Official name of the survey :
1992-93 Family Expenditure Survey.

This survey, conducted by the Israeli Central Bureau of Statistics (CBS) and funded by the Government, collected information on the components of the urban households' budgets. The findings of the Survey are used in order to determine the "weights" for the Consumer Price Index Basket. The survey is used mainly by the Central Bureau of Statistics and by the National Insurance Institute which is responsible for the connection with the Luxembourg Income Study. Its address is :

The field work was conducted over a one year period, from June 1992 to June 1993.

B. POPULATION AND SAMPLE SIZE, SAMPLING METHODS back   

1. The Sample Design and Sampling Frame

The sample was a random sample of the units within the sampling frame. The sampling was carried out in two stages. In the first stage, the localities (each over 2,000 inhabitants, thus excluding rural localities) were sampled and in the second, dwellings from the property tax file were sampled on a random basis for every locality that was selected for the sample. Each dwelling was given an equal sampling probability to be included in the Survey and all households living in the sampled dwellings were included in the sample population. In each locality the dwellings sample was randomly selected (systematically). In total 90.5 % of all dwellings were included.

Households from urban localities with less than 2.000 inhabitants have been excluded from the sample, as well as non jewish households of East Jerusalem, temporary residents, tourists, diplomats and the institutionalized population. Residents of student dormitories, absorption centers in urban localities, housing facilities for the elderly have been included. Residents of Israel and their households who were abroad for less than three have also been included. Persons who were in compulsory service in the army and single persons serving at a ship have been registered at their permanent address and have been included in the survey.

The sample included a total of 7,446 dwellings in 126 localities. Each household in the survey population had the same probability (1:180) of being selected for the sample. An "Inflating coefficient" was calculated for each household in the final sample, which would express the number of households in the population represented by it. The sample was selected from a total population of approximately 1,300,000 households.  

C. MEASURES OF DATA QUALITY back  

1. Response Rates

In total 5,212 households (weighted 1,276,662) have actually been surveyed. This sample has been made available to LIS. There were 1,530 sample units in the sampling frame were not interviewed and 127 were elimiated during editing. In total 1657 households did not participate:

D. DATA COLLECTION AND ACQUISITION back  

Participation in the survey is mandatory but in fact there are no compulsory measures applied. Privacy of respondents has been explicitly assured.

Questionnaire.

Three methods were combined:

1. Daily records in a diary for consumption;
2. Questionnaire for detailed information relating to the period preceeding the investigation month (filled out by the interviewer);
3. Questionnaire with basic information for each persons detrmined as belonging to the household (filled out by the interviewer).

One household member could answer for the household and the other members.

Respondents were asked for salary slip and other pay records in order to fill in questionnaires.

E. WEIGHTING PROCEDURES back  

To obtain estimates for the entire survey population an 'inflating co-efficient" was calculated for each household in the final sample, which express the number of households in the population represented by it. The weighting id not help adjust for missing data or other sampling errors related to income.

The sampling frame includes dwellings that do not belong to the survey population like households of diplomats, dwellings for business, empty dwelling, etc. ... . Thus, the sum of the survey weights do not equal to the total number of units in the sampling frame.

F. DETERMINATION OF SURVEY UNIT MEMBERSHIP back   

Households have been defined as a group of people who live in the same dwelling most days of the week (usually sleeping at the dwelling at least four days a week) and who share a common food budget. In most cases, the definition resembles a "family" in its ususal meaning . In special cases persons who are serving compulsory service in the army and do not have a civilian address are included. Also, single persons who are serving on a ship or flight crew, married persons who are away because of work or studies, children up to the age of 18 who study (and live) at institutions.

The overall sampling unit can be divided into subgroups in those situations where several families share the same housing unit. There are 182 households residing in 91 dwellings with two households, and 57 households residing in 18 dwellings with three households or more. Within the household, it is not always possible to identify nuclear families. If two or more families livein the same dwelling and share the same budget we cannot identify them independently. Furthermore, it is possible to specifically identify every individual in the household. These individuals can be identified by a code indicating the relation to the household head.

G. CHILDREN AND SPOUSES  back 

Children have been defined as persons under age 18 that are not a head or spouse of a household. Spouses are fefined by the household replier, independent of marital status. Parents of the household head or his spouse have been specifically defined.

H. AVAILABILITY OF BASIC SOCIAL AND DEMOGRAPHIC INFORMATION back  

I. AVAILABILITY OF LABOUR MARKET INFORMATION back   

J. AVAILABILITY OF GEOGRAPHIC INFORMATION back   

Only urban localities are included in the survey (with 2,000 inhabitants or more). Rural localities (less then 2,000 inhabitants) have been excluded.  

K. SOURCES AND AMOUNTS OF CASH INCOME back   

Household members reported in the questionnaire their income from all sources during the Survey Year. To this was added income in kind which was imputed to various items (mainly alternative interest on capital invested in housing and in vehicles).

Households only reported their gross income. Taxes were imputed afterwards. The income of individuals who did not report their income from work was generally imputed manually, according to the income of "similar cases" from the survey itself.

L. TAXES back  

Income taxes were imputed mechanically for every employed household member and pension recipient as well as for income (non work- related) subject to income taxes.

Payroll taxes (social insurance contributions) were imputed for every employed household member according to the table of rates of the National Insurance Institute.

Municipal Property Taxes have also been imputed.

M. BIBLIOGRAPHY OF MAIN PUBLICATIONS back  

Achdut, L., Awad, Y. and S. Yaakov, Reforms in the Direct Tax System in 1985-1995: Distribution of Tax Burden and Net Income from Work, Dec. 1994, Research and Planning Administration, National Insurance Institute, Israel.

Achdut, L. & Y. Awad, Tax Reforms and Tax progressivity: The Israeli Experience 1986-1995, Sept. 1995, Research and Planning Administration, National Insurance Institute, Israel.

Momi, D. @ R. Alon, Measuring the Mean Marginal Tax on the Work Income, Sept. 1995, Department of Research, Bank of Israel.

Yosef, G., The Effects of Direct and Indirect Taxes and Transfers on Inequality in Israel, Research Unit of The School of Business Administration, The college of Academic Managment Studies, Israel.