Energy Expenses in Households with Different Standards of Living

Transcription

1 TUTWPE No 140 Energy Expenses in Households with Different Standards of Living Anton Laur Centre for Economic Research at Tallinn University of Technology 101 Kopli Street, Tallinn, Estonia Phone (372) , Koidu Tenno Centre for Economic Research at Tallinn University of Technology 101 Kopli Street, Tallinn, Estonia Phone (372) , Mare Viies Centre for Economic Research at Tallinn University of Technology 101 Kopli Street, Tallinn, Estonia Phone (372) , Abstract This article provides results of an analysis of energy expenses in households with different income and expenditure levels. The authors have used the Statistical Office of Estonia Household Income and Expenditure Survey data for the period The analysis was conducted across household income deciles, expenditure deciles and also across regions (counties). For background information the authors have examined the development of electricity consumption and electricity price trends. Price dynamics of boiler fuels used in energy enterprises in the period have also been analysed, with the main focus on price trends of renewable energy sources. The structure of Estonian households expenditure has been compared to Finland s respective data in Journal of Economic Literature Classification Numbers: D1, D31, E31, P22, Q43 Keywords: energy expenses, household member, income and expenditure deciles, expenses, energy prices, renewable energy sources 59

2 Acknowledgements The authors are grateful for financial support to the Estonian Ministry of Education and Research (Project As05) and to the Es tonian Science Foundation (Grant 5926). 1. Introduction Indicators characterising households energy expenses are treated as the main components of sustainable development energy indicators in the social sphere (Energy 2005). Social dimension is revealed primarily in that energy expenses are analysed across households that have different standards of living. This article seeks to evaluate energy expenses across households with different incomes and expenditures. For analysing energy expenses in households with different living standards we used the Statistical Office of Estonia (SOE) Household Income and Expenditure Survey (HIES) data for the period Household s energy expenses are discussed as average monthly expenditure on electricity and heat (district heating and hot water) per household member. The article describes first the HIES database and methods of SOE for determining the households living standards, as well as application of these methods. The authors summarise changes in average household income and expenditure and in their distribution. This is followed by an analysis of electricity and heat expenses across income deciles, expenditure deciles and regions. The authors also analyse development of electricity and boiler fuel prices in Estonia. Development of electricity prices and dynamics of consumer price index (CPI) are analysed for the period starting in 1992 and these data are linked to energy expenses in total expenditure of households with different living standards. The last section of the article compares the structure of expenses, including the share of energy expenses in Finland and Estonia (in 2003). 2. Database Description and Methodology The Statistical Office of Estonia has been conducting the HIES regularly since July All households living legally and permanently in Estonia are considered to be the study population. The frame of the study is the population database (population register) of Andmevara Ltd., where the persons of at least 15 years of age we re included in the survey. To draw the sample, a systematic random sampling is used. The sample size has decreased since the beginning of the survey to 520 households per month (Household 2005). Since 2000, the HIES has used a new modified methodology. The sample selection was changed with the aim to get more accurate estimates at the present sample size, to reduce the non-response rate and not to use the proxy households causing a possible bias in estimates (Household 2002). HIES, requiring completion of diaries, can be considered a quite difficult survey that tends to have a rather low response rate. The estimates of parameters of the population contain a certain bias (systematic error). The bigger the non-response is, the less reliable are the results because the opinions of those who have not answered the 60

3 questionnaire are not known. The main reasons for non-interview are frame error and non-response. The frame error contains the inaccuracies of the population database of Andmevara Ltd. In 2004, the proportion of the frame error in the sample was 8.7 and across counties this varied from 5.8 to The most common reason for frame error was that the at the given address had not been inhabited (Household 2005). The non-response rate is characterising the quality of data collection (interviewing) and attitudes of the respondents. The most frequent reason for nonresponse is refusal influenced by the diary-type questionnaire. SOE is using weights for all units in the sample to calculate estimates based on the survey data. Inclusion probabilities are the most important components in the weights used to produce estimates from the survey data. They are calculated by taking representation frequencies of the addresses and household sizes into account (Tamm 2004). The difference between the characteristics calculated by the sample and the real popular values of the characteristics is measured by sampling error (random error). The bigger the sample, the smaller, in general, is the sampling error. Due to the sampling error it is impossible to make immediate conclusions about a population, knowing just the sample characteristics. Relative error (coefficient of variation of the estimate) as a quality indicator is calculated for all estimates of HIES. For 2005, relative errors for the estimates of the average income (1.7) and expenditure (1.8) can be considered quite good (in 2000 these were lower, as the sample was bigger). But relative error level varies largely by type of income and expenditure and by county, e.g. relative error for estimate of income from property was 30.7, but the proportion of this income type in total disposable income was only 0.3 in In addition to income and the HIES provides information on the main demographic and social characteristics, and on living conditions of households. The classifications and definitions of income and consumption expenditure are harmonised with those used by the Statistical Office of the European Communities (Eurostat). Results of the HIES have been published since 1996 in special collections, including structure of incomes and expenditures per household member by type of household, education, gender, age, social status and county. For determining the households living standards, the HIES has used internationally popular methods measuring the distribution of household income and expenditure. The difference between incomes and expenditures is characterised on the basis of their classification into deciles or quintiles, or by Gini coefficient calculated on the basis of this division. Households are arranged on the basis of average monthly disposable income or expenditure per household member in ascending order and are divided into ten (decile) or five (quintile) equal groups. Then the smallest income or expenditure (first) decile (quintile) is compared to the biggest income or expenditure (tenth) decile (quintile) to find the difference between their incomes and expenditures. Calculations based on HIES data show that the monthly average mo netary disposable income per household member increased 63 in and was 3309 kroons in 2005 (see Table 1). But in terms of the purchasing power of 2000, the monthly average monetary disposable income per household member in 2005 was by 526 kroons lower, thus the real growth was only 37 during this period. As the real average monetary disposable income per household member in increased 7 (Viies et al. 2001) it can be assumed that the growth of disposable income in real terms (growth per year 2 in and 7 in ) has accelerated. 61

4 While in previous years average income was only slightly higher than average then in 2005 the monthly average monetary disposable income was by 272 kroons bigger than the average monthly monetary expenditure per household member. The growth of households income is enhancing their consumption possibilities. In recent years, average monthly expenditure per household member increased significantly compared with the beginning of the period of in and 4 in The nominal average monthly expenditure per household member increased 47 in the period and amounted to 3037 kroons in 2005 (Table 1). In terms of 2000 purchasing power, the average real exp enditure per household member increased 23 in The acceleration of expenditure growth was due to a faster growth of both nominal and real incomes in Another factor influencing the expenditure growth in recent years was an increase in borrowing and a slowdown of the growth of loan repayments (Household 2005). Table 1. Monthly Average Monetary Disposable Income and Expenditure per Household Member and Their Distribution in in Estonia* Change in 2005/2000, Disposable monetary income, kroons Nominal Real Expenditure, kroons Nominal Real Income distribution ratio, times by income deciles by income quintiles Expenditure distribution ratio, times by expenditure deciles by expenditure quintiles * Total disposable income and expenditure of households participating in the HIES extended to the entire population is divided by the total number of households estimated in the same manner. No equivalence scales have been used. Source: Statistical Office of Estonia. Though the difference between urban and rural households has diminished, disposable income and expenditure per household member still varied to a large extent in Average monetary income per household member ranged from 3831 kroons in Harju County to 2352 kroons in Hiiu County and expenditure from 3674 kroons in Harju County to 2255 kroons in Ida-Viru County. Higher than average was the monetary income per household member only in Harju and Tartu County, in addition to these, also in Pärnu County. Inequality of disposable income distribution has decreased considerably during the period (Table 1). The difference between the tenth and the first decile diminished from 13.7 times in 2000 to 9.8 times in This trend is confirmed also 62

5 by the income quintile share ratio 1, which within this period fell from 6.3 to 5.6 times. It means that disposable income per household member in the first income quintile was 5.6 times smaller than that in the fifth income quintile in Despite the reduced inequality of income distribution, it is in Estonia still much more uneven than the average of EU-25 countries where the ratio was 4.8 in 2004 (Eurostat 2005). The large inequality in income and expenditure distribution reveals deep differences in households living standards. However, a positive tendency is that disposable income per household member in the first income decile has in the period increased (nearly twofold) faster than income of an average household member and also faster than that of the tenth income decile (40.4). Households in lower income deciles have on average more members than in higher income deciles, just due to the high share of households with children. And as child and family benefits have risen substantially in recent years, then income per average household member in the first income decile has also increased. Lower income deciles have also more nonworking members, especially unemployed, than average households. The difference between the tenth and first expenditure decile in terms of expenditure per household member in 2005 remained the same as it was in 2000, still being as before more than tenfold. Inequality calculations of expenditure distribution on the basis of Gini coefficient also indicate that no changes have occurred in expenditure distribution and the Gini coefficient is 36 as before. Hence, consumption possibilities of households differ to a large extent and lower income households have little possibilities to consume as much as they need after making all inevitable expenditure on food and. Average monetary expenditure per household member vary considerably depending on income or expenditure decile, especially in the first decile (in 2005 expenditure per household member in the first income decile was 1.8 times larger than in the first expenditure decile). This can be explained by much larger expenditure than income in lower income deciles or in some cases hidden incomes. For correct conclusions it must be taken into account that though data for the whole year are used, for each household it is still monthly income meaning that households in the first decile needn t earn constantly low income but only in the month of participating in the survey. According to SOE, division into deciles enhances the impact of atypical households in extreme deciles on the estimates, and the amount of households divided into ten deciles is too small for a comprehensive survey. Therefore, the SOE in 2003 abandoned more detailed analyses of household income and expenditure on the basis of deciles (Household 2004). 3. Energy Expenses in Household Expenditure A major component of household expenditure is expenses, which after food is the second largest group of monetary expenses: in 2005, food expenses accounted for 25.1 of average expenditure per household member, expenses Energy expenses, including electricity and heat, in turn accounted for nearly 40 of expenses, whereas this share has been rising over years. 1 The ratio of total equalised disposable income received by 20 of the population with the highest income (top quintile) to that received by 20 of the population with the lowest income (lowest quintile). 63

6 Households nominal expenses on electricity and heat, and their share in total expenditure and in expenses are presented for the period by income deciles (Tables 2-3), expenditure deciles (Tables 4-5) and counties (Tables 6-7). Data provided in the tables enable to conduct the following analysis. Average nominal expenses on electricity per household member by income deciles in the period (Table 2) increased from 60 kroons to 102 kroons, or 70. These expenses have grown most in the first income decile (households with the lowest income) from 37 kroons to 75 kroons, or 103. The share of electricity expenses in total expenditure and in expenses in these years has also been the highest mostly in the first income decile. Nominal expenses per household member on heat by income deciles in the period (Table 3) increased from 70 kroons to 87 kroons, or 24. Also heat expenses increased most in the first decile in this period - from 31 kroons to 46 kroons, or 48, while in the highest income deciles these expenses grew little and at the end of the period decreased again. The share of heat expenses in total expenditure and in expenses is in general higher in the first income deciles. If to compare increase in household expenses on electricity and heat in this period and the growth of consumer price index (CPI) it appears that the growth of energy expenses has been faster, especially that of electricity. Hence, there has also been a real growth of energy expenses. If to take into consideration the basic index of CPI for the period , the average real expenses per household member on electricity increased 43 and on heat 4.3 in this period. Expenses across expenditure deciles on both electricity and heat divide even more unevenly than between income deciles. While electricity expenses in the period increased 103 in the first income decile and their share in total expenditure from 3.5 to 5 percent, and in expenses from 22 to 29 percent (Table 2), then electricity expenses in the first expenditure decile increased 132, their share in total expenditure from 4.4 to 7 percent, in expenses from 24 to 35 percent (Table 4). As for heat, the picture is somewhat different. Though nominal heat expenses in the period increased most in the first expenditure decile from 19 to 37 kroons, or 95, the share of heat expenses both in total expenditure and in expenses was the highest in the fourth expenditure decile, whereas these shares did not increase in the period of analysis (Table 5). Nominal expenses on electricity across counties (Table 6) per household member and as a share in total expenditure and in expenses was the biggest mainly in Hiiu County (in 2001 exceptionally in Rapla County). Electricity expenses per household member in Hiiu County increased from 84 kroons in 2000 to 136 kroons in 2005, or 62. The share of electricity expenses in Hiiu County was the biggest in 2004: in total expenditure 6 and in expenses 53. Electricity expenses have been the smallest in Ida-Viru County; their share in total expenditure and in expenses has been the lowest in Tallinn and also in Ida-Viru County. These data quite well conform to electricity consumption data per capita consumption of electricity by households has in most of the years been the highest in Hiiu County and the lowest in Ida-Viru County. In Figure 1 we have compared per capita consumption of electricity by households and annual average expenses on electricity per household member across counties in We can see that the expenditure curve quite flexibly repeats the variation of electricity consumption by households across counties. This Figure also indicates the highest level of electricity expenses in Hiiu County, which is also in accordance with bigger consumption there (Harju County was by a narrow margin the leader in electricity consumption per capita this year). It is not so easy to explain the 64

7 high level of electricity consumption by households in Hiiu County one reason may be shortage (local fuels) or inaccessibility (natural gas) of other energy resources. Nominal expenses on heat across counties (Table 7) are as a rule the lowest in Hiiu County (where electricity expenses are the biggest). Heat expenses in terms of absolute value are the highest in Tallinn, as a share in total expenditure and in expenses in Ida-Viru County. This is according to expectations because, as mentioned above, these are primarily expenses on district heating and both in Tallinn and in Ida- Viru County the share of district heating in heat supply is high. Figure 1. Estonian Households Electricity Consumption and Electricity Expenses by Counties in kwh per capita Tallinn Harju Hiiu Ida-Viru Jõgeva Järva Lääne Lääne - Viru Põlva Pärnu Rapla Saare EEK per household member Tartu Valga Viljandi Võru 0 Households' electricity consumption Electricity expenses Source: Statistical Office of Estonia 65

8 Table 2. Household Expenses on Electricity and Share in Total and Dwelling Expenditure by Income Deciles in (Monthly Average per Household Member) Electri- Share in Electri- Share in Electri- Share in Electri- Share in Electri- Share in Electri- Share in Income city total city total city total city total city total city total deciles expenses, and expenses, and expenses, and expenses, and expenses, and expenses, and kroons kroons kroons kroons kroons kroons I II III IV V VI VII VIII IX X Average Source: Statistical Office of Estonia 66

9 Table 3. Household Expenses on Heat and Share in Total and Dwelling Expenditure by Income Deciles in (Monthly Average per Household Member) Heat Share in Heat Share in Heat Share in Heat Share in Heat Share in Heat Share in Income expenses, total expenses, total expenses, total expenses, total expenses, total expenses, total deciles kroons and kroons and kroons and kroons and kroons and kroons and I II III IV V VI VII VIII IX X Average Source: Statistical Office of Estonia 67

10 Table 4. Household Expenses on Electricity and Share in Total and Dwelling Expenditure by Expenditure Deciles in (Monthly Average per Household Member) Electri- Share in Electri- Share in Electri- Share in Electri- Share in Electri- Share in Electri- Share in Expenditure city total city total city total city total city total city total deciles expenses, and expenses, and expenses, and expenses, and expenses, and expenses, and kroons kroons kroons kroons kroons kroons I II III IV V VI VII VIII IX X Average Source: Statistical Office of Estonia 68

11 Table 5. Household Expenses on Heat and Share in Total and Dwelling Expenditure by Expenditure Deciles in (Monthly Average per Household Member) Heat Share in Heat Share in Heat Share in Heat Share in Heat Share in Heat Share in Expenditure expenses, total expenses, total expenses, total expenses, total expenses, total expenses, total deciles kroons and kroons and kroons and kroons and kroons and kroons and I II III IV V VI VII VIII IX X Average Source: Statistical Office of Estonia 69

12 Table 6. Household Expenses on Electricity and Share in Total and Dwelling Expenditure by Counties in (Monthly Average per Household Member) Counties Electri- Share in Electri- Share in Electri- Share in Electri- Share in Electri- Share in Electricity total city total city total city total city total city expenses, and expenses, and expenses, and expenses, and expenses, and expenses, kroons kroons kroons kroons kroons kroons Harju incl.tallinn Harju Hiiu Ida-Viru Jõgeva Järva Lääne Lääne-Viru Põlva Pärnu Rapla Saare Tartu Valga Viljandi Võru Average Source: Statistical Office of Estonia 2005 Share in total and 70

13 Table 7. Household Expenses on Heat and Share in Total and Dwelling Expenditure by Counties in (Monthly Average per Household Member) Counties Heat Share in Heat Share in Heat Share in Heat Share in Heat Share in Heat expenses, total expenses, total expenses, total expenses, total expenses, total expenses, kroons and kroons and kroons and kroons and kroons and kroons Harju incl.tallinn Harju Hiiu Ida-Viru Jõgeva Järva Lääne Lääne-Viru Põlva Pärnu Rapla Saare Tartu Valga Viljandi Võru Average Source: Statistical Office of Estonia 2005 Share in total and 71

14 4. Price Developments of Electricity and Boiler Fuels Growth of electricity expenses in household expenditure in the period of analysis depended on both electricity consumption by households and rise of prices. Electricity consumption by households has been relatively stable since 2000, having increased around 10 in the period Ensuing from the decline in the number of Estonian population, consumption by households per capita increased 12 in this period. Stable growth of electricity consumption by households is a sign of rising standard of living of the population. However, electricity price rise for households was much faster in this period, thus exerting a bigger effect on households electricity expenses. Dynamics of the average nominal price of electricity and basic tariff for households (both excluding VAT) in the period is depicted in Table 8, development trends of consumer price index and basic tariffs of electricity for households for the same period in Figure 2. It seems that the basic tariff of electricity for households has risen 8.4 times in this period (from 12.7 sents/kwh to sents/kwh). CPI growth (inflation) was until 2001 faster than that of the basic tariff of electricity for households, but since 2001 it has been rising faster than the rate of inflation. The basic index of CPI has throughout the period been Hence, real growth of the basic tariff of electricity for households in the period was 23. In , the basic tariff for households in nominal terms increased nearly 1.7 times and in real terms 1.4 times. Table 8. Nominal Price Developments of Electricity (VAT Excluded) Date of implementation Average price for electricity *, Sents/kWh Basic tariff for households, sents/kwh * Since 1999 average electricity price is not used Source: Estonian Energy

15 Figure 2. Development Trends of the Basic Tariff for Households and Consumer Price Index (CPI) in Source: Estonian Energy 2004 CPI Basic tariff for households Household expenses on heat are influenced primarily by changes in fuel prices used for the production of heat. Unlike electricity expenses, the effect of changes in consumption is smaller here, because improvement of the standard of living (e.g., better quality of s) may even reduce consumption of heat. Consumption of heat is much more influenced also by differences in weather across years (Tenno and Laur 2000). Data on price changes of the main boiler fuels in the period are given in Table 9. If to compare prices of local renewable fuels (per unit of energy) and their dynamics with prices of fossil fuels, we can see that in 2000 prices of sod peat, firewood and wood chips were approximately 2 times lower than, for example, the price of heavy fuel oil and times lower than the price of natural gas, but thereafter the difference has diminished rapidly. Table 9 suggests that the price rise of wood fuels and also that of sod peat was much faster until A sharp rise of oil prices in the world market which started in 2005 and the ensuing price rise of shale oil have increased the price difference between wood-peat and fossil fuels again. Still, the price rise of firewood for the whole period, for example, exceeds also the result of fuel oil price shock. This is a clear sign of the limited resources of wood fuels against the background of growing demand. This hazard and the importance of assessing wood resources was pointed out in an article in 2000 (Muiste and Kask 2000). The price rise of local renewable biofuels in the period of analysis, coupled with the recent sharp price rise of shale oil, may considerably increase the energy expenses of rural households who consume heat produced in small boiler houses, as well as jeopardise the efficiency of a shift to using biofuels in many of such boiler houses. 73

16 Table 9. Average Prices (Excl. VAT) of Boiler Fuels in Power and Boiler Plants, EEK/MWh Fuel Price Change /2000 Hard coal Sod peat Firewood Wood chips Natural gas Heavy fuel oil Shale oil Light fuel oil Source: Estonian Energy Comparison of Dwelling Expenses with Finland It is conspicuous that in comparison with many European countries, primarily so-called old EU member states, the share of energy expenses in Estonian households expenses is remarkably higher. For example, household expenditure on electricity in Finland accounted for only about 9 of expenses in 2003, in Estonia nearly 22. So as to explain this problem more profoundly, we have provided in Table 10 comparative data for Finland and Estonia on the structure of expenses. We can see that a reason for the big difference is obviously the dominating share of the socalled real estate component (actual and imputed rent) in Finnish expenses. Total rent accounted for nearly 70 of these expenses there in In Estonia rent accounted only for 6 of expenses in 2003, whereas only actual rent was taken into account. The imputed rent (i.e. rent to be paid by owner of in case he was the tenant) calculation began in Estonia only in 2004 in connection with accession to the European Union. Therefore, the structure of expenses should change also in Estonia from 2004 in the direction of falling share of energy expenses. Actually, as above data indicate, the share of expenditure on electricity in 2004 and 2005 remained almost on the level of 2003, only the share of heat expenses has fallen very slightly (by ca 2). Table 10. Structure of Dwelling Expenses in 2003, Expense item Finland Estonia DWELLING EXPENDITURE (total) incl. factual rent imputed rent maintenance and repair water etc. services, gas etc. fuels and heat Electricity Sources: Household Living Niveau 2004; 74

17 6. Concluding Remarks The analysis conducted by us indicated that electricity and heat expenses together accounted for a remarkable share of expenses in the years of analysis on average 38 to 44. The share of these expenses in total monetary expenditure by households was fluctuating at around 7. In population groups and counties with smaller incomes (first income deciles) electricity and heat expenses together account for nearly 50 of expenses (in some cases even more) and up to 10 of total expenditure. These population groups are also most influenced by the continuing price rise of electricity and district heat (in the case of heat also a rise of VAT to 18 from 1 July 2007). Household expenses on energy largely depend on developments of electricity and heat prices. The electricity price has always been regulated by the government and therefore exact information is available on its dynamics. The basic tariff of electricity for households has risen nearly 1.7 times in the period of analysis, the real price considering inflation (CPI growth) 1.4 times. The electricity price rise and increasing electricity consumption have caused a significant rise of household electricity expenses (41 in real terms over the period ). Regulation of heat prices has been amended repeatedly: until 1998 it was the domain of local governments; from 1998 until 17 May 2001, all heat selling enterprises had to get approval to change the price from the Energy Market Inspectorate. After 17 May 2001, this is applicable only to producers selling more than 50,000 MWh heat in an economic year. Therefore it is not possible to provide an explicit and comprehensive overview of heat prices. Still, heat producers are from year to year improving the database organised by the Estonian Power and Heat Association (EPHA). One possibility of reducing energy expenses in household expenditure in addition to energy saving is to increase the use of cheaper local fuels, including renewable sources, especially for production of heat. Analysis of heat prices on the basis of EPHA database indicates that heat produced from firewood, wood chips, sawdust and in some cases also from peat is often cheaper than heat produced from fuel oils and even from natural gas. Considering the continuing rise of world market prices of imported fuels, the use of renewable fuels for heat production should be definitely promoted and the share of renewable energy increased. The use of renewable fuels has also an additional direct effect from saving pollution charges primarily due to remarkable reduction of air pollution, as well as an additional macroeconomic effect from lower need for imported fuels. At the same time, a disturbing problem is faster than expected price rise of wood fuels and peat in recent years, the main reason for which is the growing export pressure on these energy resources. This may jeopardise management of projects that have been launched already or of energy enterprises that have been reconstructed to use renewable fuels and hinder increasing the further use of renewable fuels. Obviously it is time to think about using state regulation in this area. 75

18 References Estonian Energy Tallinn: Ministry of Economic Affairs and Communication. Eurostat Energy Indicators for Sustainable Development: Guidelines and Methodologies Vienna: International Atomic Energy Agency, UN Department of Economic and Social Affairs, IEA, Eurostat and EEA. Household Living Niveau Yearbook. Tallinn: Statistical Office of Estonia. Household Living Niveau Yearbook. Tallinn: Statistical Office of Estonia. Household Living Niveau Yearbook. Tallinn: Statistical Office of Estonia. Muiste, P. and Ü. Kask Biomass biokütus bioenergia puitkütus. In: V. Tiit, ed. Taastuvate energiaallikate uurimine ja kasutamine. Tartu: Eesti Põllumajandusülikooli Kirjastus, Tamm, E Leibkonna elujärje uuringu tulemuste täpsuse hindamisest. Eesti Statistikaseltsi teabevihik: Statistika majanduses ja majandusteaduses. Tallinn: Statistikaamet, Tenno, K. and A. Laur Eesti regionaalne sotsiaalmajanduslik areng ja energiaressursside kasutamine. Tallinn: TTÜ Eesti Majanduse Instituut. Preprint 59. Viies, M., Kukk, A. and M. Kreitzberg Household income and consumption. In: Social Trends 2. Tallinn: Statistical Office of Estonia, Statistics Finland. 76