It has been a number of years since I looked in on Mauritius, so I decided to do an update on their universal age pensions. I was surprised to find that the World Bank is once again advocating targeting of the universal benefits, but this time more tepidly. The IMF has joined the World Bank on this issue, and is a vocal proponent of means tests for Mauritian pensions and other social transfers.
First, some background information.
Universal age pensions in Mauritius date from 1958. Only New Zealand has a longer history of universal pensions. Mauritius lapsed from universality on two occasions. The first lapse, from 1965 until 1976, happened without international aid or interference. During this period, government denied basic pensions to those who had enough income to be liable for payment of tax. This test disqualified approximately five per cent of those who would otherwise qualify for a pension by virtue of age and residence.
The second lapse was from December 2004 to June 2005, and was a product in no small measure of World Bank assistance. The World Bank had carried out a “pension dialogue” in Mauritius that produced seven studies, one of which I was able to obtain: “Pensions in Paradise: Modernizing the Mauritius Retirement Income Schemes” (by Edward Whitehouse and John Piggott, with contributions from Yvonne Sin, August 2001). The ‘pension dialogue’ culminated in June 2004 with the release of a 103-page official World Bank report.
The World Bank in 2004 was concerned that the Basic Retirement Pension (BRP) in Mauritius is granted without an income test, indeed without a retirement test, since pensioners often continue to work at their old jobs. The following paragraph captures the spirit of this lengthy report:
On the basic pension, the options for means testing and affluence testing need to be explored in greater depth. The revamping of the social aid scheme needs to be considered (as a basis for means testing), as would self selection on the part of recipients. On affluence testing, the option of using the NPF [National Pension Fund, the tier 2 contributory scheme] administration for reporting of non-BRP income should be given serious consideration, especially given the limited reach of direct taxation. Already, the NPF needs to collect more detailed information on total income of the contributors apart from just the basic salary. Such information could be put to use in screening BRP recipients.
World Bank, Mauritius: Modernizing an Advanced Pension System (Report No. 29588-MU, 30 June 2004), p. 39.
The Minister of Finance, just before the World Bank released its report, announced the end of universal pensions, and the dawn of a new era of targeting in Mauritius.
Government transfers are simply not sustainable if they are universal and open-ended. This puts a limit on our capacity to do more for the poorer segment of the population. I believe that this is an issue that calls for a bold decision. I am therefore implementing this year a targeted approach to Basic Pensions. As from October 2004, payment of Basic Retirement Pension (BRP) will be limited to persons with monthly income not exceeding Rs 20,000 [US$700]. However, there will be a tapering of benefits for those with monthly income in the range of Rs 16,000 to Rs 20,000. Targeting for other basic pensions will be introduced as from January next.
Pravind Jugnauth, Budget Speech 2004-2005, 11 June 2004.
“Other basic pensions” refers to universal pensions for widows, orphans and the severely disabled. The introduction of targeting did not work out quite as planned. First, the BRP remained universal for those 90 years of age and older. Second, BRP targeting did not begin until December of 2004, and targeting of other pensions apparently never took off. Third, and most important, the governing coalition lost the general elections of July 2005, and the new government moved quickly to restore universal pensions.
My Government … [is] determin[ed]… to ensure that our elders enjoy dignity and respect in their retirement. In this context Government will end the humiliation previously imposed on pensioners by abolishing the targeted approach and reinstating [the] universal pension to all pensioners.
Mauritius, Government Programme 2005-2010, Address by the President of the Republic, 29 July 2005.
Now, in 2012, we are beginning to see a re-run of this movie.
The World Bank, in a report of 27 February 2012, acknowledges that Mauritius has a low rate of poverty (8.5% of total population, equivalent to 106,000 people), “reflecting the fact that Mauritius is a country with a well-developed social welfare system that favors equality”. But they fail to acknowledge that universal transfers might be an important part of the egalitarian welfare system of Mauritius. For the World Bank, it seems that targeting always trumps universality.
A key paragraph of the report (#70, pp. 25-26) provides four explanations why “Mauritius is not realizing the greatest possible benefits” from expenditure of 4.4% of its GDP on social assistance, including universal transfers.
First, Mauritius continues to spend most of its resources on untargeted subsidies and benefits. The largest of these is the universal pension, but the government also subsidizes the price of rice, flour, and cooking gas. Except for the rice subsidy, the distributional incidence of these subsidies is regressive. Second, social assistance programming is very fragmented, which results in considerable administrative inefficiencies. …. Third, in those programs that are targeted, coverage of the poorest is quite low. For example, the Social Aid program served 45,000 clients in 2008. This should be equivalent to 40 percent of the estimated number of poor households in the country, but the Household Budget Survey reported that in practice the program covered fewer than 5 percent of the poor. Only 28 percent of Social Aid payments go to the poor, while 36 percent of benefits go to those in the wealthiest two quintiles. Finally, monitoring and evaluation of activities is weak ….
World Bank, Mauritius – First Public Sector Performance Development Policy Loan Program (Report No. 61571-MU, 27 February 2012).
Let’s summarize. Social Aid, which has a budget equal to 0.2% of GDP, is a targeted programme, but fails to reach the poor, and disproportionately helps the wealthy. So, since most of the remaining amount for social assistance (4.2% of GDP) is distributed as universal cash transfers, lets target those and hope that the outcome is better than it is with Social Aid. There is no discussion of administrative costs, which will without doubt increase following a move from universality to targeting.
At least the World Bank is much less confident, much more tepid in its advice today, than it was in 2004. The IMF, in contrast, is very blunt, but its basic message differs little from that of the World Bank.
The current social protection system is costly, but not well-targeted. …. The poor receive only about 13 percent of total social protection payments. ….
Staff recommended that the authorities intensify their efforts to improve the targeting of social protection expenditures. ….
The authorities agreed that targeting could be improved, but pointed out that political economy considerations were important. Some of the relatively inefficient programs (basic retirement pension and subsidies) have a long history in Mauritius and are backed by considerable societal consensus. They explained that it would take further communication efforts to prepare the ground for such reforms in order to avoid adverse reactions. They welcomed the involvement of the World Bank in designing appropriate targeting mechanisms. (pp. 16-17)
Social Protection System
The Mauritian social protection system needs to improve to ensure that the poor benefit more from economic development. Social protection expenditures at close to 4.5 percent of GDP in 2010 are relatively high. Non-contributory and non-targeted pensions account for 1.3 percent of GDP. Only about 50 percent of direct and indirect beneficiaries of Social Aid (the main means-tested social assistance program) live below the poverty line. Old age pension benefits overwhelming go to non-poor households (only 22 percent of recipients of old age pensions are poor). Furthermore, 39 percent of the basic retirement pension benefits go to the richest 20 percent (Table V.1). The two richest quintiles of the population receive close to 58 percent of all social protection benefits, and the richest 20 percent alone receive 37 percent. Eliminating transfers to the richest 40 percent of the population could save significant sums, which could be deployed more effectively in the pro-poor sectors of primary and secondary education and health. (p. 44)
IMF, Mauritius: Staff Report for the 2012 Article IV Consultation (Country Report No. 12/62, 23 March 2012).
The World Bank and IMF have learned nothing from past failures to target social expenditures in Mauritius. The World Bank in its 2004 report overlooked an obvious way to reduce social expenditures without targeting: increase the age of eligibility for a BRP from 60 to 65 years. Well, the government of Mauritius initiated this transition to a new pension age in 2005, at the same time it restored universality. This cost-cutting measure enjoyed wide acceptance by the public. The BRP continues on a sustainable path, without degrading means tests.