Thought du Jour

Back in 2005, when the Mayor of Mexico City (Andrés Manuel Lopez Obrador) launched his campaign for the presidency, the incumbent President attacked AMLO’s plan to offer provide cash pensions to every resident of Mexico from the age of 70. Fox was a member of the conservative PAN. Its candidate, Vicente Calderón, won a disputed election in which AMLO finished a close second.

President Vicente Fox yesterday criticized the programme of support to the elderly provided by the government of the city of Mexico….

With that we cannot construct schools, universities nor hospitals, the chief executive complained.

…. “To me it seems terribly unjust that some persons, solely because they happen to be elderly, receive money from those who are working”, he explained.

For the head of Government what should be encouraged, in the case of pensions and retirement, is that “during their productive life people do just a little bit of saving, because otherwise a chilling amount of resources will be needed” [from the State].

…. [T]he president announced that he would submit to the Congress an initiative to bring those who work in the informal sector into the tax collection system [i.e., force them to pay taxes].

“Apoyar con dinero a ancianos es ‘terriblemente injusto’, expresa Fox“, La Jornada (Mexico, D.F.), 13 March 2005.

Just a few hours after President Fox spoke …, the Mayor of Mexico City, Andrés Manuel Lopez Obrador, emphasized that the universal citizen’s pension constitutes a measure of justice for those who have worked and contributed to the support of the country.

…. During his speech… he [the Mayor] pointed out that “even the World Bank advises other governments to copy the universal citizen’s pension developed in Mexico City. And they (the federal government) now say that it is unjust “.

…. He also said that “technocrats only bring bad examples from abroad, but they should also notice that European countries grant universal pensions from the age of 65 years”.

Bertha Teresa Ramirez, “López Obrador: con austeridad, viable en todo el país la pensión universal“, La Jornada (Mexico, D.F.), 14 March 2005.

Originally posted in 2005, when TdJ was distributed only as group email. Translated freely from the Spanish by Larry Willmore. The links above are to the full articles. If you are able to read Spanish, you will enjoy them. I have translated only short segments of each article.

This, in Spanish, is a key statement that President Fox made in his radio address to the nation:

A mí me parece terriblemente injusto que a otros, simple y sencillamente por estar como adultos mayores, se les cubra con el dinero precisamente de quienes trabajan.

And this is my translation:

To me it seems terribly unjust that some persons, solely because they happen to be elderly, receive money from those who are working.

In Mexico it had been traditional for incumbent Presidents to remain aloof from politics, but Fox did not respect this tradition. He seized every opportunity to attack AMLO and support the PAN candidate.

Incidentally, AMLO is once again running for President, promising a universal age pension that reaches all Mexicans. The recent initiative of Calderón – extending 70+ pensions to urban areas – takes the wind out of AMLO’s sails. The timing could not be worse (for AMLO).

This post examines possible stigma attached to universal pensions in Mexico City, and discusses briefly a proposed extension of the universal 70+ pension scheme to urban communities in Mexico. (Currently only residents of rural areas are eligible for a 70+ pension.)

I always suspected that there might be some self-targeting of the ostensibly universal age pensions in Mexico City, because it is undignified for a wealthy resident to accept benefits that – for him or her – are tiny. The cash pension is small, less than a minimum wage. The other benefits (social events, free health care and free public transportation) are of little interest to Mexicans of means. The wealthy in Mexico have their own social clubs, never travel in buses or the modern but crowded metro, and they invariably use private, not public, clinics and hospitals.

A Mexican researcher, Marco A. López Silva, confirms my suspicions. His excellent article is available only in Spanish,  so I have translated large parts of it into English. For those able to read Spanish, I recommend that you read the full essay, which contains much more of interest than I include here. Also, Mr López’s words read much better in the original than they do in my quick translation.

Mr López finds evidence of considerable self-targeting in 2001, the first year of Mexico City’s universal pension, but only modest self-targeting ten years later. This is what I would have predicted. Municipal officials in the beginning deliberately promoted the scheme in low-income areas. Mr López, who seems to have access to inside information, claims that the government of Mexico City, when they launched universal pensions, had a budget for only 250 thousand pensioners, and they knew that 328 thousand were old enough to claim an age pension. I would expect the “stigma” effect to be important only for the wealthiest three to five percent of the population.

[W]hen [Mexico City's] universal pension scheme was launched [in 2001], some of its proponents … argued that the pension scheme was self-targeted. …. The reasoning was more or less the following: applications to join the program had to be presented in government-run health clinics, and since wealthy seniors never use these clinics (and probably do not even know where they are located), they would not go there to seek benefits. At the time I heard other arguments: it was said, for example, that the program took advantage of the “stigma” effect. This effect is those who are well-off are “ashamed” to be seen receiving help intended for or needed by those who are poor. It was also said that wealthy seniors would not want to waste time in a queue to ask for benefits – a rather weak argument, because retirees’ [time] is likely to have a low opportunity cost.

Well, it turns out that data thrown up during the first phase of implementation lend support to these arguments. In the graph below, the districts of Mexico City are ordered according to their level of Human Development (X axis) and the pension program’s coverage of old people living in each district. To contrast this graph I used data published by the United Nations Development Program (for 2004), official information published by the Government of the Federal District on the number of old people receiving pensions in each district, and population data from the 2010 census. As you can see, in 2001 the “wealthier” districts (such as Benito Juarez and Miguel Hidalgo) had much lower coverage ratios than the “poorer” districts (such as Milpa Alta or Tláhuac). The difference is enormous: more than 100 percentage points in the extreme cases. In fact, the graph reveals an almost linear correlation between the two variables: the greater the development in a district, the lower the coverage ratio.

Graph: Pension Coverage vs HDI, 2001
Y-axis: Coverage (Beneficiaries/Population >70 years)
X-axis: Human Development Index, 2004

[....]   As the program expanded, the apparent effect of self-targeting became blurred. The graph below is like the previous one, except that the [coverage] data are for 2011. As you can see, the difference in coverage between the poorest delegations (eg Milpa Alta, with 106.4%) and richest (Benito Juarez, 95.5%) is now quite small.

So what happened to the stigma effect? If it did exist, it became weaker as more people obtained benefits (if my neighbour is not poor and has enrolled in the program, it is less shameful for me to ask for the same benefits; the more people participate, the weaker the “stigma”).  [....]

[The federal 70+ pension scheme is limited to rural areas, and is based on Mexico City's universal pension.] It began to operate in 2007, … providing universal coverage to those who were at least 70 years old and were living in communities no larger than 2,500 people. In 2008 … coverage was extended to communities up to 20,000 inhabitants. In 2009 coverage increased again to include communities of up to 30 000 inhabitants. At that level of geographic coverage, 55% of the elderly are poor, and the rest are not. If the proposal of the President [Felipe Calderón, to extend the 70+ pension scheme in 2012 to the entire country] is accepted, the percentage of non-poor beneficiaries is likely to increase to 60%. But is that really important? If Mexico today spends $170 billion [pesos] annually on gasoline (petrol) subsidies, is it not possible to spend $27 billion [pesos] a year to guarantee a pension, very small, for all Mexicans who helped make this country what it is?

The worst thing is that, in the coming weeks, ideological arguments will be used to mask specific interests of the political class. The PAN, which for years opposed universal programs (how can we forget that statement of [President Vicente] Fox, “to me it seems terribly unfair that some persons, simply because of their age, are showered with tax money taken from those who work”?) has decided to convert to universalism … and it suits them to expand the [70+] program in an election year. The PRD [a party of the left] will also support the proposal …. And the PRI [the traditional party, now in opposition] will oppose it, because they prefer to provide more resources to their governors, especially in an election year.

We’ll see who wins. Hopefully the old folks …

Marco A. López Silva, “Despilfarradores versus cuentachiles“, La cuadratura del círculo, 27 October 2011.

The proposed urban 70+ scheme differs from the rural scheme – and Mexico City’s scheme – in that it will be pension-tested so, in effect, a universal minimum pension rather than a universal pension.

Yesterday I wrote that Yucatán is the only state of Mexico to follow the example of Mexico City (and the federal 70+ rural scheme) by offering a universal pension to elderly residents.

I was wrong. I have since discovered that Chiapas has provided universal age pensions since January 2007 and, unlike Yucatán, has done so with great success. The programme is called Amanecer. Its full name is “Adultos Mayores Nuestra Esperanza y Certidumbre”, but is rarely used. The Amanecer scheme provides 238 thousand seniors with monthly pensions of $550 pesos ((US$41). To Qualify, applicants have to provide only proof of age (64 years or older) and proof of residence in the state. The programme is truly universal, and reaches an estimated 98% of the age-qualified population.

The federal 70+ scheme operates in Chiapas, providing universal pensions to everyone from the age of 70 who live in communities with a population smaller than 30,000. There is no mention, in the rules for benefits, that those receiving other pension income are ineligible for Amanecer, and the number of Amanecer beneficiaries suggests that many in Chiapas, aged 70 years and older, might be receiving two pensions each month, for a total of $1,050 pesos. If any reader is in Chiapas, or knows someone in Chiapas, I would very much like to find out if my suspicion is correct.

Using information available on the internet, I am in the process of assembling information on noncontributory pensions in each state of Mexico. This is an ongoing project. Chiapas has the best scheme that I have found so far. Nothing has been written about it, even though the programme should be an example for the rest of Mexico, and for the world. Chiapas is, after all, the poorest state of the Mexican union. It ranks 32nd in the Human Development Index (HDI) for Mexico’s 31 states and Federal District, and 32nd also in the per capita GDP list. Yucatán ranks higher: 19th in the HDI list and 17th in the per capita GDP list.

Total absence of information and analysis is reason enough to travel to Chiapas to study this amazing programme. Chiapas is very tempting, especially today when I am writing from Ottawa, where the temperature is a frigid -21 degrees celsius. But I leave the task to a younger scholar.

As TdJ recently reminded readers, Mexico’s federal government provides universal age benefits to all rural residents of the country through a programme known as “70+”, because 70 years is the age of eligibility. “Rural” was defined as any community with fewer than 2,500 inhabitants when the programme began, in 2007. The following year the allowable size increased to 20,000, and in 2009 it increased again, to 30,000. It remains at that limit for the moment, but the current President wants to extend the programme to all urban areas of Mexico sometime this year (2012).

Only Yucatan, of Mexico’s 31 states with Chiapas, has followed the example of Mexico City and the federal 70+ scheme by providing universal cash benefits with no eligibility requirements other than age and residence. The monthly pension is $550 pesos (US$41), somewhat higher than the 70+ pension of $500 pesos. Thirteen of the 31 states provide no non-contributory pensions at all to elderly residents; the other 1617 provide means-tested benefits rather than universal pensions.

Yucatan’s scheme, known as “Reconocer Urbano”, dates from September 2007 and is universal, but only in intent. Because of budget restrictions, the state government is phasing in the programme very gradually. This presents a rare opportunity for researchers to measure the impact of social pensions on variables of interest by looking at changes over time in a treated community (one where all the elderly receive pensions), in order to compare these changes to those observed in a similar, but untreated community that serves as control.

Social scientists from the RAND Corporation, a US-based think tank, seized the opportunity and organised a project to study the impact of “Reconocer Urbano” in the Yucatan. Working with the state government, they selected Valladolid, the third largest city in Yucatan (population about 46,000) as the treated community, and Motul, with 21,500 inhabitants, as the control.

The choice of the control in retrospect was unfortunate. Universal pensions in Valladolid – the treated community – began in December 2008. Within weeks, the federal government expanded its 70+ programme to include towns such as Motul, with populations below the new 30,000 cutoff. More than 800 residents of Motul applied for 70+ benefits in February 2008. On August 4th they received their first pension payment, a lump sum for the seven months of February-August ($3,500 pesos). Starting in September 2008, they began to receive regular bi-monthly payments of $1,000 pesos. Pensions are slightly higher in Valladolid, and they did begin somewhat sooner, but otherwise there is no difference between the pension scheme of the ‘treated’ community and that of the ‘control’.

The RAND researchers did not let that deter them. They conducted a “baseline survey” in August-November 2008 of residents aged 70 years or older, living in the two towns, and conducted a follow-up survey in June and July of 2009. The follow-up survey took place just before the elderly of Motul received their pensions, but they were definitely expecting to receive them, which might affect the use of Motul as a control. Also, six months is a very short period within which to see significant results. I was thus surprised by the findings.

But I will let the researchers speak for themselves. (As always, you may click on the link to read the full report.)

Our study designs and evaluates a new non-contributory social security program in the State of Yucatan, Mexico. …. In the current paper we provide evidence of the impact of the program based on information collected six months after the implementation of the program ….

Two cities in the Northeast of the State of Yucatan, Valladolid and Motul, were selected for the first experimental phase of the implementation of the social security program. Valladolid was assigned to treatment and Motul to control. ….

Eligible adults in the treatment town, Valladolid, started receiving the social security benefit in December 2008. The first follow-up surveys in both Motul and Valladolid were conducted in June and July 2009, approximately six months after the treatment town received the intervention. ….

To avoid overlap with other Federal government programs that are being implemented in towns with less than 20,000 inhabitants, the third phase of Reconocer Urbano is being implemented in 11 municipalities with more than 20,000 inhabitants. The roll-out of the pension program is done in stages. The first phase started in Valladolid in December 2008 with 1,047 beneficiaries. ….

The initial results are a decline in hunger, cutting meals, and running out of food due to the lack of money. The elderly recipients are spending more on food, visits to the doctor, and medicines and show improvement in memory function. We also find a decline in alcohol consumption. There is some indication of improved subjective well-being and a decline in acute conditions. For many of the indicators we would expect substantially larger effects over a longer period. We collected a second follow-up survey in Valladolid and Motul in 2010, 1.5 years after Valladolid started receiving the treatment and approximately one year after the control town started receiving a treatment from the federal government program (70 y más). Although the fact that Motul now also gets treated, makes it unsuitable any longer as a control town, the treatment of Motul introduces intertemporal variation in treatment that can be taken advantage of in further analyses.

Emma Aguila, Arie Kapteyn, Rosalba Robles and Beverly Weidmer, “Experimental Analysis of the Health and Well-Being Effects of a Non-Contributory Social Security Program“, RAND Working Paper WR-903, 3 November 2011.

I do not understand this fascination with learning what poor people do with small pensions. Results are inevitably similar to those reported here: recipients of pensions eat better, so are less hungry, and they spend more on medical care. Very often recipients (especially females) also spend a portion of their pension money on grandchildren, but that question was not included in this study.

It surprised me that pensions had a negative impact on consumption of alcoholic beverages. I would have anticipated no impact, or a positive impact, but not a negative impact. This result is so unexpected that it is worth reporting in detail:

- The number of respondents that drink alcoholic beverages in the control town Motul (the control town) increased while the number remained constant in the treatment town.

- The number of drinks per day consumed decreased in Valladolid and increased in Motul. There is a 42 percent decline in the number of drinks per day reported in Valladolid (the treatment town).

So, in the treatment town, there was no change in the number of drinkers, and their average consumption fell sharply. In the control town, the number of drinkers rose, as did their average consumption. Something might have caused residents of the control town to drink more, but I don’t think it was lack of pensions.

A minor complaint is the authors’ insistence on labelling Yucatan’s universal pension scheme as “social security” (even in the title!). The Mexican Social Security Institute (Instituto Mexicano del Seguro Social, IMSS) has nothing to do with social pensions. IMSS is a federal organisation that provides health care and retirement pensions throughout the country to those who make the required contributions. IMSS is not in the business of providing noncontributory pensions of any kind.

A major complaint is that Motul is not credible as a control. The researchers carried out a second follow-up survey in both towns in 2010, but do not report the results. I suspect there was nothing of interest to report.

The RAND study of noncontributory pensions in Yucatan is continuing. I will watch for future reports.

Update: Perhaps drinking by elderly residents of the control town increased because they were celebrating the fact that they were about to receive regular pensions from the federal 70+ programme.  Just a thought that occurred to me.