Singapore

Title: Household Expenditure Survey, 1997/98

GENERAL INFORMATION AND BASIC DEFINITIONS

1. Background

Further information is available from Singapore Department of Statistics of 100 High Street-05-01 The Treasury Singapore (179434), phone: 65 6332 7134, fax: 65 6332 7689, e-mail: Yap_Yee_Liong@singstat.gov,sg.
Data have been collected by household survey since 1957. 1 2

2. Purpose and coverage

The following purposes are considered to be very important or of some importance:

Geographic coverage: National without any specific geographic areas excluded.

Population coverage:
The following types of household are included in the data collection:
The following types of household are excluded in the data collection:

Units:
Dwelling units are used in the sample selection and are characterised by:
Data are recorded for the household unit which is characterised by:

Unit members: 3 Usual residents temporarily living away from the dwelling are included, if away continuously for less than 6 months. Visitors (not usual residents) temporarily living in the dwelling are included, if staying continuously for more than 6 months. Domestic staff living in same dwelling/compound are included in the unit. Renters living in same dwelling/compound are included in the unit. Boarders living in same dwelling/compound are included in the unit.

Head of unit:
The concept of head of household/other unit is used in this survey and is characterised by:

3. Reference periods

The time period to which income and/or expenditure statistics refer when released/published is from 11/97 to 10/98. These data are collected in the period from 11//9 to 10/98. This survey is conducted 5-yearly. The statistics are published 5-yearly.

SURVEY METHODOLOGY

4. Sample design

The Primary, Secondary and Ultimate Sampling Units are enumeration area/district, dwelling and address respectively.

Stratification:
Areas/districts were stratified using the following criteria:
Households/Consumption Unit, Income Unit, Family Unit were stratified using the following criteria:

The sampling frames for the Primary Sampling Unit (PSU) and Ultimate Sampling Unit (USU) were a register of addresses and a set of administrative registers (or combination of such registers) respectively. Primary Sampling Units (PSU) were selected using 2-stage stratified design and Ultimate Sampling Units (USU) were selected using probability proportional to size. The sample size was 9000 households or other units. The overall response rate for the survey was 88 percent. Errors/biases were minimized by using an updated sampling frame.4

Enumeration procedure: Enumeration uses a single round survey design in which each reporting unit is enumerated only once. The sample is divided into 26 representative sub-samples. No action is taken to select a smaller set of reporting units for more detailed questioning.

DATA COLLECTION, CONCEPTS, DEFINITIONS AND CLASSIFICATIONS

5. Income data

Income data are collected. Receipts have to be regular and recurring to be considered as income. Income excludes receipts resulting from the sale or reduction of assets and/or from incurring liabilities. Income excludes receipts that are not currently available to the unit.

The following receipts are collected separately:
The following receipts are collected but not separately:
The following receipts are not collected:
The following receipts are classified as income from paid employment:
The following receipts are classified as income from self-employment:
The following receipts are classified as property income:
The following receipts are classified as other employment-related income:
The following receipts are classified as other income but not employment-related:
The following receipts are excluded for other reasons:
The following receipts are collected using the last month as the reference period:
The following receipts are collected using the last 12 months as the reference period:
The following receipts are collected using the 'last payment' and 'period covered':
The reference period for:

Income data were collected separately for each person receiving income. Components of income for an individual were collected directly from each individual or from one person reporting for all individuals. Negative values (business losses) were excluded when computing self-employment income.

6. Expenditure data

Data collection method:
Diaries are used to collect expenditure data and are maintained for two weeks.

Diaries are used to collect expenditure on the following items:

Interviews are used to collect expenditure data with the respondent completing the interview by a mix of recall and using documentation.

Data for the following expenditure items are collected by referring:
Data for the following expenditure items are not collected by interview because they are collected by diary:
The following types of expenditure items are classified as 'Household consumption expenditure':
The following types of expenditure items are classified as 'Actual Household consumption':
The following types of expenditure items are classified as 'Non-consumption household expenditure':
The following types of expenditure items are classified as 'Non-household expenditure':
The following types of expenditure items are classified as 'Investment/liability expenditure':
The following types of expenditure items are excluded for other reasons:
The following types of expenditure items are collected on an 'acquisitions' basis:
The following types of expenditure items are collected on a 'payments' basis:
The following types of expenditure items are collected on some other basis:

Classification: A national classification that is not consistent with COICOP is used for classifying expenditure and has 1000 separate categories at the finest level.

7. Other Data Collection Issues

The following other topics are covered:

Households are requested to indicate whether durable goods are new or second-hand when their acquisition is recorded.

Non-response: There is no substitution for non-response, whether by non-contact or by refusal.

Non-response is reduced using the following measures:4
Reporting errors are reduced using the following measures:

DATA PROCESSING, ANALYSIS AND DISSEMINATION

Answers are not pre-coded on data collection forms, but are coded by office staff.

The responses are edited by the following:

Extreme values are retained without change. 5 In-kind receipts are included in the estimates, but consumption of own production is not. In-kind receipts are valued using: the respondent's estimate.

Treatment of owner-occupied housing: Values of owner-occupied housing are included in the total income/expenditure estimates and are valued using the rental equivalent net of housing costs usually paid by owners.

The following information is used in this valuation:

Treatment of selected groups/values in analysis: Some households/units are excluded from analysis because of incomplete response. Missing values are imputed only on some occasions where administrative records or relevant information is available. Supplementary sources are not used to adjust estimates for under- or over-reporting. Etnic groups are excluded from data analysis.

Weighting:
No weighting factors are used for sampling, non-response or benchmarking.

Sampling errors: Sampling errors are computed for Total Income/Expenditure aggregates only and these sampling errors are published.

Tabulation and Analysis: Statistics are presented showing averages per month and statistics are analysed and tabulated for households and for individuals separately. 6

The following classifications are used for tabulation and analysis of income statistics:
The following classifications are used for tabulation and analysis of expenditure statistics:

Documentation and Dissemination:
Published survey reports: Report on the Household Expenditure Survey 1997/98, 12/2000
Published methodological information: Report on the Household Expenditure Survey 1997/98, 12/2000

Additional statistics (or special tables on request) are available for public use, free of charge and with charges (non-published tables migth be provided with charges based on full cost-recovery).7 Separate tables are published for households with wages/salaries as main source of income. Files of unidentifiable unit data are not available (or available on request) for public use.


(1) Part of the data has been collected in a population census. (2) 5-yearly after 1972/73. (3) In situations where the usual resident is the sole breadwinner living with spouse , or an only child living with elderly parents, he/she will be included even if away continuously more than 6 months. (4) Sufficient training of field interviewers. Random spot checks carried out by field supervisors. (4) Substantial publicity was generated for the survey, as well as prompt dispatch of reminder letters to encourage participation. Annex: Classification of Expenditure items Household Expenditure Survey 1997/1998 (5) Or amended. (6) Statistics for individuals are tabulated only on Investments/Asset Ownership. (7) Ad-hoc writeups would be made and published as information papers/articles.