Main contents start here
Q&A about the Family Income and Expenditure Survey(Answers)
The Family Income and Expenditure Survey (FIES) is a sample survey conducted for the Statistics Bureau each month. It aims to grasp actual conditions of income and expenditures of household in order to provide bases for social and economic policies. The survey covers households excluding one-person student households, etc. According to the 2010 Population Census, they amount to 50.18 million households and 96.6 % of all households.
In the FIES, sample households are selected based on statistical methodologies so that they represent all households in the whole country. More concretely, about 9,000 households are sampled through a three-stage stratified sampling method (a). A sample household records daily income and expenditures on the family account books for six months (three months for one-person households).
(a) The three-stage stratified sampling method is a method in which samples are selected through the following three stages.
In the FIES, at the primary stage, cities, towns and villages (municipalities in short) are grouped into 168 strata according to certain criteria, and one survey municipality is selected in each stratum, 168 municipalities being selected in total.
At the secondary stage, survey districts are selected randomly in each survey municipality.
At the tertiary stage, survey households are selected using random number tables from the list of all households in each survey district.
The results of the FIES tell us how much and on what kinds of goods and services the households spend their income, and how household expenditure patterns are influenced by the differences of household characteristics, e.g. annual income level, number of household members, household head's age and occupation.
Moreover, the monthly results of the FIES provide important data to deduce consumer spending, which is the major factor indicating business conditions.
As to workers' households and no-occupation households, the results about income are also available.
To grasp the actual conditions of life of the people, the results of the FIES are widely used not only by central and local governments but also by private companies, research institutes and labour unions. They are utilized as basic data and indicators for:
- Analysis for economic and social policy planning;
- Monitoring economic trends and business conditions;
- Determining wage level in central and local governments, private companies, etc.;
- Forecasting consumer demand for goods and services;
- Estimation of private final expenditure in the National Accounts of Japan (SNA); and
- Determining the Consumer Price Index (CPI) item categories and computing the weights assigned to the categories for aggregation.
For example, ministries and agencies use the results of the FIES for:
- Economic analysis (Annual Report on the Japanese Economy and Public Finance, Monthly Economic Report)
- Analysis on conditions of life of the people (White Paper on the National Lifestyle)
- Estimation of the SNA (Quarterly Estimates of GDP)
- Planning of policy visions and future prospects of economy
- Basic material for consumer administration;
- Indexes of Business Conditions
Ministry of Internal Affairs and Communications
- WHITE PAPER Information and Communications in Japan
Ministry of Finance Japan
- Basic material for determining various tax rates and tax credits such as deductions from income
Ministry of Health, Labour and Welfare
- Determination of standards for providing public assistance
- Review of various pension schemes
- Determination of social welfare service charges such as medical fees
- Analysis on labour economics (White Paper on the Labour Economy);
- Basic material for coordinating labour problems
Ministry of Agriculture, Forestry and Fisheries
- Comparison of living standards between farm and non-farm households
- Analysis of supply and demand on food output
Ministry of Economy, Trade and Industry
- Economic analysis (White Paper on International Economy and Trade)
- Demand analysis for small and medium enterprises (White Paper on Small and Medium Enterprises in Japan)
- Basic material for formulation of industrial visions
National Personnel Authority
- Calculation of standard cost of living
- Revision of remuneration levels
- Basic material for calculating allowances such as cold-district allowance
- Basic material for Prefectural Economic Accounts
- Revision of wage levels
Universities and research institutes
- Econometric analysis on household income and expenditures
- Analysis of actual consumption and forecast of future demand by region, consumer's attribute and commodity
- Revision of wage levels
4. Although the FIES does not cover one-person households and households engaged in agriculture, forestry and fisheries, are the results of the FIES able to represent household consumption of entire Japan ?
The current FIES covers one-person households and the households engaged in agriculture, forestry and fisheries. The survey unit is the household excluding one-person student households, etc. in the entire area of Japan, which amount to 50.18 million households accounting for about 96.6 % of all households according to the 2010 Population Census.
For many years, the coverage of the FIES had been two-or-more-person households excluding those engaged in agriculture, forestry and fisheries. However, households engaged in agriculture, forestry and fisheries have been included in the coverage of the FIES and the results both including them and not including them have been tabulated and released since January 2000.
As to one-person households, the Income and Expenditure Survey for One-person Households was initiated in 1995. It aimed to compliment the FIES, and yearly average figures released from 1995 to 2001. Half-yearly averages released also from the first half of 1998 to 2001. The sample was extended to one-person households living in boarding houses and dormitories and compilation of quarterly average was started in January 2000. Furthermore, we estimated the income and expenditures of total households, which included one-person households and households engaged in agriculture, forestry and fisheries, with the results of both the FIES and the Income and Expenditure Survey for One-person Households on a quarterly basis from 2000 to 2001.
In January 2002, the Income and Expenditure Survey for One-person Households was combined with the FIES. These results are connected with the results of one-person households and total households in the FIES.
In this way, the Statistics Bureau has been improving the coverage of statistics concerning household consumption. Users are recommended to refer to above-mentioned statistics to grasp the movement of household consumption.
5. Are the sample households of the FIES biased towards peculiar households such as government officials' ?
The FIES sample households are not biased too much towards certain household groups. Based on statistical methodologies, the sample households of the FIES are selected randomly. If a sample household could not cooperate on the survey owing to unavoidable circumstances, the substitute was selected randomly, too.
Besides the FIES, the number of "government officials' households"(b) can be obtained from the Labour Force Survey that investigates the number of persons engaged in work, unemployed persons, etc. every month. The Labour Force Survey showed that "government officials' households" for 2013 numbered to 2.11 million accounting for 5.9% of all two-or-more-person households (35.62 million), while that of the FIES was 6.7%. The slight difference between the two figures is considered to be attributable to the conceptual differences between the two surveys such as the definition of household head: the main earner among household members in the FIES while a person representing the household in the Labour Force Survey.
(b) "Government officials' households" are those whose household head is employed by ministries and agencies, courts of justice, etc. of the central government, municipal offices, etc. of local governments, as well as in hospitals, schools and enterprises operated by central and local governments. Public corporations, and public financial institutions and foreign governments' facilities stationed in Japan are excluded here.
Quantities purchased by households for foods are basic material for food policies. For example, these data are utilized to look into demand and supply of foods.
In addition, quantities are used for calculation of average purchase prices, which are the quotient of expenditures divided by quantities. Average purchase prices are utilized for analyzing influences of consumer behavior on consumer spending. For example, according to the 2009 results as to meat, average purchase prices of beef, pork and chicken all fell down from the previous year. A similar tendency was appeared in some items of other foods and clothes.
As shown above, quantities consumed by households are important in government policy planning as well as in analysis on consumer behavior.
7. How does the Statistics Bureau conduct the FIES accurately under such circumstances as the number of households which keep family account books has been decreasing ?
In the FIES, the sample households are selected from all households regardless of having a custom of keeping family account books or not. The sample households are provided the family account books specially designed for the FIES, visited by the investigators and explained in detail how to record their daily income and disbursements.
They have only need to record simply their daily income and disbursements on their account books. No requirement of classification or summation, this recording method saves them trouble and improves the accuracy of the survey.
Moreover, regarding workers' households and no-occupation households, the sample households can easily find and correct their errors on daily recordings by checking the entry of "today's balance" with the cash they have in hand.
Therefore, even if the number of households which keep family account books is decreasing, we can continue to conduct the FIES accurately.
8. How do you make it possible to survey income and expenditures of whole household when it includes the grown-up children of household head ?
In the FIES, the sample households are required to record every income and disbursement of all the household members (Incomes are only for workers' households and no-occupation households.).
The keeper of the FIES's family account book needs to record all incomes and expenditures of his or her household members, e.g. those of the grown-up children who are employed and the students who have part-time jobs. When the household members spend their own earnings not for family expenses but for themselves and the keeper of the family account book hard to know details of their spending, he or she records their total earnings as the household income and their spending as "pocket money (of which, detailed uses unknown)".
The FIES is conducted in this way so as to cover all receipts and disbursements by all household members of the sample households.
As to the individual income and expenditure of each household member, we have investigated the details for persons aged 18 years or over except for family-account-book keepers in the National Survey of Family Income and Expenditure. We will examine the results and make analyses on this matter.
Tips for understanding the FIES results
9. Why do the data from the FIES show a different trend from department store sales statistics or supermarket sales statistics ?
There are some differences between the FIES and retail sales statistics. For example, the data from the FIES contain all expenditures at every type of business (e.g. department stores, supermarkets, convenience stores, mail-order business). They also include expenditures on services and goods which are usually purchased at places other than department stores and supermarkets; such as school fees, house rents, automobiles, electricity, gas, piped water, gasoline, etc. On the other hand, data from retail sales statistics include purchases by juridical persons which were outside of the scope of the FIES. Therefore, you should take these differences into consideration when you'd like to compare the FIES with retail sales statistics.
10. How accurate are the data from the FIES with only 8,000 samples of two-or-more-person households ?
In sample surveys, such as the FIES, samples (e.g. survey households) are selected randomly by statistical methods to represent the entire population (e.g. all households). However, there is still a "sampling error", difference between the survey estimate and the true value. Magnitude of "sampling error" is measured by "a rate of standard error". In many cases, the sampling error is within the limits of its rate of standard error.
Rates of standard errors of the FIES are regularly estimated by the three-stage stratified sampling method and presented in the Annual Report on the FIES. As to the average consumption expenditures of two-or-more-person households in 2017, the monthly rate of standard error was varied from 1.1% to 1.5% and the yearly rate of standard error was 0.4% .
The rates of standard error(%)(Tow-or-more-person Households)
|Year and month||Consumption
|Num. of tabulated
Note that the more the number of samples decreases, the larger the standard error becomes: figures shown for particular households and less-frequently purchased items are subject to larger sampling variability. In case of analyzing data of such categories, it is better to reduce standard errors to use the mean values of several months, or to use added figures of similar categories.
11. Why do the data from the FIES show a different trend from the private final consumption expenditure in the Quarterly Estimates of GDP (QE) of the National Accounts of Japan ?
The National Accounts of Japan (SNA) are macroeconomic statistics and compiled on the basis of the System of National Accounts, which is the international standards recommended by the United Nations. The private final consumption expenditure is not identical with the consumption expenditures in the FIES in concept and scope. Major differences are as follows:
(1) The FIES grasp the consumption expenditures as a mean amount per household, while the private final consumption expenditure in the SNA is estimated as the total amount of consumption expenditures of the whole country.
(2) The consumption expenditures in the FIES include rent actually paid by households, while the private final consumption expenditure in the SNA includes imputed rent, the estimated value that an owner of his own dwelling would pay for the equivalent rental services.
The FIES are processed for the estimation of the private final consumption expenditure in QE as follows:
(3) The consumption expenditures in the FIES include transfers to other households as gifts and remittance, while these are excluded in QE.
(4) For housing expenditures, the consumption expenditures in the FIES include repairs and maintenance expenses, while they are excluded in QE because they are contained in imputed rent.
(5) In the QE, expenditure on automobiles is estimated based on the number of newly registered cars instead of the data from FIES.
Moreover, the private final consumption expenditure in QE has been estimated by calculating the weighted average of the demand-side estimates, which is estimated by the FIES and other statistics, and supply-side estimates since the estimation method of QE was revised in August 2002. Users need to consider these differences and characteristics when they use these statistics for grasping the household consumption.
12. Why does the income in the FIES show a different trend from the wages in the Monthly Labour Survey ?
The income in the FIES and the wages in the Monthly Labour Survey (MLS) are different in scope, while the former includes not only wages and salaries, but also returns of rented houses, public pensions, gifts in money, etc.
In addition, wages and salaries in the FIES include wages from short-term employment and side-job, while total cash earnings in the MLS is the average of wages and salaries per regular employee working in an establishment.
Moreover, the MLS does not cover establishments with less than 5 regular employees and government employees. Therefore, the income in the FIES shows a different trend from the wages in the MLS.
13. Is it hard for the FIES users to grasp the timely consumption trend because the data are released later than other statistics ?
The FIES monthly results are released as early as the results of other major economic statistics. For instance, the monthly results of the two-or-more-person households in the FIES are released at the end of the next month after the survey, as a rule.
The results are widely disseminated through the Internet concurrently with their release.
14. Do the consumption expenditures in the FIES decrease only because the number of household members decreases ?
The number of household members tends to decrease in the long run, partly because of the trend towards the nuclear family. It results in decrease of household consumption, so users should take it into consideration when they use the figures of consumption expenditures as the time series data.
In this connection, we release the index of consumption expenditure level (adjusted by the distribution of household by number of household members ) every month as a reference series. We adjust the differences of the number of household members and days in each month as follows: We calculate the weighted averages of expenditures on the basis of the distribution of households by number of household members at the base year, which is the latest year ending with 0 or 5 and is revised in every 5 years. Then we convert them into the amounts per 30.4 days (equal to 365 days divided by 12) and they are indexed against the value of the base year as 100. The index thus calculated is deflated by the CPI to arrive at the index of consumption expenditure level.
We have been releasing the index of consumption expenditure level (adjusted by the distribution of household by number of household members and age group of household head) since January 2008, which is adjusted also by the differences of the age group of the household head.
15. Is the FIES able to catch up with the recent trends in consumer behavior such as an increase in expenditure on goods and services related to information and technology ?
The FIES can grasp consumer behavior accurately because the sample households need to record all receipts and disbursements on family account books, including purchases of goods and services related to IT.
We improve the methods of tabulation in the FIES to reveal recent trends in consumer behavior. The major improvements from 2000 are as follows:
- The Income and Expenditure Classification was revised. Some categories were divided to set up new categories such as PCs, low-malt beer, mobile phone charges and so on.
- Tabulation of expenditure by day from the 1st day through the last day of each month was started for the purpose of analysis on daily consumer behavior, for example, effects of revisions of taxation and charging systems on expenditures.
- Statistics for the specific household groups such as seniors' household and a married couple without children were improved.
The major improvements from 2002 are as follows:
- Compilation of quarterly average figures for total households was started by combining the results of the FIES with those of the Income and Expenditure Survey for One-person Households in order to complement the FIES statistics that exclude one-person households.
- We started to survey savings and liabilities of households by Savings Schedule because the consumer behavior is influenced by how much savings and liabilities they have. The quarterly results are available.
- "The Survey of Household Economy" was newly conducted in October 2001 in order to capture large expenditure and IT related expenditure more adequately. We have released the composite index of consumption expenditures, which is composed of the results of the FIES and the Survey of Household Economy, since January 2002.