Personal Consumption and Poverty Indicators

Mapping and Estimation of the Geographical Distribution of Risk of Poverty and Social Exclusion for Small Areas of the Republic of Croatia

Income Approach

DESCRIPTION OF THE PROJECT

The Croatian Bureau of Statistics, in cooperation with the World Bank, conducted the project entitled Mapping and Estimation of Geographical Distribution of Risk of Poverty and Social Exclusion for Small Areas of the Republic of Croatia. This activity is related to an initiative of the European Commission and the World Bank to estimate risk of poverty for small areas (NUTS 2 level, NUTS 3 level, or lower) in all EU Member States.

The main objective of this project is the development of a detailed
geo-referenced database that provides information about the geographic distribution of poverty and social exclusion. This database may be then used to create policies that focus on reducing poverty, promoting social inclusion and regional development. Furthermore, this project provides valuable statistical data that will help in the use of European structural and investment funds and ensure effective allocation of available funds intended for mentioned purposes.

Standard household surveys, Statistics on Income and Living Conditions (EU-SILC) and the Household Budget Survey (HBS), conducted by the Croatian Bureau of Statistics, are main sources of information on poverty and social exclusion. Since these surveys include only a few thousand households in the sample, they are inadequate for assessing poverty and social exclusion for small geographic areas and provide reliable results only at the level of the Republic of Croatia.

Therefore, a project in which the World Bank experts have developed an analytical model for a Small Area Estimation (SEA) as well as the software for calculating these estimates (PovMap software) has been launched. The small area estimation model enables at-risk-of-poverty rate estimates for lower geographical levels and their illustrative presentation in a geographic map form. The model is set up in a way that it combines data from a survey with the Census of Population, Households and Dwellings data that provide coverage for lower geographical levels.

It is important to note that, due to the difference in the approaches used in estimating the results, they are not comparable with the results of other surveys on poverty (the SILC Survey).

Users should note that the at-risk-of-poverty estimates for towns and municipalities as well as districts of the City of Zagreb with less than 15 000 households are less accurate, according to the methodology used in the SAE model (Elbers, Lanjouw and Lanjouw – ELL, 2002). According to the Census 2011 data, there were a total of 548 of these types of towns and municipalities in the Republic of Croatia.

DATA SOURCE

Census of Population, Households and Dwellings

The Census of Population, Households and Dwellings in 2011 was carried out in the Republic of Croatia in the period from 1^st to 28^th April 2011, according to the status on 31^st March 2011 at midnight, which is considered the critical moment of the Census.

Statistics on Income and Living Conditions 2012 (EU-SILC 2012)

The survey collects data on gross and net income of households and all household members, data on educational status of persons, activity status and employment, health care and childcare, data on financial and material status of the household as well as data on other aspects of living conditions of households.

The EU-SILC survey is a sample survey of randomly selected private households. The sample frame used in 2012 for the selection of dwellings occupied by private households was based on the 2001 Census data. The total sample consisted of 8 742 households. The sample selected for 2012 consisted of 5 853 successfully interviewed private households and the response rate at the household level was 73.5%.

The reference period for income data, the most important part of this survey, was a previous calendar year, 2011.

Since this project combines survey data with the Census of Population, Households and Dwellings in 2011, both data sources refer to the same time period.

INCOME APPROACH

The Survey provides for quality information on the household income, but the sample of households interviewed in 2011 is not sufficient for getting reliable data for lower territorial levels.

On the other hand, the Population Census data are comprehensive – full geographical coverage and full coverage of households and population – but do not contain information on the household income. Combined with the Survey data, they facilitate the estimation of wellbeing at the household level and enable the calculation of poverty rates even for lower levels.

Small area estimation model, which is used for calculation, overcomes these limitations by making use of both data sources’ advantages.

The calculation according to income approach is based on the concept of relative poverty, which takes into account the disposable household income, number of household members (household size) and distribution of income within the population. The main indicator is the at-risk-of-poverty rate, which represents a percentage of persons with the equivalised total disposable income below the at-risk-of-poverty threshold. The at-risk-of-poverty rate shows how many persons have an income that is below the at-risk-of-poverty threshold rather than an actual number of persons in poverty. The at-risk-of-poverty threshold is set at 60% of a medium value (median) of equivalised disposable income of all households.

The first step in creating the small area estimation model was to define variables common in both surveys, such as the household size, distribution of household members by sex and age, education status, activity status, housing characteristics (surface area of a dwelling, number of rooms) and geographical location. An analytical comparison of distributions has been carried out at the national level and at the level of spatial units for statistics in order to get the selected set of joint variables. The reason for this is that the selected set of variables should be comparable and consistent in order to ensure that the SILC based estimation model for small areas can be applied on the Census data in getting reliable estimations of income of the whole population. Using the PovMap programme, simulation models have been carried out and the equivalised disposable income per member was calculated for each household in the Census. The at-risk-of-poverty threshold has also been defined. On the basis of calculated equivalised disposable incomes and comparison with a corresponding at-risk-of-poverty threshold, at-risk-of-poverty estimations have been calculated for various territorial levels.

DEFINITIONS

Household is every family or other community of individuals who live together and jointly spend their income in order to meet the basic existential needs (accommodation, food etc.). A household is also considered every person who lives alone (one-person household).

Equivalised disposable income is calculated in a way that the total disposable household income is divided by equivalised household size calculated according to the modified OECD scale, in which the household head is given coefficient 1, every other adult aged 14 and over is given coefficient 0.5 and every child under 14 years of age is given coefficient 0.3. This procedure is applied in order to allot equal share to each member with respect to joint earnings.

Total disposable income of a household is the total net income received by household and all its members during the reference period, which is the previous year. It includes the income from paid employment, the income from self-employment, the property income, pension, social transfers and other receipts received from persons who are not household members.

The at-risk-of-poverty threshold is determined by calculating the equivalised income per household member for all households. After that, the middle value (median) of the income distribution is determined and 60% of the median is determined as the risk-of-poverty threshold. Persons with the income below the threshold are at the higher risk of poverty than others, but do not necessarily live in deprivation. The at-risk-of-poverty threshold is presented in kuna.

At-risk-of-poverty rate is a percentage of persons with the equivalised disposable income below the at-risk-of-poverty threshold.

COMMENTARY

According to the small area estimation model by income approach, the estimated at-risk-of-poverty rate in the Republic of Croatia amounted to 19.2% in 2011, while the at-risk-of-poverty threshold amounted to 24 000 kuna for one-person households. It is estimated that 19.2% of persons had an equivalent disposable income below the at-risk-of-poverty threshold in Croatia in 2011.

At-risk-of-poverty estimates at the level of statistical regions amounted to 20.0% in Continental Croatia and 17.4% in Adriatic Croatia.

Looking at the estimates at the level of counties, they range from 9.8% to 35.9%. The lowest at-risk-of-poverty rates were recorded in City of Zagreb (9.8%) and the County of Primorje-Gorski Kotar (11.9%), while the highest were recorded in the County of Slavonski Brod-Posavina (35.9%) and County of Virovitica-Podravina (33.4%).

For lower territorial levels, that is, for towns and municipalities as well as districts of the City of Zagreb with less than 15 000 households, at-risk-of-poverty rate estimates are less accurate. According to the Census of Population, Households and Dwellings in 2011 data, there were a total of 548 of these types of towns and municipalities in the Republic of Croatia.