Measurement of Poverty Intensity in Khuzestan Province During

Size: px

Start display at page:

Download "Measurement of Poverty Intensity in Khuzestan Province During"

Melina Sharp
5 years ago
Views:

1 Measuremet Quarterly of Joural Poverty of Itesity Quatitative Ecoomics, Summer 2009, 6(2): -26 Measuremet of Poverty Itesity i Khuzesta Provice Durig Seyyed Mortaza Afgheh (Ph.D.) ad Talea Ghaavatifat (M.Sc.) Received: 2009/5/30 Accepted: 2009/9/5 Abstract: Cosiderig the importace of poverty i ecoomic developmet issues, it is tried to estimate poverty itesity ad examie its tred via usig a applied idex called SST. This idex belogs to the family of decomposable idices ad shows the poverty itesity i terms of the product of its compoets: poverty rate, average poverty gap ratio of the poor ad Gii idex of poverty gap ratio of the populatio. This decompositio makes possible for the researcher to follow the source of chages i poverty itesity via chages i its compoets. The results reveal that the SST idex has icreased durig both i urba ad rural areas of the provice. Also, the icrease i the urba areas, is much more tha the icrease i rural areas. JEL classificatio: C02 Keywords: Poverty idex, poverty itesity * Assistat professor ad graduate studet of ecoomics at Shahid Chamra Uiversity, respectively, Ahvaz, Ira. (smafghah@yahoo.com)

2 2 Quarterly Joural of Quatitative Ecoomics 6 (2), Summer Itroductio Poverty has bee the mai issue for may social scietists ad iteratioal orgaizatio over last decades. May developed ad developig coutries have itroduced ati-poverty strategies followed by comprehesive research ad studies. Supported by World Bak, Uited Natios ad Iteratioal Moetary Fud, these coutries have implemeted variety of activities to tackle poverty. Gathered i 2000, 89 world leaders cocetrated o poverty elimiatio ad supportig huma rights for world coutries i ew milleium that led to publishig Milleium Developmet Goals (MDG). They itroduced a eight goal program for world societies to be performed i followig years. The first goal of the program is Eradicatio of extreme poverty ad huger that advised coutries to perform measures to half umber of people with less tha $ per day ad those sufferig from starvatio by 205. Cocetratig o poverty eradicatio ad icreasigly welfare level, policy makers i Ira promoted several socio-ecoomic plas. Furthermore, various academic ad official studies have bee performed. Some of these studies tackled poverty lie ad poverty idices as a whole while others cocetrated o some particular provices. Calculatig poverty lie i all provices helps policy makers to allocate resources i more efficiet coditios specially cocer with the regios that are socioecoomically i critical coditios. I this paper, thus, it is tried to calculate poverty lie ad poverty itesity i Khuzesta provice. Similar to may related studies, to estimate poverty itesity, it was ecessary to calculate poverty lie. The ext step was measurig poverty idex based o two mai features: ) beig justifiable i theory; ad 2) beig uderstadable by policy makers. To meet these two features, therefore, SST approach was employed to estimate poverty itesity.

3 Measuremet of Poverty Itesity 3 2. Research Hypotheses: - It seems that poverty itesity i Khuzesta provice (both urba ad rural areas) has icreased durig 997 to It seems that poverty itesity i urba areas of Khuzesta provice has vastly icreased compared with rural areas i the research period. 3. Ideal characteristics for a poverty itesity Idex A ideal poverty idex should be symmetrically replicatio. That is, the value of the poverty idex for the combiatio of two idetical populatios is the same as the value of the poverty idex for each of the two populatios. A ideal poverty idex should also be a cotiuous fuctio of idividual icomes. Furthermore, it takes a higher value whe a trasfer happes from a poor idividual to a rich oe. Regardig these characteristics, Shorrocks (995) suggested a modified Se Idex to estimate the poverty itesity. Zheg (997) argued that this modificatio is idetical to the limit of Tho s modified Se Idex. Thus, this idex is called Se-Shorrocks-Tho or SST idex (Xu, 998). The SST idex is symmetrically replicatio, cotiuous i idividual icomes, homogeeous of degree zero i idividual icomes ad relative poverty lie. It is cosistet with the trasfer axiom (that a acceptable measure of poverty should always icrease if resources are take from poor people ad give to richer oes). It is covetioal for the values of a idex ragig betwee zero ad oe to have a logical geometric iterpretatio (Xu, 998). 4. SST Idex ad It s Decompositio The SST idex ad its compoets are very useful i ecoomic studies but the relatio betwee this idex ad its compoets should be explaied i more details. If the average poverty gap ratios ad the Gii idex of poverty gap ratios are ot estimated properly, the results caot be reliable ad most probably are ot efficiet (Xu ad Osberg, 200).

4 4 Quarterly Joural of Quatitative Ecoomics 6 (2), Summer 2009 I most studies, it is argued that the poverty rate (Head-Cout ratio) ad the average poverty gap ratio of the poor (the icome gap ratio) violate some importat axioms (trasfer axiom i particular) of a ideal poverty idex. Clearly, these axioms play essetial roles i poverty measuremets. Although either poverty rate ad poverty gap is separately a suitable measuremet for poverty evaluatio, both together, are capable of explaiig poverty level. Based o SST decompositio performed by Osberg-Xu ad empirical ad regioal evidece durig 997, 998 ad 2000, cumulative percetage of chages i poverty gap ad average poverty gap ratios reflect almost all of the percetage chages i poverty itesity over time (Osberg ad Xu, 200). Due to the importace of SST idex ad its decompositio, it is ecessary to explai them i more details. Suppose icomes ( y s) i for a populatio of size sorted ascedig so that y y2... y. If the poverty lie of the populatio is cosidered as z ad the umber of people whose icome are lower tha poverty lie as q, so for the i th perso ( i =,..., ) the poverty gap ratio ( x i ) is defied as: z yi z yi xi = if 0 () z z z yi x i = 0 if 0 (2) z As is see, the poverty gap ratio is zero for those who are ot poor. Shorrocks (995), the, suggested SST idex of poverty itesity as: z yi P( y; = (2 2i + ) (3) 2 i= z or P ( y; = ( 2 2i + ) x 2 i. (4) i= Here p ( y; is the SST idex. Also, Shorrocks (995) idicates the followig: P ( y; = µ ( x)( + G( x)) (5)

5 Measuremet of Poverty Itesity 5 where µ (x) ad G (x) are the average of poverty gap ratio ad the Gii idex, respectively, for the distributio of poverty gap ratios. The average of poverty gap ratios is defied by: µ ( x) = x i (6) i= Equatio (6) ca be broke dow ito poverty rate q H = (7) ad average poverty gap ratio of the poor (icome gap ratio): I = x i (8) q i= Furthermore, Shorrocks has employed equality of poverty gap ratios of populatio ( + G( x) ) to estimate SST idex. Here G(x) is defied as: i i G( x) = or i= ( ) j= x j + i 2 j= 2 µ ( x) x j (2 2i + ) x 2 i µ ( x) i= G( x) = (0) that x i s are sorted ascedig like x x2... x. Cosequetly, the equatio (5) ca be developed as: P ( y; = HI( + G( x)) () Thus, the poverty itesity cotais the poverty rate (H ), the average poverty gap ratio of the poor (I), ad the measure of the iequality of the poverty gap ratios of the populatio ( + G( x)). This decompositio eables ecoomists to aalyze the right sources of a chage i the value of poverty itesity based o chages i each of these compoets over time. Thus, takig the atural logarithm of both sides of equatio () results: LP ( y; = LH + LI + L( + G( x)) (2) (9)

6 6 Quarterly Joural of Quatitative Ecoomics 6 (2), Summer 2009 where L ( + G( x)) is a approximatio of G(x) based o the first-order Taylor series expasio. Let A = A A, where A is the amout of A at the previous period. Rewritig the equatio (2), we have: LP ( y; = LH + LI + L( + G( x)) (3) where L ( + G( x)) is a approximatio of G(x). Equatio (3) shows that the chages percetage i p( y; is the sum of the chages percetage i H, I ad G (x) (Xu ad Osberg, 200). 5. Literature review A review o iteral issues shows a wide rage of academic studies o topic. Bagheri ad Kavad (2006), for istace, have used SST idex to estimate poverty lie i Ira for years 2003 ad Give a daily cosumptio of at least 2300 calories, they have calculated a poverty lie for both urba ad rural areas. Usig data from Statistics Ceter of Ira ad based o miimum ecessary calories for a perso (2300), they first estimated poverty lie for urba ad rural areas, the, measured SST idex ad estimated its three compoets, i.e. poverty rate, average poverty gap ratio of the poor ad Gii idex of poverty gap ratio of the populatio. The results revealed that poverty itesity icreased from 5.9% i 2003 to 7% i 2004 that is 7% growth i poverty itesity i two years. SST compoets reveal that poverty rate had the mai effect o the calculated poverty itesity. I rural areas, however, poverty itesity decreased from 4.9% i 2003 to 4% i The reductio of poverty itesity i rural areas was due to the decrease i both poverty rate ad poverty gap ratio of the poor. There are also plety of studies o topic at iteratioal level. Xu (998), for istace, has examied poverty itesity usig SST idex for years 969, 979 ad 988 i the Uited States. First, he reviewed theoretical cotext of SST idex, ad cocluded that this idex had bee developed durig previous decades especially due to the features that a ideal poverty idex has to have. He also suggested a useful geometric iterpretatio.

7 Measuremet of Poverty Itesity 7 Furthermore, he accepted that this idex was symmetric, mootoic, cotiuous, homogeeous of zero degree i icomes ad poverty lie ad cosistet with the trasfer axiom. I his research, Xu, employed per capita icome (before taxpayig) ad Pael Study of Icome Dyamics (PSID) of the Uited States of America data to calculate SST idex. Xu s research is also based o poverty lie that was calculated by Smeedig (99), though he devoted parts of his research to estimate poverty lie. Based o Smeedig approach, the poverty lie is the same as 50% of icome media. Thus, Xu computed SST idex for years 996, 979 ad 988 that were 2.0%, 3.35% ad 5.57% respectively. As is show, the poverty itesity i the Uited State has bee icreased durig the research period. Osberg ad Xu (999) used half the media of a icome as relative poverty lie ad estimated the Se-Shorrocks-Tho measure of poverty itesity i Caadia provices for 984, 989 ad The outcomes show that poverty itesity decreased i Otario i late 980 to the level that is i North Europe, while icreased sigificatly i 994. This chage was due to a developmet i govermet social security supports i 980 ad a cut i 994 respectively. However, Price Edward Islad had a better performace i poverty itesity reductio. At the atioal level, though, the poverty itesity decreased durig 980 decade but icreased agai from 994 owards. I their case study about Chia, Osberg ad Xu (2008), preferred to use relative poverty tool like 50% of icome media compared with absolute poverty criteria, i.e. $ per day per perso. They, the, employed Chiese Household Icome Projected (CHIP) data to calculate SST idex i 995. They focused o rural areas of some provices. They disregarded Beijig due to lack of eough rural populatio. The results showed vast chages i SST idex ad its three compoets. The poverty rate i rural areas chaged betwee 6.9% ad 9.7%. The scope of average chages i poverty gap was betwee 38.9 ad 7 percet. This large variatio is comparable with the

8 8 Quarterly Joural of Quatitative Ecoomics 6 (2), Summer 2009 variatio i poverty gap that is betwee.567 ad.962. However, based o Gii idex plus oe across the rural areas of Chiese provices the variatio was relatively large compared with similar data observed i developed coutries, while was small if poverty rate or poverty gap is based for compariso. 6. Iterpretatio of SST Idex ad its compoets The scope of SST idex is ragig betwee zero ad oe. Larger values idicate more poverty itesity ad vice versa. 6.. Poverty Rate: This idex is also betwee zero people oe. Zero idicates a coditio i which there are o people i society ad o the opposite side oe meas that all people i the society are poor. Therefore, more poverty rate meas there should be more poor people i the society Average poverty gap ratio of the poor: Average poverty gap ratio of the poor reflects the depth of poverty i society. The larger value of this idex idicates more poverty i the society Gii Idex for the poverty gap ratio: This Gii idex measures iequality amog poor ad, like other idices i this research, is betwee zero ad oe. Zero reflects a coditio of full equality amog poor while oe shows maximum iequality amog poor. 7. Method of estimatig SST Idex The ecessary data -icome ad expediture of rural ad urba households- were collected from Statistics Ceter of Ira. To estimate poverty itesity, the absolute poverty lie is employed that is defied as the miimum level of calorie per day per perso i Ira (2300). Furthermore, to calculate SST compoets, total household expeditures are employed. The area of this research is rural ad urba households i Khuzesta Provice for the period of 997 to 2006.

9 Measuremet of Poverty Itesity 9 To start estimatig poverty itesity idex, the total aual expeditures of the sample were divided by the dimesio of household to calculate per capita expediture. The outcomes, the, were divided by 2 (moths) to reach total mothly per capita expediture. Now the computed poverty lie could be compared with the household expeditures ad thus poverty rate ca be estimated. Applyig the formula explaied i sectio four, the average poverty gap ratio of the poor ad Gii idex for populatio are the estimated. Havig computed all related compoets, i.e. average poverty gap ratio, poverty rate ad the Gii plus oe idex of poverty gap ratio of populatio, the poverty itesity was estimated. The outcomes are revealed i tables ad Results ad iterpretatio of urba areas of Khuzesta provice As is show i table, poverty itesity i urba areas has icreased from 3.35% i 997 to 7.2% i 998. A review o SST compoets reveals that the icrease i poverty itesity is due mostly to a icrease i poverty lie; from 6.95% to 3.48% ad to poverty gap, from 24.6% to 27.94%. I other words, the 66.24% icrease i poverty rate ad 2.73% icrease i poverty rate ad 2.73% icrease i poverty gap are accoutable for the icrease i poverty itesity. However, the Gii plus oe idex has a small decrease from.95 to.9 showig that this idex

10 20 Quarterly Joural of Quatitative Ecoomics 6 (2), Summer 2009 Table : SST idex ad its compoets, percetage chages i SST idex Based o: Researcher s Calculatios Urba SST SST s Compoets Chages i SST idex ad it s compoets H I +G(x) Δl(SST) Δl(H) Δl(I) Δl(+G(x)) has had o sigificat effect o SST idex durig 997 ad 998. A compariso betwee years 999 ad 998 shows a mior reductio i poverty itesity from 7.2% to 6.9% that seems to be mostly due to a decrease i poverty gap from 27.94% to 24.33% ad the mior decrease i oe plus Gii idex meas that its impact o poverty itesity ca be igored. However, the small icrease i poverty rate from 3.48% to 4.85% reveals that it has o effect o poverty itesity durig the two years. The data for years 2000 to 2003 reveals a icreasig tred i poverty itesity. The most resposible factor for this tred is poverty rate as is show i table. However, the reductio i the oe plus Gii idex has o sigificat effect o the icrease of poverty itesity. Decrease i SST idex from 22.86% i 2004 to 7.95% i 2005 is seemigly a reflectio of the reductio i both poverty rate ad average poverty gap; i.e. from 40.77% to 37.78% ad from 32.8% to 26.94% respectively. The equality idex, agai, has o sigificat effect o poverty itesity reductio. Fially, the SST idex i 2006 reveals a relatively sharp icrease compared with SST idex i 2005 from 7.95% to 32.46%. Icrease i both poverty rate (from 37.78% to 52.5%)

11 Measuremet of Poverty Itesity 2 ad poverty gap ratio of the poor (from 26.94% to 37.7%) are the mai reaso for the SST icrease. There is also a 7.25% reductio i the oe plus Gii idex of poverty gap of populatio that shows o impact o SST icrease. 9. Results ad iterpretatio of rural areas of Khuzesta provice As is show i table 2, the icrease i poverty itesity from 4.35% i 997 to 30.24% i 998 is mostly due to icrease i both poverty rate ad average poverty gap ratio of the poor. However, Gii idex plus oe has o effect o poverty itesity icrease. A compariso betwee years 999 ad 998 reveals a sharp decrease i poverty itesity that seems to be a reflectio of a distict drop i both poverty rate ad average poverty gap ratio (from 52.3% to 09.9% ad from 35.52% to 7.6% respectively). Agai, the oe plus Gii idex has o impact o SST reductio. The results from years 2000 to 2006 show a zigzag movemet i SST idex ad its compoets. A mior drop i 200 compared with 2000 followed by a upward tred betwee years 2000 to 2004 the a sharp drop i 2005 (from 2.97% i 2004 to 8.4% i 2005) ad evetually small icrease i 2006 shows a relatively fluctuate tred i SST idex as metioed above.

12 22 Quarterly Joural of Quatitative Ecoomics 6 (2), Summer 2009 Table 2: SST idex ad its compoets, Percetage chages i SST idex Based o: Researcher s Calculatios Rural SST SST s compoets Chages i SST idex ad it s compoets H I +G(x) Δl(SST) Δl(H) Δl(I) Δl(+G(x)) Geometric Iterpretatio of SST Idex I this sectio, a example is applied to explai how to decompose SST idex ad how it ca be preseted via geometric iterpretatio. Suppose we have the followig data sorted i ascedig order, i.e. y = 3, y = 9, 2 y = 3 ad y = 4 5 ad poverty lie z is supposed to be 0 i this example. The computed poverty gap usig equatios () ad (2) are x = 0. 7, x 2 = 0., x = 0ad x = Thus, the SST idex is computed as follows: P( y; = (2 2i + ) x 2 i i= = (7(0.7) + 5(0.)) 6 = Like the Gii idex, the SST idex also has a simple geometric iterpretatio. The deprivatio profile is a fuctio of k / for a give data series of x : k D( x ; ) = k i= x i

13 Measuremet of Poverty Itesity 23 Where x refers to the sequece of x i s arraged i descedig order. Figure : The Deprivatio Profile Based o: Xu ad Osberg (200) As is show i Figure, the deprivatio profile curve starts from the origi ad reaches H, the it becomes horizotal. The poit H idicates the poverty rate ad the poit HI shows the average poverty gap ratio of the populatio. The average poverty gap ratio of the poor, I, is represeted by the slope of dotted lie O H because HI / H = I. The arc O H is similar to the Lorez curve. The degree that the arc O H deviates from the dotted O H implies the degree of iequality of deprivatio values. Therefore, the SST idex ca be show as the ratio of the area uder poverty gap profile to the area uder the lie of maximum poverty (the 45 degree lie) as is show i Figure. For the same data set, the poverty rate is H = 0. 5, the average poverty gap ratio of the poor is I = 0. 4, ad the average poverty gap ratio of the populatio is µ ( x) = Based o the equatio (9), the Gii idex is give by:

14 24 Quarterly Joural of Quatitative Ecoomics 6 (2), Summer 2009 i i xj + xj j= i 2 j= 2 G( x) = ( ) i= µ ( x) = ( ) = So P ( y; = HI( + G( x)) = ( 0.2)( ) = As equatios (4) ad () reveal, the results are idetical. Thus, the SST idex computed directly from equatio (4) is equivalet to that computed idirectly by the product of its three decomposed compoets (Xu ad Osberg, 200).. Coclusio As the mai socio-ecoomic issue i Ira durig last decades, poverty ad related issues have bee at the core of may academic studies. Not oly poverty at the atioal level is a importat issue but also differeces betwee the scope ad depth of poverty i various provices are of cocer i may academic studies. I this paper, we tried to compute poverty itesity usig SST approach that seems to be the most reliable approach by ow. The results from the data aalyzig i this paper reveal that the poverty itesity i urba areas of Khuzesta provice has had a sharp icrease from 3.35% i 997 to 32.46% i Thus, the first hypothesis of the preset research is ot rejected. To examie the reaso for such drastic icrease, the compoets of poverty itesity idex are also calculated. The outcomes show that the icrease of poverty rate from 6.95% i 997 to 52.5% i 2006 i the oe had ad the growth of poverty gap ratio of the poor from 24.6% i 997 to 37.7% i 2006, o

15 Measuremet of Poverty Itesity 25 the other had are the most effective factors for SST icrease durig the study period. However, the declie of the oe plus Gii idex from.9866 to shows that this idex has had o sigificat role o the SST idex icreases durig study period. A quick review of the computed data shows that the maximum value of urba poverty itesity is 32.46% i 2006 compared with the previous year (2005) that is 7.95% ad shows a 80% growth. Like urba areas (though lesser), the amout of SST idex i rural areas has icreased from 4.33% i 997 to 7.% i 2006, idicatig that the first hypothesis is ot rejected. Furthermore, the SST i rural areas durig the study period shows lesser chages compared with urba area that is 9% growth i rural SST idex i the same period. Thus, the secod hypothesis is ot rejected. A review o the compoets of rural SST idex reveals that both poverty rate ad poverty gap ratio of the poor have icreased from 30.93% to 36.4% ad from 25.76% to 26.76% respectively) durig the study period ad thus are resposible for the metioed icrease i SST idex. Like previous aalysis, however, the decrease i the oe plus Gii idex durig the study period shows that this idex has had o effect o the SST idex chages. The highest value of SST idex durig the study period 30.24% i 998 shows that, agai, it is affected by icrease i both poverty rate ad poverty gap ratio i 998 compared with 997. I geeral, the research results show deterioratio i socioecoomic coditio of Khuzesta provice populatio durig the study period. Due to ecoomic coditio of Khuzesta provice that has seemigly attracted most of the ivestmets i four ecoomic sectios (agriculture, idustry, service ad oil), the above results seem to be somehow iexplicable. Thus, more studies i this area are eeded to examie the reaso(s) for this dilemma coditio.

16 26 Quarterly Joural of Quatitative Ecoomics 6 (2), Summer 2009 Referece: Bagheri, F & H. Kavad. (2006). Measuremet of Poverty Itesity i Ira: The Applicatio of SST Idex. Scietific Research Quarterly of Social Welfare, Fifth Year, 20: Osberg, L. & K. Xu. (999). Poverty Itesity: How well do Caadia Provices Compare?. Caadia Public Policy, 25: 2-0. Osberg, L. & K. Xu. (2008). How Should We Measure Poverty i a Chagig World? Methodological Issues ad Chiese Case Study. Review of Developmet Ecoomics, 2: Shorrocks, A.F. (995). Revisitig the Se Poverty Idex. Ecoometrica, 63: Statistics Ceter of Ira. (2006). Icome ad Expediture of Rural ad Urba Households of Khuzesta Provice. ( ). Smeedig, T.M. (99). Cross-Natioal Comparisos of Iequality ad Poverty Positio. i L., Osberg (ed), Ecoomic Iequality ad Poverty: Iteratioal Perspective. New York: M.E. Sharpe, Ic. Tho, D. (979). O Measurig Poverty. Review of Icome ad Wealth, 25: Xu, K. (998). Statistical iferece for The Se-Shorrocks-Tho Idex of Poverty Itesity. Joural of Icome Distributio, 8: Xu, K. & L. Osberg. (200). O Se s Approach to Poverty Measures ad Recet Developmets. Chia Ecoomic Quarterly, (): Xu, K. & L. Osberg. (200). How to Decompose the Se-Shorrocks-Tho Poverty Idex: A Practitioer s Guide. Joural of Icome Distributio, 0: Xu, K. & L. Osberg. (2002). The Social Welfare Implicatios, Decomposability, ad Geometry of the Se Family of Poverty idices. Caadia Joural of Ecoomics, 35(): Zheg, B. (997). Aggregate Poverty Measures. Joural of Ecoomic Surveys, : 23-6.

An Empirical Study on the Contribution of Foreign Trade to the Economic Growth of Jiangxi Province, China

An Empirical Study on the Contribution of Foreign Trade to the Economic Growth of Jiangxi Province, China usiess, 21, 2, 183-187 doi:1.4236/ib.21.2222 Published Olie Jue 21 (http://www.scirp.org/joural/ib) 183 A Empirical Study o the Cotributio of Foreig Trade to the Ecoomic Growth of Jiagxi Provice, Chia