A Plan for a Social Register in Mauritius

A Plan for a Social Register in Mauritius

Christophe Muller

United Nations Development Program Consultant

September 2007

Acknowledgments

I am grateful to many executives and agents from the Ministry of Social Security, the Central Statistical Office, the Ministry of Finance and Economic Development and the UNDP for their support during the mission of five days in Mauritius that corresponds to this report. I particularly thank M. Lobind, Mrs. Cassimally and M. Cuvillier for their efforts that contributed to the success of the mission.

Executive Summary

We present a plan for a Social Register in Mauritius. The Social Register of Mauritius will be a large database of potential and actual social programme beneficiaries. It will be an instrument to assist the Ministry in identifying the beneficiaries of each programme and deciding the level of assistance for each beneficiary. It will also serve to evaluate social programmes and improve their performance.

The plan accounts for the data requirement and technical economic analyses that will be necessary to deal with the identified objectives: Managing the social programmes in an integrated way; Enhancing the targeting performance of programmes; Designing a unified approach to welfare and targeting criteria; Analysing cyclical and structural poverty reduction policies.

We present the data management issues and we indicate the complementary information that need to be gathered, mostly in the form of surveys. We also discuss the statistical methods, the poverty and targeting analyses to conduct, and the revision of social systems.

Finally, an organised calendar accompanied by a matrix of actions is proposed.

1. Introduction

The Ministry of Social Security (from now ‘The Ministry’) is leading the project of the ‘Social Register for Mauritius’ (SRM). The Ministry is responsible for the design and the execution of the project. Meanwhile, it will be assisted by experts from other organisations, notably at the Ministry of

Finance and Economy, and at the Central Statistical Office (CSO).

We propose a plan for this project. One objective of this document is to provide geneal guidelines for the project. It will also be useful as a memorandum of the reflections carried out so far about the project. The technical elements will be presented later in other documents.

The plan accounts for the data requirement and technical economic analyses that will be necessary to deal with the identified objectives:

-Managing the social programmes in an integrated way;

-Enhancing the targeting performance of programmes;

-Designing a unified approach to welfare and targeting criteria;

-Analysing cyclical and structural poverty reduction policies.

The Social Register of Mauritius will be a large database of potential and actual social programme beneficiaries. It will be an instrument to assist the Ministry in identifying the beneficiaries of each programme and deciding the level of assistance for each beneficiary. It will also serve to evaluate social programmes and improve their performance. This is therefore a highly ambitious project, which will require substantial means, time and effort. The whole process may take a few years in total.

An important objective of this plan is having the Ministry’s agents actually implementing the methods rather than merely relying on external and temporary expertise. The project is an opportunity to build capacity at the Ministry of Social Security, at the Central Statistical Office and in other institution such that the Ministry of Finance and Economy (MOFED).

In Section 2 we discuss the data management issues. In Section 3 we indicate the complementary information that need to be gathered, mostly in the form of surveys. We present the statistical methods in Section 4. In Section 5, we turn to the poverty and targeting analyses. In Section 6, we discuss the revision of social systems. Finally, Section 7 regroups the appendices.

Abbreviations:

CSO: Central Statistical Office.

TheMinistry: The Ministry of Social Security.

UNDP: United Nations Development Program.

SRM: Social Register of Mauritius.

MOFED: Ministry of Finance and Economic Development.

2. Data Management

2.1. The situation

The Mauritius social welfare system is complex.

On the one hand, there are two pension systems:

(1)‘Universal pensions’ are composed of: basic retirement pensions; basic widow’s pensions; basic orphan’s pensions; guardian’s allowance; child’s allowance; inmate allowance; enhanced basic retirement pensions; carer’s allowance. The basic retirement pension system is simple: all beneficiaries receive exactly the same amount.

(2)‘Contributory pensions’ include: contributory retirement pensions; contributory widow’s pensions; contributory invalid’s pensions; contributory orphan’s pensions.

(3)On the other hand, several smaller schemes and sub-schemes exist, sometimes overlapping:

1. The ‘industrial injury benefits’ includes: industrial injury allowance; disablement pension; constant attendance allowance; survivor’s pension; dependent pension; orphans’s industrial injury allowance.
2. The ‘social aid benefits’ are still more complex and heteroclites.
3. The ‘entitlement to income support scheme’ includes the following sub-schemes: rice and flour allowance; the medical unit of the Ministry; various disability schemes; social benefits granted to disabled persons; the Ilois Welfare fund; the senior citizens schemes; the senior citizen councils.

(4)Moreover, at least 26 charitable institutions are operating under the aegis of the Ministry. The ‘National Solidarity Funds’ can be divided in: cases covered under severe personal hardship; scoial werlfare; sugar industry labour welfare funds; non-governmental organisation trust funds; probation and after-care service.

On the whole, the global welfare system in Mauritius is complicated, redundant and inefficient. It is also costly to manage.

The beneficiary list that the Ministry manages can be grouped in five databases, each corresponding to a specific welfare programme:

-Universal pensions;

-Contributory pensions;

-Social aid;

-Income support;

-Support for examination fees.

In total there may be about 200 000 beneficiaries, some of which may take advantage of unplanned overlaps between programmes.

There are also a few cases of unemployment benefits, which are almost negligible (about 150 persons). They correspond to social aid delivered to poor unemployed families with head under sixty years old. The allowance is very small.

2.2. Content of the present databases

Each programme database includes a specific set of variables. Appendix A6 shows a specimen of the corresponding electronic forms. The social aid system is organised into thirty different software applications. Appendix A6 corresponds to a few screen copies of one of these applications for social aid. Let us focus on this programme to illustrate the complexity, while also the wealth of the available information.

The database corresponding to the social aid programme includes the richer set of variables among all social welfare programmes. As basic variables, it includes:

- age and gender;

-occupation at the entry in the list;

-localisation;

-selection criteria;

-matrimonial status;

-data on the spouse including her/his own complete electronic form;

-type of residence;

-other member of the household including their own complete electronic form;

-individual income source and amount;

-abandon of child;

-sickness.

The associated software allows the Ministry’s agents to link the electronic forms of different household members, provided their records are present in the social aid database. There are sometimes redundancies in these lists that include all the variables present in the Citizen Master file (see below) and other ones.

Other variables in the databases that are useful for data management and analyses are: reason of the request, type of benefit, religion, date of application and approbation, ownership status of the house, rent amount.

There is also information on the other members of the family: dependency, age, member type, ‘ancient beneficiary’, amount of ‘compassionate allowance’, food aid, sickness occurrence, details of incomes, type and amount, name and address of the employer, formal/informal sector, beginning and end of employment, class followed by the children, beneficiaries of cardio help or wheel chair, ‘separated’, full-time education.

Other information available for prisoners receiving social aid is: jail, liberation date, type of allocation, date of allocation termination, incarceration date, remand dates, conviction dates, sentence dates, sentence duration, proposed and actual release date.

Other information available for ‘beneficiaries admitted in charitable institutions’ is: institution code, reason for child abandon, family link of the tutor.

Other information available for ‘abandoned women’: name and occupation of the spouse, separated, spouse’s employer, spouse’s salary.

Generally, the termination dates do not seem to be well recorded. Many beneficiaries may continue to receive undeserved allocations. To alleviate this problem, control visits have been organised.

Various variables that are present in the files correspond to records that should be completed or updated from time to time. This is in particular the case for: household composition, income, death or migration of beneficiaries, occupation and education. However, many controls have already been included in the computer program of social aid for example. They allow for extensive cross-checking of the file information. Moreover, the history of modification of the files is recorded and could be used for further checking.

The social aid allocations depend on the level of the family income. So, income thresholds at family level are already calculated, although only for a small part of social benefits. Doubts have been emitted about the proper level of the income threshold to use.

One difficulty will be to integrate all the separated databases for different programmes in a unified system. Another one is to retrieve and link the different individuals in the same household. A third one is to relate the SRM recorded population to the whole population in the country.

For dealing with these difficulties, we can first avail of a list (available at the Ministry) of all the persons in the country over 17 years old with the following information:

-age and gender;

-religion;

-occupation at the entry in the list;

-localisation;

-age at recording;

-matrimonial status;

-date of death.

This information set is imperfect as address and occupation change over time and are not updated. Moreover, dead and migrant beneficiaries cannot always be detected, although there exist procedures to do so.

Second, there exists a ‘Citizen Master’ database where every person in the country who has a national identity card number has been recorded. This number is given at birth. This database is at the registry office (‘Etat-civil’). The national identity number itself provides interesting information about the individual, notably its date and location of birth.

Meanwhile, some important variables are missing in these files such as the nationality of the beneficiaries, or various socio-demographic, economic and education characteristics. To deal with these deficiencies, the SRM will be matched with external databases.

Finally, there are electronic forms describing the contributionsfrom employees and employers. These forms could be used to identify the activity sector of beneficiaries and include this new variable in the SRM.

2.4. The data management teams

During the period when a new computer software for the SRM will be installed and tested, a large amount of additional work will emerge which may stress the resources of the data management teams. Some solution must be found to avoid delays in the payment of benefits caused by the reduced availability of the social security officers.

Only five Ministry’s agents (‘Social Security Officers’) are presently managing the databases of social welfare programmes. Therefore, the scarcity of knowledgeable personnel about the database system will be a severe bottleneck for setting the new SRM. These persons should be freed of other tasks during the period of the setting up. There is also another team in Rose Hill for the section ‘Services’.

A chief information officer is in charge of the administration of the whole information system. He will be able to supervise the replacement of the agents occupied with tasks related with the SRM.

The Ministry supervises 42 regional offices. There are two database administrators, two information servers (Port-Louis, Rose Hill) and a hardware team.

2.5. The Software

The database software has been designed and installed by a semi-public company which equips most of the administrations in the country. This information technology package was installed in 1995. The functioning of the software package has not always been perfect. Breakdowns are frequent as well as communication slowness.

Specific software allows the Ministry’s agents to calculate the amount of the benefit which depends on: the date, the socio-demographic characteristics of the beneficiary and her/his household, and all the eligibility criteria.

Moreover, once calculated the payments are often made electronically, directly to the beneficiaries’ bank accounts.

A new software of the new SRM will be designed and installed on the computers of the Ministry of Social Security. The installation should be made jointly by the software provider and by the agents of the Ministry. The Ministry wants to avail of an active SRM, characterised by frequent updating of the procedures themselves. This must be possible with the new software.

The implementation of the new format of the SRM on the Ministry computers will be accompanied by systematic tests of the usual tasks performed by the Ministry’s agents. At this occasion, new manuals for certification forms and data entry in the Register will be implemented.

2.6. The format of the new data base

The social aid database is sophisticated and includes many useful variables. Since the agents of the Ministry are familiar with it and are perfectly in control of its management, a good approach would be to build the new format of the SRM and its screen applications around this database. In sense, the screen applications of the social aid database could be somehow extended to all the other programmes, while eliminating some variables and adding other ones.

The fundamental format of the new SRM will be based on a rectangular source file in which all the useful information will be stored. However, this rectangular file may not be visible in all applications. Indeed, it is often convenient to see only a few variables on the computer screen and sometimes under a presentation shape that is not rectangular.

Each line of the rectangular file will correspond to the record of an individual beneficiary, with all the variables describing the situation and the characteristics of this individual in regard to each of the welfare programmes. The first column will contain the national identity number. The national identity number is deemed to be of excellent quality in Mauritius. All the variables necessary to describe the eligibility criteria of all social programmes need be included in the file. Also, a column will be assigned that will include the household identifying numbers. The value of this variable will be common to all the records corresponding to individuals belonging to the same household.

The following table illustrates the structure of the rectangular file.

NIN / Household number / First variable of scheme no. 1 / … / Last variable of scheme no. 1 / First variable of scheme no. 2 / … / Last variable of scheme no. 2 / … / Last variable of the last scheme / Individual income / First additional individual characteristic / … / Last additional individual characteristic
… / … / …

NIN is the national identity number. Naturally, for a given individual many variables will be empty, notably when she/he does not participate in some programmes.

Using such a rectangular file has many advantages. First, the way the fundamental information is organised and available can be understood by any Ministry’s agent likely to be confronted by these data in her/his work. Second, the data management tasks are very simple with this format: they correspond to simple combinations of columns when defining new variables, and to simple combinations of lines when updating or correcting individual records. Third, this format is well adapted to statistical analyses. The rows correspond to individual statistical units and the columns correspond to random variables. Four, the aggregation at the household level is easy. It is enough to order the file by household number and to regroup the consecutive rows corresponding to the same household. Five, the rectangular file can be easily transformed in ASCII file and then converted in any format legible by a statistical package. In particular, the operations to convert it into a Stata file ready for statistical analysis are very simple. Six, the rectangular file can be easily matched with household survey and census databases that may be reorganised with NIN and household numbers as key identifying variables. Finally, the matching with the contributor files, also managed by the Ministry, will be easy and based on the NIN.

Another file to store is the history of all transactions done with the SRM. This file will be useful for references and cross-checks.