Transcript 1100 Bol d'Or

A Prospective Evaluation of Reforms in
Latin America
David Tuesta
Director of Pensions Studies
FIAP Annual Assembly
Lima, May 28, 2008
Main effects of the reforms
•
The reforms have contributed to fiscal consolidation. Without reforms, the explicit debt in
Latin America would have increased, on average, to levels close to 200% of GDP. In this
regard, the most solid gains arise from having avoided the impact of the future ageing
process on public accounts.
•
The mandatory individually funded pension systems have a positive effect on national
savings levels (between 1 and 5 points of the GDP) and on the growth of the GDP (0.5%).
•
A positive relationship between the equity of the Pension Funds and the savings rate can
be noted, which, in turn, enables lesser dependency on external financing.
•
It has been possible to develop capital markets.
•
Positive but mixed results in terms of coverage and replacement rates.
•
Relevance of the structural aspects of economies in the relative success of the reforms.
What does a pension system seek?
SOCIAL LEGITIMACY
 Coverage
 Sufficiency
 Sustainability
Does the pension system manage to cover
a relevant percentage of the population?
Are the pensions that will be obtained
adequate?
Which are the most
vulnerable population groups?
Are the systems financially sustainable?
ECONOMIC – FINANCIAL LEGITIMACY
Are these objectives being achieved?
A partial diagnosis
Historical evolution of coverage and the funds accumulated in the individually funded
system:
•
Worker coverage
(Members / pop. 14-64)
Pension Funds
(% of the GDP)
80%
70%
70%
60%
60%
Chile
Argentina
50%
50%
40%
Chile
40%
Mexico
30%
20%
10%
0%
Colombia
30%
Colombia
20%
Peru
10%
Peru
Argentina
Mexico
0%
Source: BBVA, FIAP, AIOS
4
A partial diagnosis
Currently, the situation observed in the pension
systems has not been totally satisfactory:
Coverage 2006
Members/Pop. 14-64
90%
CHI
80%
Coverage
70%
Coverage is still low in the majority of
countries. This can be explained by the size
of the informal economy, the degree of
economic development, confidence in the
systems and the labor market.
60%
ARG
CO
40%
30%
PE
20%
10%
0%
Pensions
The replacement rates show that average
pensions will be good for workers who
contribute regularly. Nonetheless, this will
not necessarily be so for those with low
contribution frequencies. Furthermore, this
will be more serious for those with lower
incomes.
MEX
50%
Replacement rates 2006
Pension/salary
68%
Mexico
60%
Argentina
50%
Colombia
49%
Chile
40%
Peru
Source: BBVA
0%
5
10%
20%
30%
40%
50%
60%
70%
80%
A partial diagnosis
Informality and coverage
100%
Structural factors are relevant for
the success of the reforms.
Uru
80%
CR
Pan
40%
Bra
Mex
20%
Hon ES
Nic
0%
15%
Par
Ecu
Per Gu
25%
35%
45%
55%
Bol
65%
Poverty and coverage, 2005
75%
60%
Informality
Per
50%
Col
Per capita income, poverty
levels and informality condition
the importance of current
consumption for an important
percentage of the population and
limit the long term savings
capacity.
Coverage rates
Coverage
Chi
60%
40%
Arg
30%
20%
Mex
Chi
10%
0%
0%
10%
20%
30%
40%
Coverage
50%
60%
70%
80%
A partial diagnosis
Worker coverage in Mexico
(% of total 2006)
Pension systems
5%
64%
Main system
Institutions
Defined
Benefit
(BD)
Municipalities
States
Quasigovernment
IMSSEmployer (RJP)
5%
Withour Social
Security
Coverage*
ISSSTE
26%
Source: BBVA Bancomer with information from INEGI
*/ Estimated on the basis of individuals without access to health agencies
Defined
Contribution
(CD)
IMSSAsegurador
(Afore System)
A partial diagnosis
Worker coverage in Peru *
(2005)
Number
of workers
(millions)
Covered workers
64%
1
Average
Income
(S/.)
4.7
1,371
8.2
605
Income less tha S/. 500 per month
4.0
416
Income between S/.500 and S/.800
3.6
641
Income greater than S/.800 3
0.6
1,591
Workers not covered
2
*Note: Corresponds to workers between 20 and 64
1/ Dependent and independent workers in professional or technical activities.
2/ Employees, employers, workers and domestic workers not included in 1/.
3/ Made up of self-employed workers and employers.
Source: Permanent Employment Survey 2005. Ministry of Labor.
Drawn up by: BBVA
SNP: Workers who have contributed for
less than 20 years (do not access the
minimum pension).
SPP: Workers with incomes below S/.800 and
low contribution frequency.
Daignosis with a view to the future
For the purpose of evaluating the performance of pension systems in the long term (projection horizon to
2050) in BBVA we have designed a Macro-Actuarial model which considers:
Perfil de crecimiento
Ageing profile
 Active population
 Work offer
 Potential retirees
Demography
Macro Economy
 CELADE-INEI
 Evolution of the GDP
 Level of investment
 Productivity
 Demand for formal
work
 PTF Model
Parameters of the
Pension System
Membership, contribution and benefits profile
 Contribution, commission , profitability rate
 Mortality tables, calculation of benefits
between systems
 Actuarial Model
The variables to be evaluated are:
Coverage, Level of pensions and Fiscal cost
Diagnosis with a view to the future
If systems continue operating as they currently do, in the next few years….
Old age coverage
(Retirees/Pop. 64- or above)
Replacement or Substitution Rates
(Pension/ Average salary over the last 10 years)
100%
80%
Mexico
80%
60%
60%
40%
Peru
20%
40%
20%
0%
2006 2010 2015 2020 2025 2030 2035 2040 2045 2050
Peru
Mexico
0%
2006 2010 2015 2020 2025 2030 2035 2040 2045 2050
Source: BBVA
…an improvement will be achieved with respect to coverage and replacement rates, but they will be far
below what is desirable and will imply poor performance for guaranteeing adequate protection during
old age.
10
Diagnosis with a view to the future
In 2050, in the case of Peru, the average SPP pension is four times higher than what it will be in
2010. However, a large pension spread can be observed based on income levels and frequency of
contributions by the workers..
Pensions compared to the minimum pension
(Minimum pension in 2010=100)
Based on Contribution Density
800
700
Minimum Pension
A: 90%-100%
B: 50%-90%
C: 10%-50%
Based on income
800
A
600
500
600
B
400
300
700
Minimum Pension
2: Between S/500 y S/.800
3: Greater than S/.800
3
500
400
C
300
200
200
100
100
0
2
0
2010 2015 2020 2025 2030 2035 2040 2045 2050
2010 2015 2020 2025 2030 2035 2040 2045 2050
Thus, despite the pension growth tendency, mid-income workers and workers with reduced
frequency contributions may not receive adequate pensions.
Source: SBS and BBVA projection. Drawn up by: BBVA.
Diagnosis with a view to the future
In the Mexican case, the generous replacement rates for the generation in transition will have a significant
fiscal cost, whereas in the Peruvian case, the total cost of the system will show a decreasing tendency as a
result of the termination of obligations in the public system..
Fiscal cost due to pension liabilities
Peru (2007-2050)
Mexico (2005-2050)
Current value (% of GDP 2006)
Current value (% of GDP 2004)
Public system
ISSSTE
IMSS
Private system
Transition
Reformed IMSS
Social security cost
24%
20%
4%
59%
56%
3%
90%
Public system
SNP
20530
Private system
Recognition bond
Transition
Social security cost
52%
31%
21%
6%
3%
3%
58%
Source: BBVA
12
Diagnosis with a view to the future
In synthesis…

The coverage rates of the pension system will continue to show an increasing tendency, but without
regulatory changes a significant proportion of the population will still be without coverage in old age.
It is especially urgent to adapt pension structures to the most vulnerable group of workers.
• For example, in Peru in 2050, 6 of every 10 elderly people will lack pension coverage
whereas almost half of working age people will still be outside the pension system.
• In general, attention must be paid to systems that encourage active participation in the pension
system; in other words both greater membership and an increase in the frequency of
contributions must be sought.

The level of pensions must also increase, but it has been noticed that there is an important group of
workers – of low income contribution frequency levels – who will obtain reduced pensions. In some
cases, they will not be able to cover minimum consumption levels or will not receive any
pension whatsoever..
• The replacement rates will continue to be adjusted to more realistic rates, but once again, it can
be seen that such will be too low for the groups of low income and low density contribution
workers.

The fiscal cost as a percentage of the GDP shows a decreasing tendency over time, which contributes
to the financial sustainability of the system
Can more be done?
- Economic-financial sustainability (an important pre-requisite) is being
attained little by little.
-Nonetheless, there are clear indications that this is not enough (coverage
and pension levels), which could limit the social and political viability of
the systems: what happens to individuals who do not have the capability
or the possibility of saving over 35 years?
-Need to generate consensus for progress among the main actors.
- Interaction of the public and private sectors.
- Any development in the systems cannot go against the grain of already
achieved sustainability.
Can more be done?
The following are required for tackling the coverage and pensions issues:
• Identifying the causes of the difficulty of generating income over time
(informality, labor market, economic growth).
• Recognizing the structural limitations; identifying systems that encourage
informal workers to participate in formal systems such as pensions
systems.
 How to generate systems that encourage individuals to contribute to a private
account? The existence of different problems must be acknowledged.
Depending on the socio-economic characteristics, the reactions to the incentives
could be different: income levels, need of the product/service.
 What is needed?
Can more be done?
 A good incentives program requires a good characterization of the
population.
 Start of the BBVA project for South America. A start has been made by analyzing the
Chilean case on the basis of the Chilean Social Protection Survey :
•
There must be a prior description of nonmembers:
- Low and variable income
- Self-employed, formal and informal
- Economic sectors
• Why do they not contribute?
- Because it is not mandatory
- Due to lack of money
•
Would they be prepared to contribute?
- Only 14% of this group replied
affirmatively. Willingness to contribute
is also not high among members..
• Among the benefits mentioned for
encouraging this contribution are health
(40%), education (30%) and housing (28%).
Tax benefits only received a 2% mention.
Marcel Committee Diagnosis, BBVA
Can more be done?
BBVA Study (2007) of self-employed workers in Chile:
 Objectives
• Identify the relevant characteristics of the self-employed workers and the key
factors in their decision to contribute or not to the pension systems.
• Propose concrete measures to encourage contributions on the basis of the
preferences of this type of workers for alternative savings systems.
• Contribute with proposals that serve to reduce the evasion in contributions of this
segment of the population.
 Methodology
• The model used is a binary choice probit model
• The dependent variable will assume a value of 1 if the self-employed worker contributes to
the pension system and 0 if he does not.
• The model is corrected for selection bias using Heckmann’s methodology in two stages, since
the decision to be or not to be a self-employed worker skews the estimations.
• Therefore, there is a secondary equation that estimates the probability of the working being
self-employed.
• Variables used: vector of the characteristics of alternative forms of saving; characteristics of
the individual; home and work characteristics.
Can more be done?
BBVA Study (2007)
Results of the main equation
Explicative variables
Log amount of durable b
Log amount of b capital
Log amount of childrens’ educational costs
Dummy financial access=1
Dummy member health system, free=1
Dummy member health system private = 1
Dummy has no health system =1
Log household per capita income amount
Dummy formal worker = 1
Dummy paid housing = 1
Dummy Sex. man=1
Age in years
Education in years
Dummy permanent work
Dummy Marital status. Married or living together = 1
Pseudo R2
N° Obs
N° Uncensored Obs.
Model 6
dy/dx
P-value
0.000
-0.001
0.002*
0.013
-0.246*
-0.036
-0.275*
-0.005**
0.056*
0.012
-0.014
-0.002*
0.004*
0.040*
-0.018
0.950
0.299
0.049
0.286
0.000
0.128
0.000
0.090
0.001
0.317
0.380
0.000
0.012
0.019
0.225
52%
8,226
1,885
*Significant to 5% reliability. ** Significant to 10% reliability. Notes: (a) dy/dx indicates the change in the probability of contributing as a result of a
marginal change in the explicative variables. (Evaluated on average). (b) Probit model with selection bias correction (Heckman) y heterocedasticity.
Can more be done?
BBVA Study (2007)
Results
1.
2.
3.
4.
5.
Mandatory contribution is not enough to solve the coverage and pension amount issues, since
only 49% of self-employed workers are formal and could be subject to control mechanisms.
Self-employed workers vary due to income level. They have preferences for different types of
alternative savings depending on their level of income.
•
Those with lower incomes prefer to save for housing
•
Those with high incomes prefer greater liquidity
→
An increase of 1% in alternative saving (capital goods or financial saving) reduces the
possibility of contributing.
Incentives: self-employed individuals would be willing to contribute if they obtained health,
housing or education benefits thereby.
Children’s education. In the estimated models, when income brackets are not controlled, the
variable turns out to be negative, which implies that an increase in the children’s educational
expenses increases the probability of contributing in the pension system for mid and high income
sections.
There is a connection between health and pensions. There is a positive correlation between
contributing to public health and pension systems.
Can more be done?
BBVA Study (2007)
6.
Other variables
•
•
•
•
Education: an increase in the number of years of education increases the probability of
contributing in the system between 0.4 and 0.9%.
Permanent work dummy: having an independent but permanent job, increases the
probabilities of contributing in the pension system.
Formal worker: the individual who normally pays taxes to the State also normally pays his
pension contributions, according to specific models.
Age: the probability of contributing increases between 26 and 65 years of age, which matches the
life cycle theory. The coefficients increase, in other words the probability of contributing to
pensions increases as we advance through age sections, reaching a maximum of 20% increased
probability in the 46 to 55 age section. It decreases in the in the last age section of the active
economic life of workers between 56 and 65.
Can more be done?
BBVA Study (2007)
Thus, the need to define a system of incentives to contribute is defined, based
on:
• The design of an attractive product.
• Adequate contribution.
• A well-focused subsidy system for those who do not have sufficient
income or are prematurely expelled from the labor market
Can more be done?
Suggested reforms for the Mexican and Peruvian cases:
Mexico
Coverage
→ State co-financing of voluntary contributions of selfemployed workers.
→ Establish a National Pensions System.
Pensions
→ Gradually increase mandatory contribution from 6.5% to
11.3%
→ Focalize public resources: double the social contribution for
individuals earning up to 3 minimum wages and eliminate
it for higher incomes.
→ Relax the requirements for the minimum pension (19 years
for 50% with scaled increases)
→ Destine the housing balance to retirement prior to
retirement.
Fiscal costs
→ Close defined benefit systems with a deficit
→ Use the recognition bond to limit the cost and reform the
IMSS
Others
→ Strengthen profitability and risk options for the Siefore.
→ Include new types of assets
→ Eliminate the domestic bias in investments and allow new
investment vehicles.
22
Peru
Coverage
→ 2 incentive programs which are fixed contribution systems of
S/. 30 and S/. 50 per month for low income informal and selfemployed workers
→ Mandatory enrollment for high income self-employed workers
Pensions
→ Extend and relax minimum pension requirements for low
income SPP workers.
→ Accelerate the reduction in replacement rates, more rapidly
adjusting the higher SNP pensions to levels in harmony with
contributions. Enable redirecting resources to those who earn
less.
Fiscal costs
→ The fiscal costs decreases as of 2035.
→ Improvements in coverage and solidarity benefits can be
implemented as of that year
→ Use complementary bonds for financing the measures or
solidarity contributions for financing the SPP minimum
pension
Others
→ Put Public Pension Fund Management (FCR) out to tender for
private management
→ Smooth out the reference salary used for calculating public
pensions, increasing it from 60 to 240 months.
Can more be done?
In summary, even though coverage of the systems will continue rising and pensions
also show a positive trend to more realistic figures, we put forward recommendations
that are aimed at:

Implementing systems that encourage the participation of uncovered sections
through attractive financial vehicles and with very focalized State participation
 Adapting
the programs and pension access in accordance with the
characteristics of the labor markets in countries.

Encouraging contribution throughout the life cycle.

Generating realistic pension systems (realistic substitution rates).
 Developing financial instruments that enable greater profitability of the
system

Ensuring that the measures maintain the country’s fiscal stability.
Impact of the measures
Coverage: The reforms will enable achieving significant reductions in the percentage of the
population outside the pension system.
Peru
Mexico
(Pensioners / Pop>64)
Pensioners / Pop>64)
60%
100%
With reform
With reform
80%
40%
60%
Without reform
Without reform
40%
20%
20%
0%
0%
2010 2015 2020 2025 2030 2035 2040 2045 2050
pág
24
2010 2015 2020 2025 2030 2035 2040 2045 2050
The benefits will arise from the implementation of systems that encourage the participation
of uncovered segments through attractive financial vehicles with very focalized State
participation.
Fuente: BBVA
Impact of the measures
Replacement rates : Focusing the measures enables a substantial improvement in
the substitution rates.
Mexico
Peru
80%
80%
60%
60%
40%
With reform
With reform
40%
Without reform
20%
Without reform
0%
20%
0%
201020152020202520302035204020452050
2010 2015 2020 2025 2030 2035 2040 2045 2050
Source: BBVA
Specifically, the improvements in Peru are more significant among the cohorts of
lower income members
25
Impact of the measures
The average pension in the Peruvian pension system after the reforms would be 12% higher than those
obtained in the base scenario. In the case of mid-income workers, it would increase 98% on average
between 2010 and 2050, with respect to the pensions that would be achieved without reforms..
Pensions of mid-income workers
(average pension in 2010= 100)
1,100
1,038
With reform
1,000
Base
865
900
741
800
466
638
700
365
552
285
600
479
263
500
241
388
227
400 284 332
194
300
193
200 184
100
0
2010 2015 2020 2025 2030 2035 2040 2045 2050
Source: ONP, SBS, BBVA projection. Drawn up by: BBVA
On average,
pensions 98%
higher than the
base.
Impact of the measures
Additionally, the reforms are fiscally viable:
Mexico
Perú
(2005-2050)
(2007-2050)
Current value as a percentage of the 2004 GDP
Public system
24%
20%
ISSSTE
4%
IMSS
Private system
59%
Transition
56%
3%
IMSS reformed
Pension cost
Proposals
?
New national system
?
Focusing on social contribution
? Gradual access to minimum pension
?
Co-financing
? Raising the contribution rate
90%
-1%
nd
nd
2%
nd
nd
Pension cost with reforms
89%
Variation
27
-1%
Current value as a percentage of the 2006 GDP
Public system
52%
SNP
31%
20530
21%
Private system
6%
Recognition bond
3%
Transition
3%
Pension cost
58%
Proposals
10%
-> Reduction of the replacement rates -2%
Gradual
->
access to the minimum pension 3%
-> Coverage
9%
Pension cost with reforms
Variation
68%
16%
Impact of the measures
In synthesis, we believe that:



The coverage rate can grow an additional 20% with respect to the passive
scenario.
The substitution rate (pensions) can grow by about 22%.
The annual fiscal cost is controlled and can fluctuate between 0.2 and 0.5% of the
GDP
Furthermore, it is important to emphasize that an integral approach is required for
improving the results of the pension system, which implies adopting measures outside its
scope (policies that will enable reducing the levels of informality and increasing worker
productivity will have a positive impact on coverage and pension levels).
28
Sensitivities to bear in mind in the
future of pensions
Sensitivity of the pension to a variation in one of the parameters, leaving the others as in
the passive scenario (Peruvian case) :
Pension
(from 6 % to 7%)
Profitability
(from 3 % to4%)
Productivity
(from 10 % to 11%)
(from 65 to 66)
(from 35% to 36%)
(from 14 to 15 years)
24%
18%
Contribution rate
10%
Retirement age
3%
Density
3%
Life expectancy-7%
-10%
0%
10%
20%
30%
Conclusions
Under current conditions, the reforms would generate improvements in the
main indicators in the mid-term. Nonetheless, the progress would be
insufficient for achieving adequate protection in old age.
In this regard, the proposed measures would enable achieving a significant
reduction in the percentage of the population outside the pension system.
Similarly, pensions improve in relation to the base scenario and this
improvement is significantly better for lower income workers.
Conclusions
Further to the above, it is worth mentioning that an integral approach is required to
improve the results of the pension system, which implies adopting measures outside
its scope. Thus, the implementation of policies that enable reducing informality
levels and increasing worker productivity will have a positive impact on coverage
and pension levels..
Along these lines, fostering the adequate functioning of the labor market, raising
educational and human capital levels and gauging the socio-economic
characteristics of the population, will help to substantially increase the performance
of the pension system.
The challenges still facing the pension system are formidable, so it is necessary to
take measures immediately in order to achieve adequate protection for the
population in retirement, especially for those groups that generate low income levels
during their active life.
BBVA is fully committed to the performance of studies that enable the generation of
consensus among the relevant actors in the pension system for the purpose of
consolidating its economic-financial viability and achieving social and political
legitimacy.
[email protected]