Greek Farm Households: Income inequality, poverty and distributional impact of farm income

Transcription

1 AUA Workng Paper Seres No March 2009 Greek Households: Income nequalty, poverty and dstrbutonal mpact of farm ncome Pavlos Karankolas Department of Agrcultural Economcs & Rural Development Agrcultural Unversty of Athens Stavros Zografaks Department of Agrcultural Economcs & Rural Development Agrcultural Unversty of Athens Agrcultural Unversty of Athens Department of Agrcultural Economcs & Rural Development 1

2 Greek Households: Income nequalty, poverty and dstrbutonal mpact of farm ncome 1 Authors: Pavlos Karankolas: Assstant Professor, Agrcultural Unversty of Athens, Department of Agrcultural Economcs and Rural Development, Iera Odos 75, Athens 11855, GREECE, Tel: , E-mal: pkarank@aua.gr Stavros Zografaks: Assstant Professor, Agrcultural Unversty of Athens, Department of Agrcultural Economcs and Rural Development, Iera Odos 75, Athens 11855, GREECE, Tel: , E-mal: stazog@aua.gr Abstract Ths paper examnes the ncdence of ncome nequalty and poverty, and the mpact of farm ncome on nequalty. A detaled typology of farm households (FHs) s developed, based on Household Budget Survey mcro-data. Research fndngs reveal enormous varatons among households wth respect to ncome nequalty and poverty. Whle Margnal- and Pluractve- FHs do not seem to have an ncome problem, ths s not the case for Households. Poverty s a wdespread phenomenon among Retred FHs. ncome and non-farm ncome generate a combned stablzaton effect, mtgatng the overall nequalty wthn households. Polcy mplcatons of these fndngs are dscussed n the context of welfare aspects of agrcultural polcy. Keywords: farm households, ncome nequalty, poverty, dstrbutonal mpact 1. Introducton Inequalty s typcally defned as nequalty of welfare dstrbuton, whch s usually approached through current consumpton and/or current ncome. Many emprcal studes of economc nequalty have been conducted at the level of varous soco-professonal groups of populaton, among whch, farmers are a dstnct case, due to the specal features of agrculture. The ncomes of farmers and ther households have been a prme concern of both academc and wder publc dscourse for qute a long tme. Agrcultural actvtes may be practsed on a full-tme, seasonal or part-tme bass. Hence, the noton of an agrcultural household s a multfaceted one, snce many households depend on agrculture for ther lvelhoods, though at a varyng degree. Thus, ncome-related matters n agrculture have never been one-dmensonal, 1 Ths paper has been publshed at Socal Coheson and Development, Vol. 3 (1), 2008, pp

3 as the basc unt of analyss n agrculture the farm operaton s ntertwned wth consumpton and socal unts (the farm household and farm famly, respectvely). Ths complex stuaton necesstates a measurement of household members ncome whch wll cover the return from farmng, as well as from any other actvty (selfemployment or waged), socal transfers, pensons, etc. It has been suffcently documented n the lterature that nequaltes wthn agrculture surpass nequaltes among non-agrcultural households. Also, agrcultural nequaltes and poverty dffer structurally from those n the rest of socety (Pauw, 2007; Commns, 2004; Frawley et al., 2000). Takng nto consderaton, not just farmng but all ncome sources avalable to farm household members, noteworthy ntra-household varatons are revealed. For example, Mshra and Sandretto (2002) examnng the US agrcultural sector from 1933 to 1999, fnd that the varablty n real net farm ncome has not dmnshed and that non-farm ncome has contrbuted to the reducton of varablty n total farm household ncome. ncome also consttutes the man element of the famous farm problem, n the sense that, frstly, ncome from farmng lags persstently behnd the ncome generated by the other sectors of the economy and secondly, t presents a remarkable temporal and spatal varaton (Gardner, 1992). These two factors render farm households one of the most vulnerable and low-ncome groups n socety (Hll, 1999; 2000). The elmnaton of ths ncome gap between persons employed n agrculture and n other ganful actvtes has been the focal pont of agrcultural polces n developed countres for the most part of the 20 th century (Schmtt and Boruse, 1996; OECD, 1987). Thus far, farm households are consdered a rather homogeneous category n the relevant lterature, usually beng compared wth non-farm households. Consequently, the multple forms of agrculture s ntegraton wth the (farm) household, as well as ntra-farm household characterstcs and dfferences, are nsuffcently depcted. Furthermore, any aggregate analyss tends to obscure serous dstrbutonal ssues, such as the dstrbuton of ncome wthn agrculture (Sarrs and Zografaks, 1996). Also, an examnaton of farm households n an undfferentated manner does not facltate an evaluaton of lkely mpacts of varous polcy measures at the holdng/household level. Ths s especally relevant now, gven the profound changes n Common Agrcultural Polcy (CAP) and the renewed nterest n the welfare aspect of CAP. Ths aspect, once expressed through the objectve of ensurng a far standard of lvng for the agrcultural communty, s an evergreen topc, as unequal dstrbuton of ncomes and subsdes n agrculture are contnuously consdered an essental reason for CAP s reduced effectveness (Fennel, 1997). In ths context of farm and non-farm ncome nterplay, crtcal ssues arse, concernng the ncome levels acheved by farm households, the ncdence of low ncome and poverty among them, the ncome and welfare nequaltes wthn the agrcultural sector and wthn farm households, the dentfcaton of specfc groups of farmers or farm households who don t have a real ncome problem, etc. Ths paper ams, frstly, at dentfyng ncome nequalty and poverty ncdence n Greek households. Secondly, crtcal dstrbutonal ssues are explored, such as the contrbuton of farm and non-farm ncome n the overall nequalty of households. Households are classfed accordng to ther relaton to agrculture, on the bass of the professonal status of the head of the household and the contrbuton of agrculture to famly ncome. Consequently, four dfferent types of farm households are dscerned, whch are then compared wth non-farm households. 3

4 Ths study s comprsed of fve parts. Part 2 brefly revews prevous studes on ncome nequalty and poverty n Greek farm households. In the 3 rd part, the data source and methodology are outlned, ncludng the creaton of a typology of households whch s used n the remander of the study. The results then follow, referrng to ncome nequaltes, the dstrbutonal effects of farm ncome and the decomposton of nequalty and poverty. Part 5 concludes by summarzng the results and dscussng some of ther polcy mplcatons. 2. Income nequalty and poverty n Greek farm households Most of the evdence on ncome dspartes and poverty n Greece emanate from Household Budget Surveys (HBS) data. Mtrakos and Tsakloglou (2003; 2006) found that n relatve terms, nequalty and poverty declned sgnfcantly between 1974 and 1982 whereas the changes n the perod after 1982 were smaller n sze and, sometmes, contradctory. In absolute terms, poverty declned substantally between 1974 and 1982 and, most probably, ths change contnued after 1982 too, but at a much slower rate. Accordng to Sarrs and Zografaks (1996) ncome dstrbuton wthn farm households s much more skewed compared wth ncome dstrbuton of non-farm households, and ths pattern does not seem to change over tme. Some other nterestng fndngs ndcate that the bulk of nequalty n Greece s due to dspartes wthn, rather than between, groups, even when the populaton s grouped nto a large number of small homogeneous groups (Mtrakos and Tsakloglou, 1998). The same concluson s drawn n the case of farm households, when they are grouped accordng to varous crtera, such as place of resdence, age, household sze, educatonal qualfcatons etc. (Mtrakos and Sarrs, 2003). Moreover, employment n the agrcultural sector, along wth old age, resdence n rural areas, low educatonal qualfcatons and, to a lesser extent, lack of employment have been dentfed as closely assocated wth acute poverty. Ths concluson s drawn rrespectve of the welfare ndcator, the level of the poverty lne, or the sze of the equvalence scales used n the analyss (Tsakloglou and Panopoulou, 1998) It ensues then, from all avalable HBS, that farm households n Greece are among the low-ncome groups n socety. Nonetheless, after a decade of relatve stablty of nequaltes ( ), a decrease n the ncome and well-beng dscrepancy between farm and non-farm households takes place over the perod Thus, whle the number of farm households rapdly decreased, ther per capta ncome converged at the same tme, to non-farm households by nne percentage ponts, from 74% to 83% (Mtrakos ans Sarrs, 2003; Karankolas et al., forthcomng). Another source of nformaton on total ncomes of agrcultural households orgnates from a specal project devoted to the Incomes of the Agrcultural Household Sector (IAHS). Ths project was launched by Eurostat n md-1980 s and ceased n Accordng to these data, ncome from ndependent agrcultural actvty accounted for 61.6% of the total ncome of Greek agrcultural households n 1982, a percentage whch dropped to 58.3% n 1998 (Eurostat, 2002). 3. Data and methodology Our analyss s based on the mcro-data of the most recent Household Budget Survey, whch was carred out by the Natonal Statstcal Servce of Greece n

5 2005. To accomplsh the am of ths paper, two crtera are used for the groupng of populaton. The frst s the occupatonal status of the household s head (employed n agrculture or n other actvtes). Secondly, the magntude of varous components of farm ncome n the total ncome of a household s employed, that s, ncome from the market, subsdes plus compensatons for natural dsasters, and the value of mputed agrcultural producton. Retred farmers comprse an addtonal dstnct group, on account of ther mportance wthn the total farm populaton and the totalty of poor people n Greece. As s well known, a large part of the farm populaton s aged; accordng to Eurostat 37 percent of heads of Greek agrcultural holdngs have an age of more than 65 years (Eurostat, 2007). An addtonal reason for ther separaton s that most of them stll work on agrculture after ther retrement, operatng an agrcultural holdng. Therefore the resultant typology s as follows: Table 1: Types of Households All Households Households (FHs) Plur-actve Households (PFHs) Margnal Households (MFHs) Retred Households (RFHs) Non- Households (NFHs) Descrpton Natonal Average Household Household head reports an occupaton n agrculture or fshng Household head has a non-agrcultural job; household ncome ncludes ether: ) some ncome and subsdes from agrculture, or ) some ncome from agrculture Household head has a non-agrcultural job; household ncome ncludes some subsdes from agrculture Household head gets a penson from ers Securty Organzaton Household head has a non-agrcultural job; no ncomes or subsdes from agrculture are reported In pursung the am of ths paper, the analyss s undertaken at the level of the abovementoned typology of households. After the examnaton of the composton of household ncome, nequalty s scrutnzed through the calculaton of ncome nequalty ndexes based on the dstrbutons of both per capta equvalent expendture and per capta equvalent ncome. Also, some crtcal dstrbutonal effects of farm ncome are explored, such as the concentraton of farm ncome across the spectrum of the ncome dstrbuton n absolute terms and the contrbuton of farm ncome n the reducton of the overall nequalty wthn varous types of households. Fnally, overall nequalty s decomposed to ts consttuent parts and poverty ncdence s estmated wth alternatve defntons of poverty lnes. 4. Results 4.1 Income nequaltes Table 2 shows that all types of farm households sum up to 22 percent of all households n Greece. It s nterestng to note that ther total sum (877,836) does not devate substantally from the total number of agrcultural holdngs n Greece, whch for 2005 were found to be 833,590 (Eurostat, 2007). households are only one quarter of them, or 5.5 percent of all households. Margnal farm households represent 5

6 only 2.3% of all households, comprsng the smallest category of all. As far as the household sze s concerned (members per household) farm households rank frst, exceedng the natonal average by 0.72 members, followed by plur-actve farm households. It s also worth notng that FHs are double the sze of RFHs. Table 2: Households: basc data N % Household sze Households (FHs) 217, % 3.45 Plur-actve Households (PFHs) 302, % 3.15 Margnal Households (MFHs) 90, % 2.78 Retred Households (RFHs) 266, % 1.72 Non- Households (NFHs) 3,115, % 2.73 All Households 3,992, % 2.73 FH+PFH+MFH+RFH 877, % The examnaton of relatve well-beng and ncome levels reveals mportant between-households varatons (table 3). Retred farm households are the most extreme case snce they seem to compare unfavourably wth the rest of socety n terms of ther equvalent ncome (50.4% of the average) and equvalent expendture (56.0% of the average). Note that those ndcators are expressed n equvalent terms, usng the modfed OECD scales, whch have been used n a number of emprcal poverty studes (Hagenaars et al., 1994) and assgn weghts of 1.0, 0.5 and 0.3 to the household head, each of the remanng adults and each chld n the household respectvely. At the other end of the scale, the most prosperous category seems to be nonfarm households, surpassng both natonal averages by 5.5 and 4.7 percentage ponts. Margnal farm households revolve around the all-households average, slghtly hgher for equvalent ncome and slghtly lower for equvalent expendture. Plur-actve farm households are on a par wth average equvalent ncome, but not so for equvalent expendture. Despte ther convergence trends durng the last few years, farm households stll fall short of the all-households average by 9.1 ponts n terms of equvalent ncome and by 15.9 ponts n terms of equvalent expendture. Table 3: Equvalent household expendture and ncome (amounts n ) Equvalent Expendture Equvalent Income Households (FHs) 1, , Plur-actve Households (PFHs) 1, , Margnal Households (MFHs) 1, , Retred Households (RFHs) Non- Households (NFHs) 1, , All Households 1, ,

7 Another essental dfference among types of households refers to the composton of household ncome and, more specfcally, the share of varous components of agrcultural ncome n household ncome (table 4). The pcture s qute heterogeneous. Agrcultural ncome s share n total ncome of varous types of households vares from 55.0% n FHs to 3.5% n RFHs. As expected, the hghest share s found n FHs. Ths means that a sgnfcant part of ther total ncome s derved by non-agrcultural sources, a clear ndcaton of serous changes that have taken place wthn FHs. Obvously, there s a strategy of dversfcaton of actvtes on behalf of household members, even n the case of FHs, where farmng remans the man source of ncome. But even n ths type of household, only 36.8% of total ncome s derved from the market, whereas 13.2% comes from subsdes and 5.0% from consumpton of own producton. Our results conform to Eurostat s fndngs on total ncome of agrcultural household sector (Eurostat, 2002). It also has to be noted that n 20 OECD countres, the percentage share of farm ncome n total ncome of farm households ranges from 6% n USA through 72% n the Netherlands (OECD, 2004). Moreover, n PFHs 23.0% of total ncome s drawn from agrculture, of whch 14,4% s farm ncome from the market, 5.1% from subsdes and 3.5% from agrcultural consumpton of own producton. Much lower s the contrbuton of farm ncome to the total ncome of the other two household types. In the case of MFHs, ths amounts to 6.7%, consstng of subsdes and consumpton of own producton. Recall that those households produce just tny quanttes of agrcultural products exclusvely for own-consumpton, for whch they get some subsdes. Fnally, n RFHs the only element of agrcultural ncome s the value of consumpton of own producton that represents only 3.5% of total ncome. Table 4: Income composton Hous ehold Inco me Income (wthout Subsd es) Agr. Subsd es and Compen satons Value of Agr. Consum pton of own product on Total Inco me Non- Income Households (FHs) 100,0 36,8 13,2 5,0 55,0 45,0 Plur-actve Households (PFHs) 100,0 14,4 5,1 3,5 23,0 77,0 Margnal Households (MFHs) 100,0 0,0 3,9 2,8 6,7 93,3 Retred Households (RFHs) 100,0 0,0 0,0 3,5 3,5 96,5 Non- Households (NFHs) 100,0 0,0 0,0 0,4 0,4 99,6 All Households 100,0 3,3 1,3 1,0 5,5 94,5 As already mentoned, across the perod a convergence of wellbeng and ncome levels between farm and non-farm households has taken place, as well as a steadness n the contrbuton of farm ncome to the total ncome of farm households (Karankolas et al., 2008). Obvously, the former depends on the latter. For the measurement and decomposton of nequalty, the well known ndexes from the lterature are deployed, such as G (Gn Coeffcent), T (Thel Index) and N (Mean Log Devaton) (see Annex for defntons of the ndexes). The overall 7

8 nequalty as measured by G, T and N ndexes s found to be 0.295, and respectvely, accordng to per capta equvalent expendture dstrbuton (table 5). Relatvely hgher s the overall nequalty, f the dstrbuton of per capta equvalent ncome s used (0.300, and 0.149, respectvely). Table 5: Income nequalty ndexes by type of household Income Inequalty based on the dstrbuton of: Per capta equvalent Per capta equvalent expendture ncome Gn Coeffc ent (G) Thel Index (T ) Mean Log Devato n (N) Gn Coeff cent (G) Thel Index (T ) Mean Log Devat on (N) Households (FHs) Plur-actve Households (PFHs) Margnal Households (MFHs) Retred Households (RFHs) Non- Households (NFHs) All Households RFHs exhbt the lowest nequalty of all types of households (table 5). Ther nequalty s much lower than the respectve all-households nequalty, by 29% on the bass of per capta equvalent expendture and by 48% accordng to per capta equvalent ncome. Thus, RFHs are placed at a very low well-beng and ncome level, although ther nequalty s the smallest of all other types of households. On the other hand, from the households wth an economcally actve head, PFHs exhbt the lowest nequalty whch s almost the same for the two dstrbutons employed, yet t lags behnd the natonal average by 10% or 24%. Interestngly, nequalty of MFHs s almost dentcal to all-households nequalty and hghest among all agrcultural households by per capta equvalent expendture. Of partcular mportance s the case of FHs. On the bass of per capta equvalent expendture dstrbuton, ther nequalty s lower than all-households nequalty, a possble ndcaton of convergng consumpton patterns of FHs. Yet, at the same tme, ther ncome nequalty exceeds the average all-households nequalty by 37%. Of course, ths could be accounted for by the hgh sectoral and geographcal varaton of FHs, followng the hghly heterogeneous producton specalzaton of farm holdngs and ther dverse economc performance. Hence, nequalty of FHs s the hghest among all types of agrcultural households. Fnally, nequalty of NFHs s slghtly lower than the average nequalty of all households. 4.2 Dstrbutonal effects of farm ncome It s nterestng to examne to what extent the varous types of households are concentrated manly at the bottom, the mddle or the top of the ncome dstrbuton. Table 6 provdes an answer. Households are splt nto decles on the bass of per capta equvalent ncome, so that each decle contans 10 percent of all households. In comparson to all households and non-farm households, farm households are much 8

9 more concentrated n the lowest decles, followed by MFHs and PFHs. Hence, the stronger the engagement of a household wth agrculture on a professonal bass, the hghest the possblty of t beng a low ncome household. On the other hand, 86 percent of RFHs are found n the three lowest decles, a clear ndcaton of the especally low ncome level of ths type of household. Table 6: Allocaton of households by type and decles (%) All Househ olds Households (FHs) Plur-actve Households (PFHs) Margnal Households (MFHs) Retred Households (RFHs) Non- Househol ds (NFHs) Decle Decle Decle Decle Decle Decle Decle Decle Decle Decle The next ssue to be addressed concerns the concentraton of farm ncome across the band of the ncome dstrbuton. The absolute value of farm ncome and subsdes per decle s qute dssmlar, especally n the case of farm households (Annex Table 1). In absolute terms, farm households of the top four decles obtan much hgher farm ncome and subsdes compared wth the lower sx decles. Qute dssmlar s also the dstrbuton of farm ncome across decles n PFHs and MFHs. On the other hand, the relatve contrbuton of farm ncome to total household ncome vares across decles (table 7). For example, n the case of FHs, farm ncome share n household ncome ranges from 41.6% to 64.7%; the hghest percentages are found n 10 th and 6 th decles, however n most cases ths ndex s dstrbuted rather evenly. Furthermore, n the case of PFHs, farm ncome share tends to declne as we move from low to hgh decles. As far as the subsdes share to farm ncome s concerned, t doesn t seem to follow a clear pattern of dstrbuton across decles; an exceptonal case s the 70.2 percent n the 9 th decle of FHs. Table 7: ncome and subsdes by type of households and decles (%) 9

10 All Househo lds Households (FHs) Pluractve Househol ds (PFHs) Margnal Households (MFHs) Decle 1 farm ncome/household ncome subsdes/farm ncome Decle 2 farm ncome/household ncome subsdes/farm ncome Decle 3 farm ncome/household ncome subsdes/farm ncome Decle 4 farm ncome/household ncome subsdes/farm ncome Decle 5 farm ncome/household ncome subsdes/farm ncome Decle 6 farm ncome/household ncome subsdes/farm ncome Decle 7 farm ncome/household ncome subsdes/farm ncome Decle 8 farm ncome/household ncome subsdes/farm ncome Decle 9 farm ncome/household ncome subsdes/farm ncome Decle 10 farm ncome/household ncome subsdes/farm ncome Apart from the estmaton of varous nequalty ndexes for each type of household, another prme ssue s the dentfcaton of some dstrbutonal effects of the man ncome sources. Therefore, the analyss focuses on the contrbuton of farm and non-farm ncome n the overall nequalty wthn varous types of households. As already noted, the Gn coeffcent for the total ncome of farm households s (table 5). However, the Gn coeffcent for ther farm ncome s Of course, ths could be explaned by the hghly heterogeneous condtons of agrculture whch result n an astoundng sectoral and geographcal varaton of farm ncome. At the same tme, the ncome from sources other than farmng undoubtedly contrbutes to the reducton of the overall ncome nequalty, as has already been documented n many countres (for the USA see Mshra and Sandretto, 2002). As expected each component of the total farm ncome of farm households s more unequally dstrbuted than ther sum. Gn coeffcents for agrcultural compensatons, subsdes and ncome from farmng are 0.959, and respectvely. Subsequently, despte unequal dstrbuton of each component of farm ncome, the resultng nequalty s lower when they are summed up to form total farm ncome. Evdently, ths s, to a large extent, due to techncal reasons, as the smaller the number of farm households havng one source of ncome, the hgher the respectve Gn coeffcent. The next queston to be addressed s whether farm ncome rases or reduces the overall ncome nequalty n each type of household. One would expect that the contrbuton of farm ncome to the reducton of nequalty s postvely correlated to ts share n the total ncome of varous types of households. Table 8 shows that three man fndngs can be drawn. Frstly, FHs exhbt the hghest nequalty on the bass of 10

11 equvalent household ncome. Secondly, non-farm ncome s dstrbuted more unequally wthn FHs than n any other type of households. Thrdly, both farm ncome and non-farm ncome are dstrbuted unequally wthn each type of household; for example, wthn FHs, Gn coeffcents are and 0.417, respectvely. Nonetheless, the sum of those two components of ncome, that s total ncome of FHs, s dstrbuted less unequally as s evdent from ts Gn coeffcent (0.339). Ths s due to the fact that there are FHs wth ether low farm ncome coupled wth hgh non-farm ncome or vce versa. Thus, addng farm ncome to non-farm ncome reduces nequalty from to n FHs, whereas the same coeffcent s reduced from to n the case of PFHs. Consequently, farm ncome and non-farm ncome complement each other, generatng a combned stablzaton effect at the level of the total household ncome, the effect beng analogous to farm ncome share n total household ncome. Although someone could clam that ths fndng could not be attrbuted to farm ncome alone as t also holds for other types of ncome, t nevertheless has an obvous sgnfcance for the subject examned here. Table 8: Income nequalty ndexes (Gn Coeffcent) by type of household Equvalent Income (1) Equvalent Non- Income (2) Equvalent Household Income (3) Households (FHs) Plur-actve Households (PFHs) Margnal Households (MFHs) Retred Households (RFHs) Non- Households (NFHs) All Households Decomposton of nequalty The decomposton of nequalty ndces s another valuable attrbute that enables the estmaton of each component s contrbuton to the overall nequalty. Of partcular mportance s the decomposton of the total nequalty n wthn-groups and between-groups nequalty. After splttng our populaton n the abovementoned fve groups of households, by usng the same crtera as before, the wthn-groups component accounts for almost 88%-90% of the overall nequalty, whereas the between-groups share s 10% to 12% (table 9). Ths holds rrespectve of the ndex (T or N) and the dstrbuton of nequalty employed - per capta equvalent expendture or per capta equvalent ncome. Table 9: Decomposton of nequalty Components of Inequalty based on: Per capta equvalent Per capta equvalent expendture ncome 11

12 Thel Index (T ) Mean Log Devaton (N) Thel Index (T ) Mean Log Devaton (N) Wthn-groups 90.1% 88.2% 87.9% 90.2% Between-groups 9.9% 11.8% 12.1% 9.8% 4.4 Poverty RFHs are by far the poorest households n the total populaton (table 10). In most cases of the dstrbutons employed, poverty rates n RFHs are more than double the respectve rates n the other types of households. Takng as an exemplary case the 60% poverty lne, the most commonly used threshold for poverty, 56.5 percent or 62.9 percent of RFHs fall below ths lne, whle the respectve rates for the total populaton are 17.1 and RFHs are one of the poorest socal categores, not only among pensoners but also wthn the total Greek populaton. From households wth an economcally actve head, FHs present the hghest poverty rates. At least one quarter of FHs are classfed as poor f poverty s set at 60% of the mean equvalent expendture or the mean equvalent ncome. These results confrm prevous fndngs on poverty n Greece (Mtrakos and Sarrs 2003; Tsakloglou and Panopoulou, 1998). For PFHs and MFHs poverty rates seem to be close to, or hgher than, the allhousehold average. Also, n all poverty lnes, non-farm households have a slghtly lower poverty ncdence than all-households. Table 10: Poverty rates (percentage of households below the poverty lnes) Poverty lne (% of mean equvalent expendture) 40% 50% 60% 70% Households (FHs) Plur-actve Households (PFHs) Margnal Households (MFHs) Retred Households (RFHs) Non- Households (NFHs) All Households Poverty lne (% of mean equvalent ncome) 40% 50% 60% 70% Households (FHs) Plur-actve Households (PFHs) Margnal Households (MFHs) Retred Households (RFHs) Non- Households (NFHs) All Households

13 5. Conclusons Ths paper has examned the ncdence of ncome nequalty and poverty as well as the mpact of farm ncome on nequalty and poverty n Greece, usng data from the most recent Household Budget Survey. A typology of households was developed wth a specal focus on both the agrcultural holdng operated by a household and the share of farm ncome n the composton of household ncome. The results show large varatons among households wth regard to composton of ther total ncome. The share of farm ncome n total ncome vares from 55.0% n FHs and 23.0% n PFHs to 6.7% n MFHs and 3.5% n RFHs. Thus, even n farm households, a sgnfcant part of ther total ncome s derved from nonagrcultural sources. It has been argued that the farm problem n the sense of ncome gap between agrculture and other actvtes does not exst any more (Gardner, 1992). The dsaggregated approach to household structure adopted here allows testng the valdty of ths argument. As we have seen, two household types, margnal- and pluractve- farm households, do not seem to have a real ncome problem. Whether n terms of equvalent expendture or n terms of equvalent ncome, they seem to compare favourably wth the rest of socety. Nevertheless, ths s not the case for farm households, whch, as a group, stll lag behnd the all-household average ncome, though they exhbt convergng trends over the last few years. The case of retred farm households, whch have almost half the average natonal ncome and consttute one of the poorest socal categores of the entre populaton, s somewhat dfferent. ncome presents a stable share n total ncome of farm households over tme (Karankolas et al, 1998). Ths way, t contrbutes to the reducton of temporal varaton of household ncome, as well as to the convergence of ncome levels between farm and non-farm households. However, the larger the share of farm ncome to total ncome of a household type, the more t augments ncome nequalty wthn these households. Ths observaton holds more n the case of subsdes than farm ncome derved from market. The share of farm ncome n household ncome s also postvely correlated wth the possblty of ths household havng a low ncome. Almost three quarters of RFHs are found n the three lowest decles, a clear ndcaton of the especally low ncome level of ths type of household. As far as the concentraton of farm ncome across the band of the ncome dstrbuton s concerned, the absolute value of farm ncome and subsdes per decle s qute dssmlar, especally n the case of farm households. Yet, the relatve contrbuton of farm ncome to total household ncome vares, though wthout followng a clear pattern across decles. Furthermore, farm household nequalty s the hghest among all types of agrcultural households, whereas retred farm households exhbt the lowest nequalty. Research fndngs ndcate that, although every component of household ncome s very unequally dstrbuted, the dstrbuton of ther sum total ncome of household s less unequal. Consequently, farm ncome and non-farm ncome complement each other, generatng a combned stablzaton effect at the level of the total household ncome, the effect beng analogous to the share of farm ncome n total household ncome. If ths s also true for other types of ncome, even so, t s an nterestng fndng for our analyss. Besdes, decomposton of nequalty reveals that the wthn-groups component accounts for almost 88%-90% of the overall nequalty, whereas the between-groups share s 10% to 12%. Ths means that the elmnaton of all expendture or ncome 13

14 dfferences between groups of households wll not reduce total nequalty by more than 12%. Accordngly, polces amng at the reducton of nequaltes wthn each household category wll contrbute much more to the reducton of total nequaltes. Thus, prevous fndngs on the decomposton of nequalty seem to be confrmed, though those prevous studes had used dfferent crtera for the groupng of the populaton (Mtrakos and Sarrs, 2003). Poverty s a major characterstc of some types of households. Retred farm households are one of the poorest socal categores, not only among pensoners, but also wthn the total Greek populaton, wth 56% of them fallng below the 60% poverty lne. From households wth an economcally actve head, farm households present the hghest poverty rates. At least one quarter of FHs are classfed as poor f poverty s set at 60% of the mean equvalent expendture or the mean equvalent ncome. These results are n lne wth prevous fndngs on poverty n Greece. Therefore, those results rase serous dstrbutonal and re-dstrbutonal ssues and also have some obvous polcy mplcatons. If welfare consderatons, along wth publc goods provson from agrculture, consttute a sound legtmatng bass for agrcultural polcy n the near future, the abovementoned fndngs could prove useful. It s known that CAP has always had an explct welfare aspect expressed through the objectve of a far standard of lvng of the agrcultural communty. On the other hand, welfare and dstrbutonal consderatons have been at the forefront of crtcsms aganst CAP. Ever snce the late 1980 s, unequal dstrbuton of agrcultural subsdes and ncome dspartes wthn European agrculture have been used as a legtmatve bass for radcal CAP reforms. In ths context, ncreased effectveness has been pursued through decouplng of payments to farmers, and an overall reducton n support and protecton of European agrculture. Payments decoupled from the producton and delvered on the bass of hstorcal enttlements of farmers allegedly address more effectvely the problems related wth ncome dspartes between farmers, types of producton systems, etc. Our results ndcate that the unequal dstrbuton of farm ncome or subsdes s not a problem per se. At least from a welfare and dstrbutonal pont of vew, the combnaton of farm and non-farm ncome could lessen household ncome nequaltes, pontng to an addtonal postve role of farm ncome and ts consttuent parts. Of course the pursut of the hghest degree of effectveness of agrcultural polcy, as well as the choce of the polcy nstruments most approprate to fulfll ts goals, reman open. Fnally, the hgh rates of poverty ncdence among retred farm households mples that a socal polcy amng at the reducton of poverty n the whole populaton could beneft by targetng ths specfc category of households. Further research s needed to extend the abovementoned fndngs n a longer tme perod especally before and after the major CAP reform of Even more nterestng s a smlar examnaton henceforth, so that the most recent revew of CAP (n ) wll be taken nto consderaton. References Commns, P. (2004), Poverty and Socal Excluson n Rural Areas: Characterstcs, Processes and Research Issues, Socologa Rurals, 44 (1):

15 Eurostat (2002), Income of the Agrcultural Household Sector, 2001 Report, Luxemburg: Eurostat. Eurostat (2007), Structure Indcators, structures. Fennell, R. (1997), The Common Agrcultural Polcy - Contnuty and Change, Oxford: Clarendon Press. Frawley, J., Commns, P., Scott, S. and F. Trace (2000), Low Income Households Incdence, Characterstcs and Polces, Dubln: Oak Tree Press. Gardner, B. (1992), Changng economc perspectves on the farm problem, Journal of Economc Lterature, XXX (March 1992): Hagenaars, A. J. M., de Vos, K. and Zad, M. A. (1994), Poverty Statstcs n the Late 1980s: Research based on Mcro-Data, Theme 3, Seres C, Luxembourg: Eurostat. Hll, B. (1999), Household Incomes: Perceptons and Statstcs, Journal of Rural Studes, 15 (3): Hll, B. (2000), ncomes, wealth and agrcultural polcy, 3 rd Edton, Aldershot: Ashgate. Karankolas, P., Zografaks, St. and N. Martnos (forthcomng), Income and welfare dspartes of farm and non-farm households, n Loulouds, L. and K. Krmpas (Eds), Agrcultural Polcy, Athens: Athens Academy, [n Greek]. Mshra, A. and C. Sandretto (2002), Stablty of Income and the Role of Nonfarm Income n US Agrculture, Revew of Agrcultural Economcs, 24 (1): Mtrakos T. and P. Tsakloglou (1998), Decomposng nequalty under alternatve concepts of resources: Greece 1988, Journal of Income Dstrbuton, 8 (3): Mtrakos, T. and A. Sarrs (2003), Characterstcs and Structure of Greek Households; Incomes, Consumpton Patterns and Dmensons of Inequalty and Poverty, n Sarrs, A. (Ed.), Towards a Development Strategy of Greek Agrcultural Sector, pp , Athens: Mnstry of Agrculture and Athens Unversty [n Greek]. Mtrakos T. and P. Tsakloglou (2003), Economc Inequalty and Poverty n Greece: Structure and Inter-temporal Change, n Veners, D. and Ch. Papatheodorou (Eds), Socal Polcy n Greece: Challenges and Perspectves, pp , Athens: Hellenka Grammata [n Greek]. Mtrakos T. and P. Tsakloglou (2006), Inequalty and poverty n the last quarter of the 20th century, n M. Petmesdou and E. Mossalos (eds) Socal polcy development n Greece, Ashgate Publshng Lmted, Chapter 6, pp , 2006, Aldershot: Ashgate. OECD (1987), Natonal Polces and Agrcultural Trade, Pars. OECD (2004), Household Income: Towards Better Informed Polces, OECD Observer, October Pauw, K.W. (2007), Agrculture and poverty: ng for food or farmng for money? Agrekon, 46 (2): Sarrs, A. and S. Zografaks, (1996), Agrcultural ncome statstcs and polcy: a vew from southern Europe, n Hll,B. (Ed.) Income statstcs for the agrcultural household sector, pp , Luxembourg: European Commsson, Eurostat. 15

16 Schmtt, G. and C. Boruse (1996), What about the ncome problem of agrculture n developed economes?, n Hll, B. (Ed.), Income Statstcs for the Agrcultural Household Sector, pp , Luxemburg: European Commsson, Eurostat. Tsakloglou P. και Panopoulou G. (1998), Who are the poor n Greece? Analysng poverty under alternatve concepts of resources and equvalence scales, Journal of European Socal Polcy, 8 (3): ANNEX INEQUALITY AND POVERTY INDICES Varables: Ε= Total per capta expendture of all households Ε = Total per capta expendture of household Η = Number of households n HBS h = Number of household s members π = poverty lne Ν = Number of HBS households members Ν p = Number of poor people n HBS n p = Number of poor households n HBS Households have been ranked by per capta total expendture n ascendng order E E... E E... E 1 2 n n 1 H p p where π s the poverty lne. The above nequalty automatcally denotes the number of poor households n p Varable E s defned as: E Indexes: H H h E h H h E N =1,2,3,., Η for all households n the sample 16

17 j=1,2,3,., n p for all poor households n the sample TH Thel Index (TI): H 1 h N E E Ln E E Mean Log Devaton (MLD): H 1 1 MLD Ln h E N N H 1 1 h Ln E Where the number of poor households n p s defned by (1.2) (1.4) E n p 1 2 n p 1 h E N Gn Coeffcent (G): E n p 1 N G x x x y x y (1.7) where: x j 1 H j 1 h h j j, y j 1 H j 1 h E j h E j j j, 17

18 ANNEX TABLE 1: Mean Monthly Incomes by type of household and decle ( ) All Househ olds Househ olds (FH) Pluractve Househol ds (PFH) Margnal Households (MFH) Retred Househol ds (RFH) Non- Househ olds Decle 1 non-farm ncome farm ncome wthout subsdes subsdes Decle 2 non-farm ncome farm ncome wthout subsdes subsdes Decle 3 non-farm ncome farm ncome wthout subsdes subsdes Decle 4 non-farm ncome farm ncome wthout subsdes subsdes Decle 5 non-farm ncome farm ncome wthout subsdes subsdes Decle 6 non-farm ncome farm ncome wthout subsdes subsdes Decle 7 non-farm ncome farm ncome wthout subsdes subsdes Decle 8 non-farm ncome farm ncome wthout subsdes subsdes Decle 9 non-farm ncome farm ncome wthout subsdes subsdes Decle 10 non-farm ncome farm ncome wthout subsdes subsdes