POVERTY STATISTICS IN MALAYSIA

                                                      Abdul Rahman Hasan
                                                          Sa’idah Hashim

                                                             March 2001

The views contained in this article are those of the authors and do not necessarily represent the official views of the Department of
Statistics, Malaysia.
1.   Malaysia consists of two geographical entities: Peninsular Malaysia occupies the southern
     part of the Malay Peninsular in Southeast Asia while East Malaysia (comprising Sabah and
     Sarawak ) occupies the northern quarter of the isla nd of Borneo. The two areas are separated
     by about 600 km. of the South China Sea at the closest point. Covering an area of 329,827
     square kilometers, Peninsular Malaysia takes up 40 percent of the area with 11 states and a
     Federal Territory while Sabah and Sarawak forms 22 and 38 percent of the area respectively.

2.   Based       on   the   recent    Population   Census   2000,   Malaysia’s   ethnically   heterogeneous
     population stand at 22.2 million of which 79.9 percent live in Peninsular Malaysia, 11.0
     percent in Sabah and 9.1 percent in Sarawak.              Of these, the ethnic composition of the
     population is about 58% Bumiputera, 24% Chinese, 7% Indian and 11% are Others which
     also includes non citizens.

3.   Historically, the economic and social features of the population has been moulded since the
     colonial times and that has either directly or indirectly shaped the pattern of income and
     poverty distribution among its multiracial population.         Hence, the Malaysian Government’s
     Development Policies was formulated, with the overriding objective of attaining national
     unity and fostering nation-building through the two-pronged strategy of eradicating poverty
     and restructuring society. The first prong to eradicate poverty, irrespective of race whilst the
     second sought to restructure society by eliminating the identification of race with economic

4.   The Malaysian economic growth has accelerated since achieving Independence in 1957.
     During the 1990’s, its real GDP growth exceeded 8 percent until the onset of the economic
     crisis in 1998 (-7.4%).         As a result of the measures taken to combat the ill-effects of the
     crisis, its economy subsequently shows signs of recovery commencing the third quarter of
     1999 and by 2000, achieving a real GDP growth of 8.5%.

Current Development Policies
5.   In the last decade, the country adopted the Vision 2020 concept to set the country’s strategic
     direction for the next 30 years. The successful implementation of the National Development
     Policy during the 1990s which marks the first phase of Vision 2020, has enabled the country

       to achieve rapid growth with wide-ranging transformation of its economic, social, physical
       and institutional landscape.

6.     The launching of the Third Outline Perspective Plan (OPP3) also saw the nation moving into
       the second phase of Vision 2020. During the next 10 years, the National Vision Policy
       (NVP) will form the basis of the strategies and programmes under the OPP3. The National
       Vision Policy is aimed at establishing a progressive and prosperous Bangsa Malaysia
       (Malaysian Race) that lives in harmony and engages in full and fair partnership. Economic
       growth will be promoted alongside aimed at poverty eradication and restructuring of society,
       as well as reducing social, economic, and regional imbalances.

7.     A remarkable progress has been achieved in poverty eradication. In the early 1970’s almost
       half of the households being classified as poor, and in 1997, the incidence of poverty has
       significantly declined to 6.1 percent.     However, the crisis brought a slight upward swing in
       the trend at 8.1 percent in 1998. Efforts to counter the effects of the crisis has borne positive
       results when the incidence of poverty among Malaysians was reduced to 7.5 percent in 1999.
       By 2005, the incidence of poverty was targeted to reduce further to 0.5 per cent. To realise
       this target, the anti-poverty programs will be more target-specific and address pockets of
       poverty in different regions, particularly in remote areas as well as among disadvantaged
       groups such as the Indigenous groups of Peninsular Malaysia and other minorities in Sabah
       and Sarawak. To increase programme effectiveness, existing anti-poverty programmes will
       be consolidated and revitalised under Scheme for the Development of the People’s Well-
       Being.    A key approach to eradicate poverty and improve the quality of the life of the
       population is education. In addition, the provision of health services and other social
       amenities will also contribute towards reducing poverty and enhancing the quality of life.

8.     The objective of this paper are:
     ² to briefly describe the national statistical system in Malaysia;
     ² to describe data sources pertaining to poverty statistics and
     ² to explain income data collection and derivation of poverty statistics.

National Statistical System
9.     The Department of Statistics (DOS) is the main Government agency entrusted with the
       responsibility to collect, compile and disseminate statistics.            The Department operates
      under the authority of the Statistics Act 1965 (Revised 1989) which defines its functions and
      provides the legal backing to undertake these functions.                 Statistics are collected and
      compiled from surveys undertaken by DOS and also from secondary sources and supplied to
      users both in the public and private sectors as well as international organizations. However,
      this Act do not curb other Government agencies from conducting their own data collection,
      mainly for their specific needs, e.g. the compilation of paddy data by the Ministry of

10.   DOS, which is under the Prime Minister’s Department, is staffed with about 2,000 personnel
      and operates with an annual budget of RM60 million.              In order to carry out its functions,
      DOS is organised into 17 Divisions under 3 Sectors, namely the Trade and Industries Sector,
      the Coordination and Support Sector and the Prices and Human Resources Sector.                 It has
      offices in each of the 14 States and some of the larger states are assisted by Operations
      Centres.   Functionally, the State Offices are responsible for data collection and generating
      validated state data whilst the Central Office is responsible for the production of aggregated
      National statistics as well as analysing and disseminating the output.

11.   With respect to poverty statistics and indicators, the data is obtained from various primary
      and secondary sources.        Two household surveys form the main source of primary data,
      namely the Household Income Survey (HIS) and the Household Expenditure Survey (HES).
      The HES carried out on a five year interval, aims primarily at collecting the latest
      information on the levels and patterns of consumption expenditure by private households on
      a comprehensive range of goods and services.                 This in turn provides the basis for
      determining the weights of items of household expenditure, thus forming the basis for
      updating the current price in establishing the Consumers’ Price Index and subsequently the
      expenditure pattern of private households.

The Collection of Income and Poverty Data
12.   The Household Income Survey (HIS) has been conducted periodically since 1974 with the
      latest being in 1999. The main objective of the HIS is to provide data on income
      distribution, the extent and magnitude of poverty and accessibility to basic amenities for use
      in policy formulation. Since the main trusts of the Malaysian Government Development
      Policy is for balanced development, the focus on the eradication of poverty, especially hard-
      core poverty and the reduction of relative poverty, raising the income share of the lowest 30

      percent of the households and increasing the size of the middle income group, continued to
      be emphasised. Continuos monitoring of the changing pattern of income distribution and the
      incidence of poverty is essential to assess and evaluate the success (or failure) of
      Government’s programmes and policies. The data collected from the HIS have been used in
      part for the preparation of the various Malaysia Development Plans, the Mid-Term Reviews
      of these Plans and formulation of income policies.

Sample Design
13. A two-stage stratified sample design has been adopted for the survey. The first stage
    sampling unit is the Enumeration Block (EB), which is a geographical area artificially
      created to have about 80-120 living quarters with about 600 population. The EBs are
      selected using a probability proportionate to size (PPS) linear systematic selection scheme
      which give higher probabilities for larger EBs. This systematic selection also ensures a good
      spread of the sample, thereby enhancing its representativeness. In the second stage, living
      quarters are selected from each of the sampled EBs, also on a linear systematic basis. The
      rate of the selection used is aimed at maintaining the self-weighting feature of the sample
      design. Since the living quarters are ultimate sampling units, all households (i.e. the
      reporting units) within the selected living quarters are canvassed. Geographically, the HIS
      covers both the urban and rural areas in Malaysia.      During the 1999 survey, about 40,000
      households were selected.

14.   The survey covers a representative sample of all private households.                 Institutional
      households, whose population constitute about one percent of the total population are
      excluded.   The concept of a household is based on arrangements made by persons,
      individually or in groups, for food and other essentials for living within the same living
      quarter. Thus, a household may consist of related and/or unrelated members.

Data Collection and Data Processing
15. The HIS is conducted jointly with the Labour Force/ Migration Survey (LF/MS) using a
     common questionnaire. The questionnaires on income and basic amenities are tagged on to
      the LF/M questionnaire. The socio-economic characteristics of the household, for example
      ethnicity, citizenship status, age, educational attainment, etc., are canvassed under the labour
      force section.

16.   Data collection is carried out by the 13 State offices throughout Malaysia. Staff members
      are thoroughly trained prior to the enumeration and closely monitored by experienced
      supervisory staff.     During training, the staffs are exposed to concepts, operational
      procedures, coding and checking procedures.

17.     The selected households will be visited by the field enumerator. The field enumerator
        explains to the respondent the purpose of the HIS and then proceeds on to interview the
        respondent and the other usual members of the households.

18.     When the questionnaires are returned to the State offices, they are clerically edited and
        coded. Incompleteness and other related problems are then immediately referred back to the
        field enumerators. Upon completion, the manually processed questionnaires are then passed
        to the data entry clerks for data capture and subsequently passed through a very detailed
        computer editing. Inconsistencies, errors, duplicates, missing information, etc. are identified
        for correction. This process will be repeated until no errors are identified.

Income Concepts
19. The definition of income is used in the HIS which covers both money income and income in
     kind and also receipts which are of a recurring nature and accrue to the household or to the
        members of the household regularly at annual or at more frequent intervals.

20.     Household income can be derived from wages and salaries and other receipts from
        employers, commissions, net income from self-employment, income from rent, interest,
        dividends and royalties.     Payments received regularly from pension funds, alimonies,
        scholarships, inheritance or trust funds, remittances and other cash assistance regularly
        received are also included.

21.     Household income in kind includes wage payments in kind, goods and services given free of
        charge to an employee and the value of home produce consumed within the household. The
        estimated net rental value of owner-occupied housing and rent-free housing are also treated
        as income.

Reference period
22. The reference period for income in the Household Income Survey (HIS) was the last twelve
    months. In other words, supposing the survey month was June 1999, then income received
      was recorded for the period of 1st June 1998 until 31 May 1999. The survey rounds for HIS
      1999 were June, September and November 1999. Spreading the survey over three rounds
      ensured that seasonal variation of income is accounted for.

Evaluation of the data
23. Data obtained from surveys or research based on sample selection has a probability of
        encountering two types of errors or commonly referred to as mistakes. These errors are the
        sampling and non-sampling errors.

Sampling Errors
24. Sampling errors is a result of estimating data based on probability sampling, not on census.
      Statistically, these errors are referred to as relative standard errors, denoted by RSE and are
      given in percentage. It is an indication of the precision of the parameter understudy. In
      other words, it provides a picture on the extent of variation with other sample-based
      estimates. Subsequently, the accepted level of reliability of the data has been determined by
      placing an RSE of less than 5 and 10 per cent at National and State level respectively.

25.   For HIS 1997, the mean monthly gross household income for Malaysia was RM 2,607 with
      an RSE of 1.4 percent. In other words, the standard error (SE) is approximately RM 36.00.
      With the assumption that the household income is normally distributed (although untrue in
      practice), the confidence interval for the estimated mean income can be calculated. Based
      on a 95% confidence level (α=0.05), the mean household income was found to be in the
      range of RM 2,535 – RM 2,679 monthly.

Non-sampling Error
26.   To ensure a high quality data, several steps were taken to keep non-sampling error to a
      minimum. Unlike sampling error, this error cannot be measured and can only be overcome
       through several administrative procedures. This error can arise through incomplete survey
       coverage, weakness in the frame, feedback error, non-response error, error during
       processing such as during editing, coding and data capture.

27.   Response error can occur due to differences and difficulty in interpreting questions, be it on
      the part of the enumerator or respondent. To minimise this error, intensive training was
      conducted for the enumerator as well as the supervisors. In addition, random checks were
      carried out on households which were already canvassed by the enumerator to ensure the
      validity of the information recorded. To ensure the completeness of the survey coverage, the
      sampling frame is frequently updated and the selection of living quarters done after the
      household listing exercise.

28.   For non-response errors such as refusal or no one at home, respondent substitution is done at
      the office, based on equivalent ethnic group, strata and household size. For HIS 1999, the
      response rate was 97 percent.

29.   To obtain respondents’ co-operation, a full scale publicity was launched through the
      electronic and print media. On editing and processing errors, several consistency checks
      were created either manually or computer aided, to ensure the generation of quality and
      acceptable data. By practising the various steps mentioned, it is believed that the non-
      sampling error for HIS 1999 is kept to a minimum.
30.     Apart from that, the accuracy of the HIS data can also be checked against the estimated
        expenditure data of the National Accounts through the private final consumption
        expenditure component. However, this cannot be a straightforward comparison due to
        differences in the income and expenditure concepts used in both system. For HIS, gross
        income concept is used, whereby it include the income component of current transfers,
        savings and investment. This is because, a portion of the gross household income that is
        unutilised, will be saved or invested whilst another portion will be spent on income tax,
        transfers to other household, contribution to social security schemes and other current
        transfer payments.        However, for the National Accounts, this estimate includes the
        expenditure of institutional population and unincorporated business.

      Comparison between HIS Income and Private Consumption Expenditure of National Accounts
        Year                      Private Final                            Gross Annual
                                  Consumption Expenditure                   Household Income               Ratio Income/
                                  (National Accounts)1                      (HIS)                     Expenditure
                                  (RM million)                              (RM million)                   (%)
      1997                   127,783                                          128,593                         101
      1999...............           123,257                                   120,387                         98
            Annual National Product and Expenditure Account, 1987 – 1999, Department of Statistics, 2000.

31.     By       definition2,    gross    income       is   higher     compared         to   expenditure    and     usually   the
        income/expenditure ratio should be within 80-120 percent, hence it is an indication that the
        HIS data can be accepted. It was found that the ratio between household income from HIS
        and private consumption expenditure (National Accounts) is about 98-101 percent. As such,
        it can be concluded that the income reported from the survey is of good quality.

Poverty Concept
32. When data pertaining to household income has been established, the identification of
        households categorised as poor can be determined. In the Malaysian context, two poverty
        measurements are used, namely absolute poverty and relative poverty. Absolute poverty
        approach is based on the concept of a minimum standard of living, i.e. poverty line income
        (PLI). The PLI is the estimated costs of the bundle of goods necessary to ensure that basic
        consumption requirement are met.             Hence, the expenditure on food and non-food
        components is being used in the calculation of the PLI.

Source: The Developing Economies, XXXV – 1(March 1997): 68 – 95

33.   The minimum expenditure for food component is based on daily calorie needs of 9,910 per
      day for a family of five. A Technical Committee at the Inter Agencies Planning Group
      (IAPG) level was formed to determine the various types of food components for the three
      regions of Malaysia i.e. Peninsular Malaysia, Sabah and Sarawak. The PLI on food
      components is then calculated by using the retail price of the selected foods with the
      calorific requirements of each item. At the same time, expenditure on non-food components
      such as on housing, fuel and energy, transportation, education and health is obtained from
      the HES.

Determination of PLI
34. The calculation of the PLI was first prepared for Peninsular Malaysia in 1977 and in 1978
     for Sabah and Sarawak and for subsequent years, was updated using the Consumer Price
      Index (CPI) to reflect the PLI for the respective years. An example of the Regional PLI is as

         Monthly Poverty Line Income by year and region
          Region                            1989        1992    1995    1997    1999
          Monthly Gross (Poor) PLI
          Peninsular Malaysia                 370         405     425     454     510
          Sabah                               544         582     601     626     685
          Sarawak                             452         492     516     535     584

          Monthly Gross (Hardcore) PLI
          Peninsular Malaysia                 185         203     213     227     255
          Sabah                               272         291     301     313     343
          Sarawak                             226         246     258     268     292

35.   Once the PLI has been identified, the next step would be to identify households that are in
      poverty and the magnitude of poverty that Malaysia is experiencing. Among the measures
      used in determining the magnitude of absolute poverty in Malaysia are the incidence of
      poverty (proportion of household with income below PLI), the incidence of hardcore poverty
      (proportion of household with income below 50% of PLI), Poverty Gap and Pa Index.
      Besides this, measures of central tendencies such as the mean and median of the poor by
      various socio-economic characteristics are also calculated. In the context of relative poverty,
      statistical indices such as the mean, median, Gini coefficient and income share by deciles
      were used.

Networking between Producers and Users of Poverty Data
36. While DOS is entrusted with the collection and production of poverty data based on
    International guidelines with local adaptation, the Economic Planning Unit (EPU) of the

     Prime Minister’s Department plays the role of main User. With regard to the planning,
     implementation and evaluation of poverty programs, the functions have been allocated in the
     following manner:

   ² the formulation of poverty reduction programmes in Malaysia has been led and
     coordinated by the Economic Planning Unit (EPU) of the Prime Minister’s Department as
       an integral part of the process of formulating the nation’s development plans. The process
       has been a multi-tiered one, involving village leaders at the bottom, the federal, state, and
       local levels of the public machinery, the Cabinet and at the very top, Parliament. At the
       same time, it has also been broad-based, incorpor ating not only inputs from the parties
       mentioned above but also those of universities, NGOs and individual experts. The main
       focus of the exercise has been on the achievements and shortcomings of existing
       programmes and the identification of remedial and new programmes.

   ² the implementation of poverty programmes has been carried by various Ministries through
     their operating agencies at the state and sub-state levels based on approved policies and
       programmes as well as supporting implementation guidelines.

   ² the evaluation of the progress made in poverty reduction has been carried out mainly
     through periodic Household Income Surveys (HIS) conducted by the Department of
       Statistics which provided information on household income and incidence of poverty. In
       addition, EPU and other agencies have also commissioned studies on more specific aspects
       of poverty as the need arose.

Implications of HIS Towards National Development Policies
37. The reliability, availability and the ability of an Income Survey to explain a socio-economic
     phenomenon, either directly or indirectly, influences both positively and negatively, the
     process of national policy formulation. Generally, the effect of consuming the findings of a
     systematic survey can be viewed from three important policy implications, i.e.:

     a) Performance and Analysis of the Policy
        The main reason quoted when requesting for systematic and reliable information
         pertaining to income is because it’s required for the evaluation of an implemented policy.
         In this aspect, the income data has been widely consumed for the purpose of evaluating
         the National Development Policies, especially in achieving poverty eradication goals and
         income distribution. Since the policy itself undergoes changes and adaptations through
         its implementation, there is a need that requires the methodology applied in deriving the
         data to be as versatile, so as to facilitate the accommodation of relevant additional
     b) Allocation of National Financial Resources
        Poverty eradication and the modernisation of rural areas have and continuously will
        require enormous public financial resources. Among the indicators are the level of mean
         income and income distribution by region, state and ethnicity. It has been the norm that
         additional data at disaggregated levels is very much needed to ensure the direction of a
         policy. In this aspect, the ability and reliability of the information on hand plays an
         important role.

     c) Selection and determination of priorities
        A clearly defined policy is no indication that strategy and program priorities are equally
        transparent. Consequently, survey results should also be sensitive to the selection of
         priorities. For example, a clearly defined program to overcome poverty might not have a
         clear-cut priority. That is why the ability of a survey to furnish reliable estimates is
         required as benchmark data before such priorities can be established.

Capacity building needs
38. The HIS is carried out twice during each 5-year term of the National Development
     Policy(NDP) and over the years, changes have to been made to the questionnaires to ensure
     that the information collected would reflect the current socio-economic changes experienced
     by the population as well as cross-classifying questions to ensure the validity of data
     collected. It should also be able to furnish the planners with the relevant data needed in
     formulating and strengthening poverty-reduction programmes. As such, capacity building of
     related manpower is needed at each process of income and poverty data collection,
     production and analysis, to ensure that the above-mentioned requirements are met. Training
     would also be required for the analysis of sample-based small area statistics and also specific
     model building, with the aid of PC-based techniques.

Availability of Data
39. The data is made available through the NDP reports as well as its Mid-term reports. Another
     source would be the First Outline Perspective Plan 1971 – 1990, the Second Outline
     Perspective Plan 1991 – 2000 and the Third Outline Perspective Plan 2001 - 2010. These
     can be made available from the Government Printers. However, those published data form
     only a small portion of HIS data that has been compiled by DOS and as such is insufficient
     for researchers/academicians to pursue studies pertaining to income and poverty. As is
     required by the system, comprehensive data on the subject is published by DOS for internal
     consumption of its sole user, EPU. DOS also make available the raw data to EPU for further
     analysis. So far, this raw data is not permitted for public use.

40. This paper has briefly discussed the main sources of data that has been utilised by the
     Government in poverty monitoring and some poverty measurements. It has also enlightened
     on one hand, the role of DOS in the compilation of income data and on the other hand, EPU
     as the sole user of the said data. It has also been emphasised that Malaysia has a clear and
     consistent development plan in poverty eradication and raising the income among the


To top