the cost of minimum pension guarantee

The Cost of Minimum Pension Guarantee

What is the problem we study?

• Governments often promise a minimum level of benefits under an accumulation scheme

• If future does not turn out to be rosy, what is the likelihood that the government has to foot the bill of this guarantee?

• Using the actual experience of the past eight years, and including actual features of the Mexican system, we calculate the probability distribution of such promises

• Pension System in Mexico

• Investment Regime

• Minimum Pension Guarantee

• Model

• Results

• Conclusions

Current Mexican pension regime (for the formal sector)

In 1997, Mexico moved from a defined benefits system (a la US Social Security) to a defined contribution system (a la Chile)

The system is publicly mandated but funds are privately managed

The funds are called AFORES (Administradoras de Fondos para el Retiro)

There are three components to each fund: government component, private compulsory component and private voluntary component

There is a government contribution

There is also a government guarantee

Contribution is 6.5% of the base salary (SBC)There are three component

Worker

Employer

1.125%

Monthly Contribution as percentage of SBC

5.150%

0.225%

6.5%

Govt

TOTAL

Other Contributions

This is called Cuota Social (cs). 5.5% of minimum wage indexed for inflation

Each individual can also make voluntary contribution to the system. But such funds are maintained separately

Note: There is an upper limit to the SBC equivalent to 25 times the minimum wage




• Model

• Results

• Conclusions

Investment regime under the new system

The AFORES invest in special funds called SIEFORES.

Until 2004, the SIEFORES had extremely restricted investment regimes. For example, investment in anything other than high grade inflation protected bonds were not allowed

In January 2005, new investments were allowed such as foreign bonds in currencies other than pesos as well as in other instruments provided that capital is protected from erosion (through synthetic instruments such as futures and options)

ACTUAL INVESTMENT STRUCTURE IN THE PENSION SYSTEM

SIEFORE-1

SIEFORE-2

Allowed Financial Instruments & Limits Affiliated Workers

Debt Instruments with inflation protection – lower limit 51%

Foreign Debt – up to 20%

Assigned Workes

Workers over 56 years old

Workers who choose to invest under this SIEFORE

Foreign Instruments – up to 20%

Equity Notes – up to 15%

Workers under 56 years




• Model

• Results

• Conclusions

The nature of minimum pension guarantee

The government promises to pay a minimum level of benefits if the amount of money accumulated in the worker’s account does not reach certain minimum value

What does the government promise?

The Federal Government promises the equivalent of one minimum wage to any worker who has contributed to the new system for 1250 weeks or 24 years (even if it is not continuous)

The government promises to pay an annuity of one minimum salary for life for each worker in the system

Although small today, the number of people who become beneficiaries of this promise is rising

Source: IMSS data

Number of people entitled to minimum pension guaranteethousands,2005

Government burden of such a guarantee isrising...(millions of pesos)

1997 1999 2001 2003 2005 1998 2000 2002 2004

4.34

3.5

2.8

2.2

1.7

1.3.9

.19

86

66

49

35

25

17

8

1 to 3 MW

4 to 5 MW

6 to 10 MW

More than 10 MW

....and this will continue to rise because the workers under the regime largely have low wages (with over half earning

no more than 3 times the minimum wage)

M.W.= Minimum Waves

Source: IMSS

55.68%

20.65%

15.06%

8.6%Salary level of workers under the IMSS system2004




• Model to Calculate the Cost of the M.P.G.

• Results

• Conclusions

The model is as follows: The first element describes the distribution of the variable rate of return, the second

describes the law of motion, the third is a contribution at each period taking into account the commission charged

),( 2MMMt Normalr

ttftr

t CVrVV Mte )1)(1(1

csVComAFOREC ttit )065(. ,

Equity Returns

Accumulation

Contribution

The rate of return is the actual rate that prevailed during 1997-2005 in Mexico

= 0.00757452 equivalent of 9.48% annual

= 7.12315%

Adjusted Log Likelihood = 135.017

For the riskless rate, we took the real rate of return of Bondes182 (a government bond with inflation indexation)

rf = 4.63% annual

We assume that a fraction is invested in the broad stock index and the other fraction (1- ) is invested in government bonds

Where: ComAFOREi,t Fee charged by AFORE i at time t cs Cuota Social

ttftr

t CVrVV Mte )1)(1(1

csVComAFOREC ttit )065(. ,

Monthly rate of return of the Mexican broad market index IPC was indeed Normal during 1997-2005

Underlying assumptions for running simulations

- Retirement Age. 65 years

- Contribution period. 25 and 40 years respectively

- Contribution Rate. 6.5% of base salary

- Contribution frequency. Monthly

- Commissions. We use the actual and projected commission structure

taking into account the loyalty discounts offered by some AFORES

- Inflation. We calculate everything in real terms

Measuring the guarantee costWe can conceptually think of the guarantee cost as an implicit put option for the government at the retirement age

Where

PMG: Price of a contingent annuity that pays the equivalent of one minimum salary in real terms

VT: Funds accumulated in the worker’s individual account at retirement

Payoff (T) = max{ PMG – VT , 0}

PMGVT

Pay

off

We could value the option using traditional Black Scholes option pricing model (it requires risk neutral valuation and the assumption of complete contingent markets)

Simulation

Cashflows

Average

1,000 realizations of the final amount for each level of salary assuming different levels of investment in equity

We calculate the present value for each trajectory

The average payment is the value of the option

We need to calculate the single premium contingent annuity of one minimum salary (the government guarantee)

The net premium for such an annuity is calculated using the following euqation

)1(*))1.9(.*( )12()12(97 smfääSMP xxy

Where

SM97: Minimum salary current in 1997 but brought forward to 2005

f : Administrative and acquisition fee 1%

sm: Security Margin of 2%

We also assume that at retirement the worker is married and his spouse is four years

younger than him (average in Mexico)

*We use the mortality table used by the Mexican Social

Security





• Results

• Conclusions

50%

100%

25%

75%

5 MW3 MW

10 MW 25 MW

1 MW

0

0.2

0.4

0.6

0.8

1

*Measured in multiples of minimum wage

PROBABILITY TO EXERCISE THE GUARANTEE UNDER THE ASSUMPTION THAT THE WORKER

CONTRIBUTES TO THE SYSTEM DURING 40 YEARS

Probability

Income* Equity Fraction

MARKETLambda

Percentage Invested in Equity

INCOME* 0% 25% 50% 75% 100%

1 1.000 0.549 0.268 0.224 0.231

2 1.000 0.103 0.103 0.115 0.142

3 0 0.008 0.024 0.055 0.087

5 0 0 0.005 0.022 0.042

10 0 0 0 0.003 0.009

25 0 0 0 0 0

WHEN WE REDUCE THE CONTRIBUTION TIME TO 25 YEARS, THE RESULTS CHANGE DRAMATICALLY

MARKET Lambda


INCOME*0% 25% 50% 75% 100%

1 1.000 1.000 0.967 0.829 0.728

2 1.000 1.000 0.863 0.680 0.603

3 1.000 1.000 0.694 0.499 0.443

5 1.000 0.893 0.327 0.284 0.297

10 0 0 0.034 0.058 0.099

25 0 0 0 0.001 0.0060

50%

100%

25%

75%

5 MW3 MW

10 MW 25 MW

1 MW

0

0.2

0.4

0.6

0.8

1

Equity FractionIncome*

Probability


15%

50%

75%

100%

5 MW

10 MW

1 MW3 MW

25 MW

50,000

100,000

150,000

200,000

250,000

Equity FractionIncome*

Option Price

We also calculate the cost of the option for the government

MARKET Lambda


INCOME*0% 25% 50% 75% 100%

1 216 201 158 127 116

2 181 161 103 85 81

3 147 119 63 50 54

5 78 42 21 24 32

10 0 0 1 3 7

25 0 0 0 .073 .256

Option price, 25 yearsthousands


LOS COSTOS DERIVADOS SON REDUCIDOS DE FORMA IMPORTANTE CON UNA MAYOR

FLEXIBILIDAD DE INVERSIÓN EN EQUITIES

0.00

20.00

40.00

60.00

80.00

100.00

120.00

140.00

160.00

0% 20% 40% 60% 80% 100%

Equity Fraction

If the SIEFORES could invest up to 50% in Equity:

$$63,91363,913

SIEFORE-2:$119,804$119,804

SIEFORE-1:$ 147, 491$ 147, 491

53% Cost Reduction

Guarantee Price for a worker who earns 3 M.W.





• Results

• Conclusions

CONCLUSIONS

Minimum pension guarantee is present to serve a social purpose: protecting low income individuals from falling into poverty

But it also binds the government to future costs High commission (between 20% and 30% of contribution) directly

affect the minimum pension guarantee: therefore, governments should seriously consider strategies for reducing commissions

The capital protection regulation currently in place does not eliminate all downside risks

While allowing for investment in broad stock indexes can increase upside for the workers, it can significantly increase the burden for the government especially for low income workers

Income profile of AFORE affiliates who contribute

regularlyTapen Sinha

How did income increase during 1997-2005 among affiliates?

• Main question• Why is it important?• To know how much affiliates would have in

their AFOREs during their lifetimes• To have an idea of how income changes in

a life cycle context in Mexico in the formal sector

• To disentangle the gender gap in income

Limitations

• There are few observations in some of the cells – thus, they are not reliable (example: women in the highest quintile)

• We do not get a picture of all who are contributing – only the ones that are contributing regularly

• We only have data for 1997-2005 much less than lifetime data

Original study

• With macroeconomic data

• Modigliani Bloomberg life cycle hypothesis

• Milton Friedman permanent income hypothesis

• Both posit that people smooth consumption over lifetime income

• Income has less volatility than consumption

IncomeConsumption

income

consumption

borrow? working life retire death

“Typical” income consumption profile

Age consumption income profile in the US matches the typical one......

....but the Japanese data does not match the typical...

Data

• For men and women separate samples• For each quintile, 1000 persons for each age group• 20 or below, 21-25, 26-30, 31-35, 36-40, 41-45,

46-50, 51-55, 56-60, 61 and above• Total ten categories• Thus, (in theory) we have 50,000 observations for

men and women – we have less numbers• Each person has a salary figure for August 1997 to

February 2005 every two months (46 obs)

Simple analysis

• We can collapse all the observations by examining the average salary over (almost) eight years and examine how average salary changes with all the persons put together taking into account sex of the person and the age of the person

• This will be similar to the analysis of the US and Japan (we saw in the OECD Report in the previous slides)

Estimating for the entire sample

• Estimation Equation:

• Log(SalProm) = c0 +c1.sexo +c2edad +c3edad2

• LOG(SalProm) = 4.482169548 - 0.1617908224*SEXO + 0.0318998435*EDAD - 0.0003307339208*EDAD*EDAD

• Highly significant coefficients

Separate estimates

• Males

• SalProm = -9.935 + 9.311*EDAD - 0.0887*EDAD*EDAD

• Females

• SalProm = 18.471 + 6.712*EDAD - 0.06534*EDAD*EDAD

When does the income starts to decline

• From separate estimates:

• Male 52.47

• Female 51.35

• From combined estimate:

• Highest income at 48.23

Wage Equation Estimated

4.6

4.65

4.7

4.75

4.8

4.85

4.9

4.95

5

5.05

5.1

5.15

20 25 30 35 40 45 50 55 60 65

Age

log

(Wa

ge)

male

female

48.22 years

What is wrong with the pictures?

• They are calculated from by combining all the people of different age groups

• They do not tell us anything about cohort effect• Here, we can actually observe the real trajectory

of persons over eight years• What we need to know for constructing income

profiles of actual workers is what they earn over their lifetimes – we have segments in our data

Household Savings and Income Distribution in MexicoOrazio P. Attanasio and Miguel Székely, 1998

Analyzing data controlling for cohorts and gender

• Now examine the data, controlling for "starting income" (which is a proxy for cohort) and follow them for eight years

• How should it look like?• Logically, it should depend on the economic

performance (more about that at the end)• It should also depend on cohort• For young higher growth than old

Time

Wage

Young

Old

Observations from data….that leads to….

1997 2005

20s 30s 40s 50s 60s

Wage…the following lifetime wage profile

0.5

0.55

0.6

0.65

0.7

0.75

0.8

0.85

0.9

Mar-97 Aug-98 Dec-99 May-01 Sep-02 Feb-04

m60m55m50m45m40m35m30m25m20m15

Males in Quintile 1

Each point is average

0.6

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

1.5

1.6


m60m55m50m45m40m35m30m25m20m15

Males in Quintile 2

0.7

0.9

1.1

1.3

1.5

1.7

1.9

2.1

2.3


m60m55m50m45m40m35m30m25m20m15

Males in Quintile 3

Males in Quintile 4

1

1.5

2

2.5

3

3.5


m60m55m50m45m40m35m30m25m20m15

Males in Quintile 5

2

3

4

5

6

7

8

9


m60

m55

m50

m45

m40

m35

m30

m25

m20

m15

Time

Wage rate Stylized Fact

Men in 20s

Men in 30s

Men in 40s

Men in 50s

0.5

0.55

0.6

0.65

0.7

0.75

0.8


w60w55w50w45w40w35w30w25w20w15

Females in Quintile 1


0.6

0.7

0.8

0.9

1

1.1

1.2

1.3


w60w55w50w45w40w35w30w25w20w15


0.6

0.8

1

1.2

1.4

1.6

1.8


w60w55w50w45w40w35w30w25w20w15


0.9

1.4

1.9

2.4

2.9


w60w55w50w45w40w35w30w25w20w15


2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

7


w60w55w50w45w40w35w30w25w20w15

Men and woemn 20s and 60s in the third quintile

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Aug

-97

Nov

-97

Feb

-98

May

-98

Aug

-98

Nov

-98

Feb

-99

May

-99

Aug

-99

Nov

-99

Feb

-00

May

-00

Aug

-00

Nov

-00

Feb

-01

May

-01

Aug

-01

Nov

-01

Feb

-02

May

-02

Aug

-02

Nov

-02

Feb

-03

May

-03

Aug

-03

Nov

-03

Feb

-04

May

-04

Aug

-04

Nov

-04

Feb

-05

Wag

e R

ate m60

m20

w60

w20

Stylized facts

• For the three middle quintiles, all income of all ages grow

• Men's income always stays above that of women's• For men and women in their twenties, the income

grows faster than for men in their forties, fifties and sixties

• Wage gap between men and women is expanding for women in their twenties but not for them in their sixties (see the following slide)

Women's Income as a Proportion of Men's

0.75

0.8

0.85

0.9

0.95

1

Aug-9

7

Feb-9

8

Aug-9

8

Feb-9

9

Aug-9

9

Feb-0

0

Aug-0

0

Feb-0

1

Aug-0

1

Feb-0

2

Aug-0

2

Feb-0

3

Aug-0

3

Feb-0

4

Aug-0

4

Feb-0

5

Inco

me

ratio

w60/m60w20/m20

We ignore this segmentToo few data points

Dependent Variable: M20

Variable Coefficient Std. Error t-Statistic Prob. DATE 0.036276 0.003567 10.16877 0DATE*DATE -7.20E-05 7.36E-05 -0.978116 0.3335C 0.639283 0.036348 17.58766 0

R-squared 0.970604 Mean dependent var 1.43933Durbin-Watson stat0.439952 Prob(F-statistic) 0

For the men in their twenties in the third quintile

Dependent Variable: M50

Variable Coefficient Std. Error t-Statistic Prob. DATE 0.026763 0.002397 11.16667 0DATE*DATE -0.000313 4.94E-05 -6.324 0C 1.451821 0.02442 59.45233 0

R-squared 0.915035 Mean dependent var 1.85297Durbin-Watson stat0.754211 Prob(F-statistic) 0

For the men in their fifties in the third quintile

Wage Varies With Age Group

0

0.5

1

1.5

2

2.5

Aug

-97

Feb

-98

Aug

-98

Feb

-99

Aug

-99

Feb

-00

Aug

-00

Feb

-01

Aug

-01

Feb

-02

Aug

-02

Feb

-03

Aug

-03

Feb

-04

Aug

-04

Feb

-05

Wag

e ra

te

m20q3m50q3

Applying the naïve model

Econometric issues

• Wage rate series for almost all age groups for almost all income quintiles have unit roots

• Thus, simple correlations among level series give spurious correlations

• How do we deal with the problem?• One solution: Instead of the original series,

manage the growth rate of the series

0.4

0.8

1.2

1.6

2.0

2.4

5 10 15 20 25 30 35 40 45

M15M20M25M30

M35M40M45M50

M55M60

-.3

-.2

-.1

.0

.1

.2

.3

5 10 15 20 25 30 35 40 45

DM20DM25DM30

DM35DM40DM45

DM50DM55DM60

Result

• Even though the original series look quite dispersed, the growth rate (log difference) are almost perfectly synchronized

• What would they be related to?• One possible culprit: the monthly GDP

Index produced by INEGI (actual GDP figures are only available every three months, our series is available every two)

100

110

120

130

140

150

0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2

M30

GD

PIN

DE

X

A simple model

Dependent Variable: M30Variable Coefficient Std. Error t-Statistic Prob.

GDPINDEX 0.032815 0.003322 9.878397 0C -2.3587 0.419274 -5.62567 0

R-squared 0.689228 Mean dependent var 1.774597Adjusted R-squared0.682165 S.D. dependent var 0.322315S.E. of regression0.181711 Akaike info criterion -0.5303Sum squared resid1.452829 Schwarz criterion -0.45079Log likelihood 14.19678 F-statistic 97.58273Durbin-Watson stat1.370915 Prob(F-statistic) 0

But the model is wrong

• Why?

• Using Augmented Dickey Fuller test we cannot reject the hypothesis that M30 and GDP Index have unit roots

• Next logical pass: cointegrated model

Cointegration test results

Unrestricted Cointegration Rank Test (Trace)Hypothesized Trace 0.05No. of CE(s) Eigenvalue Statistic Critical ValueProb.**

None * 0.34862 28.62281 15.49471 0.0003

We reject the null hypothesis that thereis no unit root: one unit root

What comes first? Change in GDP or Wage rise?

Pairwise Granger Causality TestsSample: 1 46Lags: 4

Null Hypothesis: Obs F-Statistic Probability

GDPINDEX does not Granger Cause M30 42 4.54527 0.00492 M30 does not Granger Cause GDPINDEX 5.01425 0.00286

• Lump sum– Usually there are many restrictions– Justified to prevent "double dipping"

• Phased withdrawal (also called programmed withdrawal)– Problem of adverse selection

• Annuities– Individual or joint– Criticism against individual annuities

Options for Retirement BenefitsOptions for Retirement Benefits

Country

Payout Market

StructureLump-sums

Allowed?Degree of

AnnuitizationDegree of

PWs

Australia Competitive Yes Very Low Low

Chile Competitive Yes, but restricted > 60% < 40%

Denmark Competitive Yes, but restricted > 50% < 50%

Sweden Monopoly No 100% 0%

Switzerland Competitive Yes, but restricted Very high 0%

Colombia Competitive Yes, but restricted

Hungary Competitive No Starts in 2013 Starts in 2013

Poland Competitive No Starts in 2009 Starts in 2009

Source: Rocha, IIFA, Istambul, 2006

Types of PayoutTypes of Payout

In Chile (best case scenario)

Year Programmed Withdrawal

Annuities Insurance

1988 32.20% 10.00% 57.80%

1990 42.10% 23.30% 34.40%

1995 46.00% 39.60% 11.20%

2000 40.60% 52.20% 5.30%

2005 37.30% 58.60% 2.80%

2006 38.30% 57.90% 2.60%

Source: SAFP, 2007

• Where is the meat of contribution?• 6.5% of base wage (upto a limit) for

retirement• 5% on housing account• 2.5% on life and disability• Workers compensation• All the money in bold red are

managed by the ubiquitous IMSS and in blue by INFONAVIT

What happens in Mexico?What happens in Mexico?....and why should the rest of the world care?....and why should the rest of the world care?

• SAR-97 reform treats differently members of the “transition cohort” from members of the “new cohort.”

• If a new entrant to the labor force after July 1997 ( “new cohort”) becomes disabled or dies, she or her beneficiaries must buy an annuity from a private insurance sector specialized annuity provider.

RulesRules

Where is the meat?Where is the meat?

Workers Number Distribution

Transitional generation

7,743,057 63 percent

New generation

4,604,993 37 percent

Total 12,348,050 100 percent

Source: IMSS, data for 2004

Source: CNSF, 2008

Types of Companies in the Annuities Market

0%

20%

40%

60%

80%

100%

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Other

Financial Groups

Foreign Subsidiaries

Number of companies

0

2

4

6

8

10

12

14

16

Life and Pension

Specialized

Source: CNSF, 2008

• The 1997 LGISMS reform stated that insurance companies licensed to practice life insurance were allowed to offer pension annuities for a five years transition period.

• After that period, those insurance companies would have had to separate annuities operations in specialized pension annuity companies.

• This explains the transition

Why all specialized companies now?Why all specialized companies now?

Concentration of top 5

60

65

70

75

80

85

90

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Source: CNSF, 2008, vertical axis in percent

Bancomer 32.8%Porvenir GNP 18.3%Génesis 14.7%Inbursa 13.3%Comercial América 6.2%

BBVA Bancomer 23.0%Pensiones Inbursa 15.8%Profuturo GNP 15.5%Pensiones Banorte 10.9%Pensiones Banamex 10.6%

• Foreign subsidiaries have come to dominate– This is a general pattern due to NAFTA

• Number of companies have shrunk somewhat – dominated by specialized companies– This is due to laws related to who can

sell• Concentration in the industry has

stayed the same since 1999

Summary of what has happenedSummary of what has happened(the tip of the iceberg)(the tip of the iceberg)

Performance of annuity market

LR = the loss ratio; OR the operational expense ratio; UR = the underwriting expense ratioSource: CNSF

0%

20%

40%

60%

80%

100%

Pensions

Accidents & Health

Life

Other P&C

Motor

Source: CNSF, 2007

The Iceberg: Pension becomes a The Iceberg: Pension becomes a significant player significant player but then it shrinks but then it shrinks

A closer look: A clear drop from the peak in 2001A closer look: A clear drop from the peak in 2001

The Rise And Fall of Annuitants

0

5,000

10,000

15,000

20,000

25,000

30,000

35,000

Eligible Insured

Annuitants

• When the new regulatory regime came into effect in Mexico, the initial plan of 1995 called for buying single premium annuities for the widows and disabled workers under the new system.

• Given that all these affiliates under the new regime would not have enough money in their accounts, the government (through the IMSS) would pay a lump-sum to pay for such annuities

• Specialized companies would enter this market

The Initial PlanThe Initial Plan

Additional benefits offered by private companiesAdditional benefits offered by private companies

Company Additional Benefits %Allianz 14.5

Inbursa 12.0

Comercial América 8.0

BBV 7.3

Banamex 7.0

GNP 7.0

Bital 7.0

Banorte 6.5

Génesis 6.3

Bancomer 3.7

Source: Own, data from 2000

• Before 1997, during the defined benefit regime, beneficiaries were paid annually and no annuity was bought

• Under the new regime, IMSS observed two facts between 1997-2000– First, cash outflow rose dramatically to pay for the

disabled workers and the widows– Second, more than a third of the disabled workers

died within three years

What happened during 1997-2000What happened during 1997-2000

• It depended on the size of the annuity• On the average, the cost was between

800,000 and 900,000 pesos per case• With over 30,000 cases in 2001, it put

severe pressure on the budget of the IMSS (which still had a pay as you go structure – the current expenditure came from current year's budget of the federal government)

How much did it cost per annuitant to IMSS?How much did it cost per annuitant to IMSS?

• A study by IMSS showed that the minimum pension guarantee is a little less than USD 200 per month

• This pension is 10% more than what people got under the old law

• The lump sum required for a worker receiving a minimum pension is of a little less than a USD 100,000.

Summary of additional outflow of IMSS due to annuitiesSummary of additional outflow of IMSS due to annuities

Comparative benefits under 1973 versus 1997 regimes

Source: Fernandez Reyes (2004)

• In order to lower the costs, IMSS adopted a new policy:

• All disability pensions were going to be treated as provisional pensions

• This would avoid the lump sum payment of buying a single premium annuity for at least two years

• Why two years?

How much did the cost rise for IMSSHow much did the cost rise for IMSS

• By law, all temporary pensions must be renewed as temporary or changed to definitive pension within two years

• After this two year period rule during 2001-2003, outflow due to single premium annuity were dramatically reduced

• Most of the pensions granted were "renewed" temporarily instead of reassigned as "definitive"

Why two years?

• Mexican law establishes that a disability pension must be granted whenever a person is unable to fulfill the exact same job they were doing before the disease or accident

• The IMSS established more stringent medical criteria before granting a permanent disability pension

• Essentially, a person had to prove to be VERY sick in order to get it.

This ruling became know as theThis ruling became know as theLast Temptation of the IMSSLast Temptation of the IMSS

• In Mexico, proving one's identity can be difficult

• For tax and social seucurity purposes– First, the names have to match exactly– Second, the identity by RFC have to be

right– Third, new identity by CURP have to be

right• Many people get multiple Ids• It affects them severely when they try

to collect their social security benefits

Common problems got compounded with tighetened Common problems got compounded with tighetened requirementsrequirements

• The IMSS did an internal study• It projected forward the cost had the

lump sum payment to buy annuities continued (called the "benchmark")

• It projected forward the cost with the new policy of payment with the tightened policy

• The comparison is shown in the next slide

Does the IMSS REALLY reduce the cost in the long run?Does the IMSS REALLY reduce the cost in the long run?

Cost of two options 2005-2068

0

100200

300

400

500600

700

800

2005 2025 2045 2065

year

(mil

lion

US

D)

Current Policy

Benchmark

IMSS study of cost comparison

• The big spike in the cost of year 2006 from the benchmark comes from assuming the costs of all the lump sums generated by the pensions that were handled as temporary but were supposed to be definitive.

• The current policy is cheaper until the year 2053

• Which one is cheaper? • We have to choose a discount rate to make a

comparison of the two vectors by converting them into scalars

Some observations of the cost projection 2005-2068Some observations of the cost projection 2005-2068

Comparison of the regimesComparison of the regimes

Rate PVCurrent PVBenchmark0% $276,993.62 $314,310.602% $132,821.26 $163,091.39

3.5% $83,659.71 $109,329.315% $56,633.42 $78,594.50

10% $23,065.99 $37,416.61

• New policy has reduced the “cost to IMSS” in the NPV sense in the order of $20 billion pesos (at a benchmark rate of 3.5%)

• However, such a reduction in cost has a price in the future

• The market for contingent annuities has collapsed

• This reduction in competition could increase the transactions cost of buying annuities in the future thus compensating for the saving that IMSS has generated

ConclusionsConclusions

• Institutional structures are important– It is possible that loopholes in reform can

allow the (bad?) old regime to play havoc in the annuities markets

• Strict regulation does not necessarily benefit annuitants– Mexican companies offering annuities have

heavy restrictions on investment regime– A quick change in market condition can

actually make matters worse – it also discourages new entrants

So what? So what? General lessonsGeneral lessons

Cultura Financiera, Información y Elasticidad de la Demanda: El estudio y la Evidencia Experimental de México

Introducción• Una mayor capacidad de elección• La privatización en tradicionales mercados de bienes

y productos • La elección permite a la gente usar la información

privada • La elección crea incentivos para la eficiencia• Ejemplos• La elección escolar• El cuidado de la salud • La seguridad social

Introducción

• En todos los mercados, la eficiencia y los resultados dependen del comportamiento del consumidor

• Altos costos de la toma de decisiones podría colocar a los consumidores en posición para determinar las características, como la marca, utilidad y publicidad más que el precio

Conclusiones

• La competencia no conduce a la eficiencia

• Hace más grande la brecha del estatus socioeconómico

• Las empresas podrían no tener incentivos para proporcionar información eficiente

Problemas de Cultura Financiera

• (1) Suponga que tiene $100 en una cuenta bancaria y el tipo de interés que otorga es del 2% por año. Después de 5 años, ¿cuánto cree que habrá en la cuenta: más de $ 102, exactamente $ 102, menos de $ 102?

• (2) Imagine que la tasa de interés en la cuenta de ahorro fue de 1% y la de inflación de 2% por año. Después de 1 año, ¿sería capaz de comprar más, exactamente lo mismo o menos de lo que hoy puede comprar con el dinero de la cuenta?

Problemas de Cultura Financiera

• (3) Si 5 personas tuvieran el número ganador de la lotería y el premio fuera de 2 millones de dólares, ¿cuánto le tocaría a cada uno?

• En EUA, menos del 50% de las personas podrían contestar (1), (2) y (3) correctamente

• Un estudio similar está haciéndose en México y Chile

Estudios

• Los trabajadores son inelásticos al precio

• Altos costos de cambio• Cambios en el trabajo implican

cambios en su duración• Efectos importantes cuando se trabaja

en equipo con los patrones

Estudios

• Los trabajadores responden a la publicidad en lugar del “precio”

• Los consumidores NO tienen visión futura

• Son demasiado sensibles a la información irrelevante

• Comportamiento diferente en los trabajadores con salarios altos

Estudios

• Hastings y Mitchell examinan los vínculos entre la actual tendencia, la cultura y decisiones financieras en los ahorros y pensiones en Chile.

• Experimento de campo y uso de datos de encuestas de hogares

• Implicaciones para el diseño de políticas

Este estudio

• Se discuten los resultados de la encuesta piloto en la ciudad de México

• Este proyecto recoge los datos de la encuesta de una muestra para examinar

• ¿Cómo se mide el impacto de la cultura financiera en la sensibilidad al precio?

• ¿Pueden los políticos usar la información y enmarcar dichas desviaciones causadas por el “analfabetismo” financiero?

Cambio de México

• Objetivos: Reducir la desigualdad presente del sistema anterior, aumentar la cobertura y el monto de las pensiones, hacer del sistema de pensiones financieramente viable

• El sistema toma lugar entre 2005 y 2008

Hasta 2008

• Los trabajadores deben elegir el fondo en el momento de entrar al mercado laboral formal

• El trabajador puede cambiar con bastante facilidad los fondos: Llamar a CONSAR En línea Llamar a las Afores

Tasas

• Las AFOREs cobraban una tasa sobre flujo y/o una tasa de equilibrio

• Comisiones variables está calculado como un porcentaje del salario

• Comisiones variables de 1% = 15.4% del salario (1/6.5 = 0.154).

• El balance de los comisiones son calculados como un porcentaje de los activos manejados

Tasas• CONSAR publica un “Balance de comisiones

equivalentes anuales” (CEF)• El balance de comisiones de una Afore podría

contener un cargo para hacer el mismo efecto al final de un año que si no carga comisiónes variables

• Hipótesis que deben ser formuladas: El balance está establecido con relación al salario del sistema, el cual sobreestima el pago de comisiones,pero hace que el CEF se vea mas en el rango de una comisión de manejo razonablemente alta

Tasas

• Con el fin de cambiar o elegir una Afore, los trabajadores deben firmar un documento constando que ellos han visto y entendido la tabla equivalente de últimos tres años de las comisiones equivalentes.

Cuestionario

• Serie de preguntas sobre el conocimiento del sistema, la elección de comportamiento, la cultura financiera

• Seguido por las opciones hipotéticas • Comisiones anuales solamente• Comisiones sólo en pesos por año• Comisiones sólo en pesos por 10 años• Comisiones y tasas de retorno por un año• Comisiones y tasas de retorno por 1 y 3 años• Cada trabajador es seleccionado aleatoriamente para

obtener 4 de estos cálculos de balance

Cuestionario

• Identificación

• Precios (rendimientos) y nombres de empresas a través de encuestas al azar y dentro de la encuesta

• Preguntan por las 3 primeras empresas clasificados

Cuestionario

• Universidad de Yale: • Encuesta sobre el Ahorro para el Retiro • Estamos realizando una encuesta para entender

mejor la perspectiva de los trabajadores Mexicanos sobre el ahorro para retiro. Sus respuestas de esta encuesta serán completamente confidenciales y usaremos solamente con fines de análisis e investigación en nuestra universidad.

Cuestionario

• La encuesta le quitará aproximadamente 10 minutos de su tiempo. Por favor lea instrucciones para cada pregunta y señale su respuesta. Nos interesa su opinión, y para ello, no hay respuestas correctas o incorrectas.

Cuestionario

• Ahora, nos gustaría pedirle que elija una AFORE en base a la información presentada. Suponga que su hermano favorito, José, está considerando cuál AFORE elegir. José tiene 40 años de edad, gana $3,300 pesos al mes, y actualmente tiene una cuenta de ahorro para el retiro con un saldo de $17,500 pesos. José le ha pedido su consejo sobre cuál AFORE debería de elegir.

Cuestionario

• Suponga que usted sabe que las AFORES cobran una comisión anual, y la CONSAR le ha facilitado una tabla que indica la comisión anual que cada AFORE le cobraría a su hermano José.

Cuestionario

• De acuerdo a su actual conocimiento sobre las AFORES y a la siguiente tabla, que le ha sido facilitada por la CONSAR, indique las TRES MEJORES AFORES que le recomendaría elegir a su hermano José. En la casilla en blanco a la izquierda del nombre de la AFORE, escriba el número 1 para la mejor opción, el número 2 para la segunda mejor opción, y el número 3 para la tercera mejor opción.

Conclusión

• Información simplificada tiene un impacto causal en la elección del comportamiento de los “analfabetos” financieros

• Revisando las comisiones y retornos reduce la elasticidad del precio de la demanada

• Los “analfabetos” financieros son un gran segmento del mercado, y el cambio en su comportamiento puede tener grandes efectos sobre las tasas

• Observando la elección y demanada se puede entender los incentivos de las firmas para diseñar políticas que minimícen las distorciones en mercados privados para bienes públicos

the cost of minimum pension guarantee

Documents