Ministry of Economic Development and Productivity & Regional Development
1990 HOUSING AND POPULATION CENSUS RESULTS
Relative Development Index - Economic Indicator Issue No. 233
FOR MUNICIPAL WARDS AND VILLAGE COUNCIL AREAS ACCORDING TO BOUNDARIES REVISED IN 1991-1992
This paper shows a set of key socio-economic characteristics derived from the 1990 Census but presented according to the new boundaries of local administrative areas (Municipal Wards and Village Council Areas) as redefined in 1991 and 1992.
An attempt has also been made to calculate a Relative Development Index for each Municipal Council Ward (MCW) and Village Council Area (VCA). The 20 wards, 124 VCAs and Rodrigues have then been ranked according to this index.
2. 1990 census data for small areas
The increasing activity in economic and social planning are placing new demands on statistical information in general and on housing and population census data in particular, with growing interest in small area data. For census purposes, it is convenient to use small geographical areas (called enumeration areas) not only for ease of enumeration on the field, but also for control purposes at the various stages of editing, coding, data capture, verification and processing. However, the smallest geographical entity for which the 1990 census data were published was the smallest officially defined administrative region: this was the Municipal Council Ward (MCW) for regions within Municipalities, and Village Council Areas (VCA) for regions within district councils. There were also some regions which fell neither within Municipalities nor within Village Council Areas. These were broken down into subdivisions according to the location of inhabited agglomerations so that census data could be published separately for these subdivisions as well.
3. Revision of small area boundaries
One year after the 1990 census, major alterations were brought to the boundaries of the MCWs followed in 1992 by similar changes in VCA boundaries. This exercise also ensured that all regions within the island of Mauritius were allocated either to a MCW or VCA. Moreover, the regional data published in the 1990 census reports became outdated.
4. 1990 census data according to revised small-area boundaries
There was therefore the need to work out estimates according to the new boundaries to assist local authorities in their budgetting and planning activities. These estimates were derived by recoding the location of census enumeration areas, or parts thereof, to reflect the revised boundaries of the local administrative areas. Although this method does not achieve 100% accuracy it gives sufficiently reliable results for most purposes. A 100% accuracy can be achieved only by recoding each household to the new areas and this is an almost impossible task given the time that it would take.
5. Publication of 1990 census data according to revised boundaries
The publication of all 1990 census reports according to the revised local administrative boundaries would be costly and time-consuming and may not be warranted. It has been thought that a set of key indicators derived from the 1990 census would best serve the general needs of most users of small area statistics. These are shown in Tables 1
while the related concepts and definitions are described at Appendix I. The map of the Island of Mauritius at Appendix II shows the location of MCWs and VCAs in coded form, and Appendix III gives a description of these codes.
There are no administrative sub-regions in Rodrigues and the 14 zones defined at the 1990 Census have been maintained in this paper. Appendix IV shows the location of these zones, and the localities within each zone are listed at Appendix V.
6. Relative Development Index for small areas
The need has been expressed for a set of social indicators which will accurately measure social phenomena, reflect the quality of life in Mauritius, and enable sociological analysis at a level which is conducive to policy formulation. Work has already started on the project, and one proposal put forward has been the calculation of a Relative Social Index for different geographical sub-divisions, following the methodology of the Human Development Index (HDI) of the United Nations Development Programme (UNDP).
The HDI of the UNDP is a country level index. The application of the methodology to geographical sub-divisions of a country requires data at the sub-regional level. Unfortunately, important indicators like life expectancy at birth and per capita real income, which are crucial components of the HDI index, are not available for geographical sub-divisions of the country.
Furthermore, the only readily accessible source of comprehensive sub-regional data is the 1990 Census. Any index intended to measure the current relative development of sub-regions will necessarily be restricted to variables which can be derived from the Census.
It was therefore decided to select, from the 1990 Census, a set of variables which can be used to reflect the extent of relative development of Municipal Council Wards (MCW) and Village Council Areas (VCA) to construct a Relative Development Index. The following twelve variables were identified:
(i) % of households having piped water,
(ii) % of households having electricity,
(iii) % of households having flush toilet,
(iv) % of households living in dwellings made of concrete,
(v) % of households having one or more rooms used for living purposes per person,
(vi) % of households who own their dwelling,
(vii) % of population aged 18 years and over having at least School
Certificate or equivalent educational attainment,
(viii) primary school enrolment ratio,
(ix) secondary school enrolment ratio,
(x) literacy rate of population aged 12 years and over,
(xi) employment rate for population aged 12 years and over,
(xii) % of the employed population who are legislators, senior officials and managers, or
professionals, or technicians and associate professionals.
The values of the variables for each MCW and VCA (listed in alphabetical order) are shown in Table 5
The methodology of the HDI was used to construct a Relative Development Index for each administrative area (MCW, VCA and Rodrigues). If Xij is the value of variable i for the area j, then a development indicator based on variable i can be defined as
Hij = (Xij - Min Xij)/(Max Xij - Min Xij)
where Min Xij is the minimum and Max Xij is the maximum value of Xij over all k (=145) areas.
It is to be noted that Hij = 0 for the least developed area and 1 for the most developed area. An overall development index for each area (Hj) is then obtained as a simple average of the values of the twelve indicators. A National Development Index can be calculated by adding each MCW and VCA index using their population as weight.
It is to be noted that the twelve selected variables are indicators of development (and not deprivation), so that the derived index is a measure of relative development in line with the HDI.
The values of Hij for each of the twelve variables and 145 administrative areas are shown in Table 6
together with the overall development index for each area. Table 7
shows administrative areas ranked in descending order of "development" (as measured by the overall index) together with their population size at the 1990 Census. Chart I shows the geographical distribution of administrative areas by rank order of "development". Areas with highest development (ranks 1 - 20) are unshaded whilst those with lowest development (ranks 121 - 145) are completely shaded. This map should be interpreted in conjunction with the map in Chart II showing the population density by administrative area.
(iv) Limitations of the Relative Development Index
Like the HDI for countries, the Relative Development Index for regions is not a cardinal index and no special significance should be attached to the values of the index for different MCWs and VCAs. The index values serve only to rank the administrative areas according to their relative development.
The HDI of UNDP is based on a few variables which have been selected for their adequacy to reflect specific dimensions of human development. The Relative Development Index had to be limited to variables available from the 1990 Census. Some important dimensions of development such as health and longevity, and access to various services are not captured. Furthermore, some of the variables are correlated and may be redundant, particularly among those relating to living conditions and education. However, the actual values of each variable shown in Table 5, and the corresponding index values shown in Table 6
, allow the use of other combinations and weightings.
(v) Uses of the Relative Development Index
It is hoped that in spite of the limitations of the index, the ranking of MCWs and VCAs based on it, will be helpful to central and local authorities in targetting areas for surveys on poverty and deprivation, and identifying priority areas for integrated development.
It is also expected that this publication will give rise to discussions and suggestions for future improvements in the construction of socio-economic indices for areas. There may be a need for several indices to address specific areas of concern like deprivation, health, education and economic resources. The variables included in the present Relative Development Index are limited to those investigated at the Census. Data may need to be collected on other variables to cover all aspects of a given socio-economic dimension. Furthermore, expertise in multivariate techniques has to be developed to guide the choice of variables that adequately summarise the information available.
(vi) Main findings
An analysis of the results in Table 7
and the map in Chart I indicates that:
(a) Rodrigues has the lowest development score.
(b) In the island of Mauritius, the Village Council Areas with the lowest scores are located mostly in the East, South and West.
(c) Although the densely populated urban wards are relatively more developed, there are a few like Wards 6, 5, and 1 of Port Louis, Ward 1 of Beau Bassin - Rose Hill and Ward 3 of Quatre Bornes which have low development rankings because of the relatively low scores mainly with respect to construction materials of dwellings, overcrowding, education level and type of occupation.
(d) There are some rural areas which are relatively highly developed; in particular, the VCAs of Moka, Camp Thorel, Long Mountain, St. Aubin, Morcellement St. André, St Pierre and Souillac figure among the first 20 most developed areas.
(e) In general, the least developed areas are sparsely populated, and the resultant high cost per person of facilities and services conducive to development may explain their unavailability or inaccessibility.
(f) The 24 least developed VCAs in the island of Mauritius can be grouped into 7 regions, or pockets, which are shown completely shaded in Chart I. These regions, which are described below, account for 10% of the population (100,300) spread over 28% of the area (525 sq. km.).
||Village Council Areas
Gros Cailloux, Petite
Baie du Cap, Bel Ombre
Chamarel, Case Noyale,
Grande Rivière Noire,
Bananes, Mare Chicose,
Olivia, Quatre Soeurs,
Bambous Virieux, Bois
des Amourettes, Grand
||Poste de Flacq
||Arsenal, Terre Rouge
||Island of Rodrigues
||Island of Mauritius
||Republic of Mauritius
Central Statistical Office
Appendix II Map of Island of Mauritius showing codes for Municipal Council Areas
and Village Council Areas
Appendix III Description of codes shown on Appendix II
Appendix IV Map of Rodrigues showing the 14 zones for which data are tabulated