Official name of the survey:
INKOMSTFÖRDELNINGSUNDERSOKNINGEN 1987(HINK 1981)
(Income Distribution Survey 1981)

Administrative unit responsible for survey:

Funding for the survey is provided for in the Swedish national budget. The data are stored at, and can be obtained from:

Copies of the original codebook and other documentation can also be obtained from the above address.

For additional information to that provided here on various aspects of the survey the following people (all at the above address) can be contacted:

Sample Design:

Jan Eriksson
Phone: 019-1766753

Data collection:

Marianne Andersson
Phone: 019-176573

Income Information

Kjell Jansson
Phone: 019-176540

The Income Distribution Survey is a multipurpose survey. The main purpose is to measure the distribution of income and wealth among the non-institutionalized family units as well as among individuals. Other objectives are to show the composition of income and wealth and show the trends in the composition and distribution of income and wealth. Data in the survey is also used for modelling income and distribution of income.

The Income Distribution Survey was first carried out in 1973 and it has since then been undertaken every year. In 1982-83 there was a general revision of the survey. The Income Distribution Survey is now available on microdata for the years 1975, 1978 and 1980 and all consecutive years.

There are three main origins of the data. First there is an interview with one of the members in the family, second, data is collected from the income tax return and third, data is also collected from several administrative sources.

The process of data collection for the 1981 survey started in January 1982 and was completed in March 1982.

The survey is published as official statistics every year. For the 1981 survey there were three Statistical Reports. These are mentioned in the bibliography in the Appendix together with other important publications based on the HINK survey.

The HINK 1981 microdata have been made available to the LIS project under the condition that LIS protects the identity of all individuals in the survey. In general, permission to use the microdata is not given to other people, organisations or researchers. In some cases researchers can use the microdata under the same conditions as the LIS project.

LIS data are available for 1981 and 1987.

B POPULATION AND SAMPLE SIZE, SAMPLING METHODS back

The survey population is all people of age 18 or older not living in institutions. People living in Sweden for more than six months in 1981 belong to the survey population. Persons of age 18 or more but who have lived less than 6 months in Sweden are excluded (for instance people who died before July or people who have emigrated).

Starting with the sampling person families are constructed.

People who belong less than six months to the family where the sampling person lives, or who divorced or died before July, and people who married the sampling person after July, are excluded from the family. Children of age 17 or less always belong to the family irrespective of the six month criteria. Children of age 18 or more constitute a family-unit by themselves if he or she was selected as a sampling unit.

The sampling frame for the HINK 1981 is the "Total Population Register" (TPR), which contains all individuals 18 years of age or older. The sampling frame includes all people living and registered in Sweden in the middle of 1981. Consequently, the sampling frame includes people living in institutions. Later in the survey these people are excluded since they do not belong to the population desired.

1. The Sample Design

The 1981 sample design is a two stage stratified sample design. In the first step different types of persons are selected, and in the second step they are selected by their 1979 income level.

Table 1
Sample and sample frame for panel 7 of the Income Distribution Survey 1981

Stratum Gross family Sample Population

income

Table 2
Sample and sampleframe for panel 8 in the Income Distribution Survey 1981

Stratum Gross family Sample Population

The gross sample consisted of 10,046 persons. 433 of these did not belong to the targeted population. The net sample was 9,625. Data for all these sampling units are available in the LIS database.

Data for income, transfers and taxes are available for all 9,625 sampling units. Individuals in prison and other institutions were not eligible to be interviewed. No further information is available concerning these people.

All geographic areas of the country were included in the sample design .

Estimates of sampling variability (standard errors) have been computed for this survey. Available estimates are reported in official statistical reports.

C. MEASURES OF DATA QUALITY back

1. Item and Group Nonresponse Rates

Data from the interview are not available for all units. The nonresponse rate is 15.3%. Families where all people were of age 71 or more were excluded from the interview in 1981. At this time there was no interest in knowing about the living conditions or employment for older people. Income data were nevertheless collected.

For those sampling units where an interview was sought the overall nonresponse rate was 13.2%, which represents a total of 1,310 cases. Excluding older people, nonresponse was predominantly amongst single males and younger persons.

2. Reporting and Under-Reporting and Income Data Quality

Income data in the survey are collected from income tax return forms and administrative sources.
Weighted macrolevel figures from the survey are compared with macro data from different registers.
Disposable income is compared with National Accounts. Comparisons have also been done for wealth and tax data.
Great effort have been put on checks for consistency errors. For example, people who work should also have income from that work. Families with children will also receive a child allowance. There was no need to impute income data after these checks.
For data other than income data there are sometimes missing values. In some cases there will be an imputed value and in other cases there will be none. If a person has income from work of a certain level we also impute working time and categorize it as either short part-time, long part-time or full time work.

During the process of checking data from our three sources -- interview, tax return form and administrative registers -- inconsistencies are examined and some corrections are made.

Household composition

Data in the sample are from 1 July 1981. Households in the survey consist of one adult or a married couple with or without children less than 18 years. In the interview people are asked if they live together even if they are not married. The purpose is to find cohabiting couples, irrespective of whether they are married or not. In the case of non-response there is only information if people are married or not. In these cases the survey will consider a person who is not married but cohabiting as a a single person. Single persons are overestimated and those who are cohabiting are underestimated.

If a child of age 18 or more is sampled and is living with his or her parents, he/she is considered a single person. In the same way, if a parent has been sampled and a child of age 18 or more is living at home then the child will not be considered as belonging to the family.

D. DATA COLLECTION AND ACQUISITION back

1. Method of Data Collection

In 1981 data were collected by mail. For people who did not answer by mail there was a telephone interview. In some cases an interviewer visited the household. Participation in the survey was voluntary.

The respondent could deny us permission to obtain data from other sources. If so the sampling unit would be a non-response. The SCB assured the respondents that their privacy would be maintained.

2. Structure of the Survey Instrument

Some questions considered the whole family as a unit and others considered all members as an investigation unit. Each member was not required to answer the questions for him/herself. Respondents were not asked to consult or use documents, such as pay records. Income data was collected from other sources.

E. WEIGHTING PROCEDURES back

In the HINK81 survey weights have been assigned to each sample case.

Although the survey frame was comprised of individuals, the objective of the survey was to estimate variables for both households and individuals. The weights were assigned in order to compensate for sampling errors which occurred because of the different units used in the sampling frame and the analytic unit of the survey. These weighting procedures helped adjust for missing data or other nonsampling errors in the income data.

The sum of the weights is equal to the total number of units in the sampling frame excluding people in institutions. The sampling frame included all people listed in the Total Population Register. Since the purpose is not to study people living in institutions, the weights of the sampling units are adjusted for this.

The sum of the weights will therefore be a little less than the sampling frame.

There is another problem with estimating income for the units. The period of investigation is the whole year. People living more than 6 but less than 12 months are within the sample. But these people do not have full year income. The same problem arises when people immigrate to Sweden in the period of January to June, where the income measure is not for the full year.

F. DETERMINATION OF SURVEY UNIT MEMBERSHIP back 

The basic units of aggregation which were used in this survey were individuals and households.

These were further aggregated into groups based on sex, age, profession and income.

Households were defined as either one adult or two adults of different sex (equal or more than 18 years old) with or without children (equal or less than 17 years old). If the household contained children older than 17 years, these children were excluded from the household.

Children away from home and attending a school, or husbands away from home at work are included as members.

The overall sampling units cannot be divided into subgroups in those situations where several families share the same housing unit.

All households members can be identified in the survey.

G. CHILDREN AND SPOUSES back

In this survey children are defined as all persons under the age of 18 who have never been married.

Only children of the unit head (or spouse) are registered as children within the household. A child of someone other than a household member or of someone living outside the household is not included in the survey.

In this survey spouses are defined as people who are cohabiting, as well as people who are married.

H. AVAILABILITY OF BASIC SOCIAL AND DEMOGRAPHIC INFORMATION back

In Table .1 are summarized the basic social and demographic information which is available in the HINK 87.

Table .1:
Basic Social and Demographic Information

I. AVAILABILITY OF LABOUR MARKET INFORMATION back

Labour force status information is available for this survey. In the interview all persons were included. Labour force status information is collected from different administrative registers. For non-response some summarized data are coded as follows:

Table .2:
Labor Market Information

J. AVAILABILITY OF GEOGRAPHIC INFORMATION back

The geographic codes in the LIS data are H-regions

1. Stockholm and surroundings
8. Göteborg(Gothenburg) and surroundings
9. Malm" and surroundings
3. Cities and surroundings with more than 90,000 inhabitants
4. -"- 27,000 - 90,000 and with 300,000 inhabitants within a distance of 100 kilometres
5. As in 4. but less than 300,000 within a distance of 100 kilometres.
6. Cities < 27,000 inhabitants.

 

K. SOURCES AND AMOUNTS OF CASH INCOME back

Sources and amounts of income are recorded for an annual period. The amounts recorded were the actual amounts received. Income sources and amounts are recorded for individuals, not households.

L. TAXES back

Taxation information is available for all persons in this survey

M. BIBLIOGRAPHY OF MAIN PUBLICATIONS back