme on employer withholding of con-
tributions in environments in which most employment relations are informal. Better design 
of matching defined contribution schemes could increase the level of pension savings by the 
nearly 20 million middle-income workers in the informal sector in the three countries. Such 
an approach is preferable to ex post solutions to the growing problem of old age as the region 
ages rapidly, which are not sustainable and strengthen incentives for informality.
C
olombia, Mexico, and Peru implemented structural reforms of their pension systems 
at almost the same time. Between 1994 and 1997, all three countries introduced man-
datory funded systems for retirement based on individual defined contribution accounts. 
In Colombia and Peru, the reform allowed participants to choose between the existing 
defined benefit pay-as-you-go system (with some parametric adjustments) and the new 
defined contribution scheme. In contrast, in Mexico the introduction of privately run sav-
ings involved closing the pay-as-you-go system.
1
These reforms had three explicit objectives: to ease the pressure generated by the 
pension system on public expenditures, to provide pensions that are more directly linked 
to the history of contributions, and to create incentives for greater participation in the 
formal labor market and therefore the pension system. Although the first two objectives 
were to a large extent achieved, the reform fell short in addressing the problem of low 
coverage.
2
 Coverage rates (measured as the number of contributors as a share of the eco-
nomically active population) remain low (28 percent in Colombia, 27 percent in Mexico, 
and 22 percent in Peru in 2010). These levels of participation in the pension system are 
not only low but slightly below those observed before the implementation of the reforms 
in Mexico and Peru.
The authors would like to thank Consuelo del Carmen Hoyo (BBVA Research) and Christian 
Daude and Juan de Laiglesia (Organisation for Economic Co-operation and Development Devel-
opment Centre) for their helpful comments and suggestions, as well as Maria Claudia Llanes and 
Rosario Sanchez (BBVA Research) for their assistance in the empirical analysis for Colombia and 
Peru. The views expressed herein are the sole responsibility of the authors and do not necessarily 
reflect the opinions of their institutions.

194 
MATCHING CONTRIBUTIONS FOR PENSIONS: A REVIEW OF INTERNATIONAL EXPERIENCE
The lack of coverage by large segments of the population represents an enormous 
public policy issue in these countries. It reflects the often very difficult challenges associ-
ated with achieving high coverage rates among low- and middle-income groups and in 
environments characterized by high levels of informal employment, which are common at 
this level of development (figure 10.1).
3
FIGURE 10.1  Pension coverage in Colombia, Mexico, and Peru 
20% 
30% 
25% 
28% 
27% 
22% 
0
20
40
60
80
100