Word Pro - SA - Title page
Social Limitation, Crime, and Perceived Quality of Life Patterns in South Africa
Nattavudh Powdthavee* University of Warwick, January, 2003
Abstract
Is happiness pattern structurally the same in the poor countries as they are in the rich countries? Using a cross-sectional data from SALDRU93 survey, we show that relationships between subjective well-being and socioeconomic variables have a similar structure in South Africa as they are in the developed countries. The significance in correlations can vary between-group of social norms, however, if the discrepancies in norm characteristics are great. In addition, basic living standard indicators such as durable goods correlate positively and significantly with the reported level of life satisfaction. Crime committed on an individual, on the other hand, is associated with lower well-being in general. Lastly, the perceived quality of life is perhaps partly relative in many respects: owning a motor vehicle is correlated with higher satisfaction levels, but the positive effect reduces with the number of motor vehicles within a given environment, for example.
* Address: Department of Economics, University of Warwick, Coventry, UK, CV4 7AL. Telephone: +44 2476 28240; E-mail: [email protected]. I am grateful to Andrew Oswald, Jeff Round, Carol Graham, Peter Law, Norman Ireland, Robin Naylor, Rea Lydon, Jonathan Gardner, Pedro Martins, Alexandros Zangelidis , Maureen Pauls , and the participants at the RES Easter School in Birmingham, April, 2002 for helpful comments. The usual disclaimer applies.
Table of Contents
1
Introduction
1
2
On Subjective Well-being Research
2.1
Concept and Past Work
2
2.2
Towards a Developing Economy Framework
5
3
On South Africa and Data Description
3.1
General Back Ground
6
3.2
The South African Integrated Household Survey
8
4
Empirical Strategy and Preliminary Results
4.1
Basic Models of Life Satisfaction
9
4.1.1
Consumer Durables and Life Satisfaction
11
4.2
Personal Attributes and Subjective Well-being
12
4.2.1
Income Analysis
15
4.3
Social Norms: Between- and Within-Group Life Satisfaction
17
4.3.1
Comparison of Household Size and Income: First Evidence of Norm Effects
21
4.4
Testing for Social Limitation and Relative Consumption
27
4.5
Crime and Quality of Life
31
4.5.1
Testing for the Impact of Crime on Perceived Quality of Life
32
4.5.2
The Effect of Community Crime Rates on Life Satisfaction
35
5
Summary of Conclusions
37
References
40
a
List of Tables & Figures
Table 1
The Distribution of Perceived Quality of Life Responses
i
Table 1a
Happiness Responses in USA: 1972-1994
i
Table 1b
Life Satisfaction Responses in Europe: 1975-1992
i
Table 2
Life Satisfaction Equations with Household Variables
ii
as Independent Variables for South Africa, 1993 Table 3
Correlation between Reported Durable Goods
iii
Table 4
Life Satisfaction Equations with Personal Variables
iv
Table 5
Valuations in Household Income of Life Events
v
Table 6
Summaries of Data Categorised by Different Norm Sets
vi
Table 7
Life Satisfaction Equations on (1) Black and (2) White Sub-samples
vii
Table 8
Life Satisfaction Equations on (1) Rural and (2) Urban Sub-samples
viii
Table 9
Life Satisfaction Equations on (1) Rural Black and (2) Urban Black
ix
Table 10
Happiness Equations with Comparison Household Size Variable
Table 10.1
- estimated on a Full Sample
x
Table 10.2
- estimated on Black and White Sub-samples
xi
Table 10.3
- estimated on Urban and Rural Black Sub-samples
xii
Table 11
The Perceived Quality of Life Equations and Comparison Income Variable
xiii
Table 11a
- estimated on Urban Blacks with Other Wealth Variables removed
xiii
Table 12
Life Satisfaction Equations with Interaction Dummies of Consumer Durables
xiv
Table 13
Tests for Social Limitation and Relative Consumption on Motor Vehicle
xv
Table 14
Crime Rates and the Perceived Quality of Life
Table 14.1
Reported Victims of Crimes against Regions and Races
xvi
Table 14.2
Different Reported Crimes against Regions and Races
xvi
Table 15
Cross-tabulation of Crime and the Perceived Quality of Life
xviii
Table 16
Life Satisfaction Equations with Victim of Crime Variable
xviii
Table 17
Life Satisfaction Equations with ‘Relevant Others’ Crime Rate Variable
Table 17.1
- estimated on a Full sample and Regional Sub-samples
xix
Table 17.2
- estimated on Racial Sub-samples
xx
Table 17.3
- estimated on Regional Sub-samples with a Control for Race
xx
Figure 1
Proportion of Crime Types against Regions
xvii
Figure 2
Proportion of Crime Types against Races
xvii
b
4
Lqwurgxfwlrq
Dq dgyhuw iru R{idp4 dsshdov dvnv shrsoh Zkdw gr zh guhdp iru rxu fkloguhqB Li zh zhuh wkhq wr vwrs dqg wklqn derxw wkh txhvwlrq iru d plqxwh/ prvw ri xv zrxog suredeo| uhvsrqg zlwk vxffhvv dqg khdowk1 Krzhyhu/ wkh prvw qdwxudo uhvsrqvh frplqj iurp hyhu| rqh ri xv zrxog suredeo| eh wkh kdsslqhvv iru rxu fkloguhq1
Wkh txhvwlrq wkhq lv zkdw frqvwlwxwhv kdsslqhvvB
D uhylhz ri
uhvhdufk rq zhoo0ehlqj e| Zlovrq +4 d > orj \ > RwkhuKK > Frppxqlw|S > S huvrqdo , H+KS me @ + . . 2 . , . orj \ . RwkhuKK . Frppxqlw|S . S huvrqdo 5:
+8,
zklfk lv wkh h{shfwhg kdsslqhvv ohyho ri krxvhkrog wkdw srvvhvvhv dw ohdvw rqh xqlw ri frqvxphu gxudeoh e d 1 Qrwh wkdw e d uhsuhvhqwv d yhfwru ri gxpp| yduldeoh iru gxudeoh jrrg +4 iru kdyh/ dqg 3 iru qrw0kdyh,/ zkloh d lv d yhfwru ri wkh phdq ri uhohydqw frkruwv dvvhw1 Djdlq/ rxu lqirupdwlrq rq cuhohydqw frkruwv* kdv ehhq wdnhq iurp wkh frppxqlw| lqirupdwlrq jlyhq lq wkh vxuyh|1
Wkh
mxvwl fdwlrq lq wdnlqj wkh frppxqlw| phdq iru judqwhg lq rxu whvw iru uhodwlyh frqvxpswlrq fdq eh h{sodlqhg dv iroorzhg1
Dv frqvxphu gxudeohv lq txhvwlrq duh jrrgv wkdw uhtxluh lqiuhtxhqw
uhsodfhphqw lq gh qlwlrq dqg duh prvwo| nhsw dqg xwlolvh dw krph/ lw lv srvvleoh wr dvvxph iurp wkhlu sk|vlfdo dsshdudqfhv wkdw wkh rqo| shrsoh zkr duh prvw olnho| wr eh qrwlfhg dqg eh d hfwhg iurp rxu frqvxpswlrq ohyhov duh wkrvh zkr kdsshqhg wr eh olylqj forvh e| ru zlwklq wkh vdph frppxqlw|1 E| dvvxplqj wkdw wkh deryh vwdwhphqw lv ydolg zh fdq wkhq jr rq dqg phdvxuh d jhrjudsklfdoo| xvlqj wkh lqirupdwlrq jdwkhuhg iurp wkh foxvwhuhg frppxqlw| vdpsoh1
Olnh wkh phdq ohyhov ri
krxvhkrog vl}h dqg hgxfdwlrq/ qrwh wkdw rxu d lv dovr d frqwlqxrxv yduldeoh/ dqg lv doorzhg wr ydu| ehwzhhq gl huhqw krxvhkrogv1 Wkh wklug yduldeoh/ +e d d ,/ lv lqwurgxfhg dv d pdlq h hfw dqg lqwhudfwhg zlwk wkh krxvhkrog*v rzq srvvhvvlrq ri gxudeoh dvvhw m1 Zh kdyh dovr lqfoxghg wkh uhohydqw frkruwv* orj lqfrph yduldeoh/ orj \ > dv zhoo lq wkh uhjuhvvlrq wr fdswxuh rwkhu frpsdulvrq h hfwv suhvhqfh lq wkh orfdo frppxqlw| wkdw fdqqrw eh slfnhg xs iurp rxu olvw ri gxudeoh jrrgv1 Hpslulfdoo|/ zh h{shfw lq htxdwlrq +7, wr eh srvlwlyh dqg vljql fdqw iru gxudeohv wkdw hqwhu rxu olih vdwlvidfwlrq htxdwlrqv vljql fdqwo|1 Khqfh/ zh dqwlflsdwh wr eh srvlwlyh dqg vljql fdqw
dw wkh srsxodwlrq ohyho iru prwru yhklfoh/ jh|vhu/ whohskrqh/ dqg WY1 Wkh vljq ri frh!flhqw 2 / krzhyhu/ lv pruh gl!fxow wr suhglfw1 Zh gr qrw kdyh sulru lqirupdwlrq rq wkh uhodwlrqvklsv ehwzhhq uhohydqw frkruwv* srvvhvvlrq ri vrph gxudeohv dqg wkh zhoo0ehlqj ri wkrvh zkr gr qrw srvvhvv dq| ri wkh jlyhq gxudeoh vhwv1 Iru h{dpsoh/ zh fdqqrw suhglfw ehiruhkdqg krz wkh phdq ri orfdo fdu xvdjh zrxog d hfw wkh zhoo0ehlqj ri wkrvh zkr gr qrw rzq d fdu1 Kljk dyhudjh fdu rzqhuvkls pd| kdyh d qhjdwlyh lpsdfw rq wkh sv|fkrorjlfdo zhoo0ehlqj rq wkrvh zkr gr qrw srvvhvv dq| prwru yhklfoh wkurxjk vkhhu hqy| h hfw dqg qhjdwlyh h{whuqdolwlhv/ vxfk dv dlu srooxwlrq dqg wud!f mdp1 Qhyhuwkhohvv/ srvlwlyh h{whuqdolwlhv wkdw lqfuhdvh sv|fkrorjlfdo zhoo0ehlqj fdq dovr frh{lvw zlwk rwkhu qhjdwlyh h hfwv1 Duhdv zlwk kljk phdq ri orfdo fdu xvdjh duh olnho| wr eh pruh survshurxv/ frqwdlq ehwwhu idflolwlhv dqg surylgh rwkhu xqrevhuydeoh ehqh wv iru wkh qrq0fdu rzqhuv wkdq duhdv zlwk orzhu phdq1
Khqfh/ ghshqglqj rq wkh w|sh ri jrrgv dqg wkhlu h{whuqdolwlhv dqg rwkhu frqwuro
yduldeohv lq wkh uhjuhvvru/ 2 fdq hlwkhu eh srvlwlyh ru qhjdwlyh1 Zh suhglfw wkh lqwhudfwlrq whup lq htxdwlrq +7, wr eh qhjdwlyh dqg vljql fdqw iru jrrgv wkdw/
dffruglqj wr Kluvfk +4
Nattavudh Powdthavee* University of Warwick, January, 2003
Abstract
Is happiness pattern structurally the same in the poor countries as they are in the rich countries? Using a cross-sectional data from SALDRU93 survey, we show that relationships between subjective well-being and socioeconomic variables have a similar structure in South Africa as they are in the developed countries. The significance in correlations can vary between-group of social norms, however, if the discrepancies in norm characteristics are great. In addition, basic living standard indicators such as durable goods correlate positively and significantly with the reported level of life satisfaction. Crime committed on an individual, on the other hand, is associated with lower well-being in general. Lastly, the perceived quality of life is perhaps partly relative in many respects: owning a motor vehicle is correlated with higher satisfaction levels, but the positive effect reduces with the number of motor vehicles within a given environment, for example.
* Address: Department of Economics, University of Warwick, Coventry, UK, CV4 7AL. Telephone: +44 2476 28240; E-mail: [email protected]. I am grateful to Andrew Oswald, Jeff Round, Carol Graham, Peter Law, Norman Ireland, Robin Naylor, Rea Lydon, Jonathan Gardner, Pedro Martins, Alexandros Zangelidis , Maureen Pauls , and the participants at the RES Easter School in Birmingham, April, 2002 for helpful comments. The usual disclaimer applies.
Table of Contents
1
Introduction
1
2
On Subjective Well-being Research
2.1
Concept and Past Work
2
2.2
Towards a Developing Economy Framework
5
3
On South Africa and Data Description
3.1
General Back Ground
6
3.2
The South African Integrated Household Survey
8
4
Empirical Strategy and Preliminary Results
4.1
Basic Models of Life Satisfaction
9
4.1.1
Consumer Durables and Life Satisfaction
11
4.2
Personal Attributes and Subjective Well-being
12
4.2.1
Income Analysis
15
4.3
Social Norms: Between- and Within-Group Life Satisfaction
17
4.3.1
Comparison of Household Size and Income: First Evidence of Norm Effects
21
4.4
Testing for Social Limitation and Relative Consumption
27
4.5
Crime and Quality of Life
31
4.5.1
Testing for the Impact of Crime on Perceived Quality of Life
32
4.5.2
The Effect of Community Crime Rates on Life Satisfaction
35
5
Summary of Conclusions
37
References
40
a
List of Tables & Figures
Table 1
The Distribution of Perceived Quality of Life Responses
i
Table 1a
Happiness Responses in USA: 1972-1994
i
Table 1b
Life Satisfaction Responses in Europe: 1975-1992
i
Table 2
Life Satisfaction Equations with Household Variables
ii
as Independent Variables for South Africa, 1993 Table 3
Correlation between Reported Durable Goods
iii
Table 4
Life Satisfaction Equations with Personal Variables
iv
Table 5
Valuations in Household Income of Life Events
v
Table 6
Summaries of Data Categorised by Different Norm Sets
vi
Table 7
Life Satisfaction Equations on (1) Black and (2) White Sub-samples
vii
Table 8
Life Satisfaction Equations on (1) Rural and (2) Urban Sub-samples
viii
Table 9
Life Satisfaction Equations on (1) Rural Black and (2) Urban Black
ix
Table 10
Happiness Equations with Comparison Household Size Variable
Table 10.1
- estimated on a Full Sample
x
Table 10.2
- estimated on Black and White Sub-samples
xi
Table 10.3
- estimated on Urban and Rural Black Sub-samples
xii
Table 11
The Perceived Quality of Life Equations and Comparison Income Variable
xiii
Table 11a
- estimated on Urban Blacks with Other Wealth Variables removed
xiii
Table 12
Life Satisfaction Equations with Interaction Dummies of Consumer Durables
xiv
Table 13
Tests for Social Limitation and Relative Consumption on Motor Vehicle
xv
Table 14
Crime Rates and the Perceived Quality of Life
Table 14.1
Reported Victims of Crimes against Regions and Races
xvi
Table 14.2
Different Reported Crimes against Regions and Races
xvi
Table 15
Cross-tabulation of Crime and the Perceived Quality of Life
xviii
Table 16
Life Satisfaction Equations with Victim of Crime Variable
xviii
Table 17
Life Satisfaction Equations with ‘Relevant Others’ Crime Rate Variable
Table 17.1
- estimated on a Full sample and Regional Sub-samples
xix
Table 17.2
- estimated on Racial Sub-samples
xx
Table 17.3
- estimated on Regional Sub-samples with a Control for Race
xx
Figure 1
Proportion of Crime Types against Regions
xvii
Figure 2
Proportion of Crime Types against Races
xvii
b
4
Lqwurgxfwlrq
Dq dgyhuw iru R{idp4 dsshdov dvnv shrsoh Zkdw gr zh guhdp iru rxu fkloguhqB Li zh zhuh wkhq wr vwrs dqg wklqn derxw wkh txhvwlrq iru d plqxwh/ prvw ri xv zrxog suredeo| uhvsrqg zlwk vxffhvv dqg khdowk1 Krzhyhu/ wkh prvw qdwxudo uhvsrqvh frplqj iurp hyhu| rqh ri xv zrxog suredeo| eh wkh kdsslqhvv iru rxu fkloguhq1
Wkh txhvwlrq wkhq lv zkdw frqvwlwxwhv kdsslqhvvB
D uhylhz ri
uhvhdufk rq zhoo0ehlqj e| Zlovrq +4 d > orj \ > RwkhuKK > Frppxqlw|S > S huvrqdo , H+KS me @ + . . 2 . , . orj \ . RwkhuKK . Frppxqlw|S . S huvrqdo 5:
+8,
zklfk lv wkh h{shfwhg kdsslqhvv ohyho ri krxvhkrog wkdw srvvhvvhv dw ohdvw rqh xqlw ri frqvxphu gxudeoh e d 1 Qrwh wkdw e d uhsuhvhqwv d yhfwru ri gxpp| yduldeoh iru gxudeoh jrrg +4 iru kdyh/ dqg 3 iru qrw0kdyh,/ zkloh d lv d yhfwru ri wkh phdq ri uhohydqw frkruwv dvvhw1 Djdlq/ rxu lqirupdwlrq rq cuhohydqw frkruwv* kdv ehhq wdnhq iurp wkh frppxqlw| lqirupdwlrq jlyhq lq wkh vxuyh|1
Wkh
mxvwl fdwlrq lq wdnlqj wkh frppxqlw| phdq iru judqwhg lq rxu whvw iru uhodwlyh frqvxpswlrq fdq eh h{sodlqhg dv iroorzhg1
Dv frqvxphu gxudeohv lq txhvwlrq duh jrrgv wkdw uhtxluh lqiuhtxhqw
uhsodfhphqw lq gh qlwlrq dqg duh prvwo| nhsw dqg xwlolvh dw krph/ lw lv srvvleoh wr dvvxph iurp wkhlu sk|vlfdo dsshdudqfhv wkdw wkh rqo| shrsoh zkr duh prvw olnho| wr eh qrwlfhg dqg eh d hfwhg iurp rxu frqvxpswlrq ohyhov duh wkrvh zkr kdsshqhg wr eh olylqj forvh e| ru zlwklq wkh vdph frppxqlw|1 E| dvvxplqj wkdw wkh deryh vwdwhphqw lv ydolg zh fdq wkhq jr rq dqg phdvxuh d jhrjudsklfdoo| xvlqj wkh lqirupdwlrq jdwkhuhg iurp wkh foxvwhuhg frppxqlw| vdpsoh1
Olnh wkh phdq ohyhov ri
krxvhkrog vl}h dqg hgxfdwlrq/ qrwh wkdw rxu d lv dovr d frqwlqxrxv yduldeoh/ dqg lv doorzhg wr ydu| ehwzhhq gl huhqw krxvhkrogv1 Wkh wklug yduldeoh/ +e d d ,/ lv lqwurgxfhg dv d pdlq h hfw dqg lqwhudfwhg zlwk wkh krxvhkrog*v rzq srvvhvvlrq ri gxudeoh dvvhw m1 Zh kdyh dovr lqfoxghg wkh uhohydqw frkruwv* orj lqfrph yduldeoh/ orj \ > dv zhoo lq wkh uhjuhvvlrq wr fdswxuh rwkhu frpsdulvrq h hfwv suhvhqfh lq wkh orfdo frppxqlw| wkdw fdqqrw eh slfnhg xs iurp rxu olvw ri gxudeoh jrrgv1 Hpslulfdoo|/ zh h{shfw lq htxdwlrq +7, wr eh srvlwlyh dqg vljql fdqw iru gxudeohv wkdw hqwhu rxu olih vdwlvidfwlrq htxdwlrqv vljql fdqwo|1 Khqfh/ zh dqwlflsdwh wr eh srvlwlyh dqg vljql fdqw
dw wkh srsxodwlrq ohyho iru prwru yhklfoh/ jh|vhu/ whohskrqh/ dqg WY1 Wkh vljq ri frh!flhqw 2 / krzhyhu/ lv pruh gl!fxow wr suhglfw1 Zh gr qrw kdyh sulru lqirupdwlrq rq wkh uhodwlrqvklsv ehwzhhq uhohydqw frkruwv* srvvhvvlrq ri vrph gxudeohv dqg wkh zhoo0ehlqj ri wkrvh zkr gr qrw srvvhvv dq| ri wkh jlyhq gxudeoh vhwv1 Iru h{dpsoh/ zh fdqqrw suhglfw ehiruhkdqg krz wkh phdq ri orfdo fdu xvdjh zrxog d hfw wkh zhoo0ehlqj ri wkrvh zkr gr qrw rzq d fdu1 Kljk dyhudjh fdu rzqhuvkls pd| kdyh d qhjdwlyh lpsdfw rq wkh sv|fkrorjlfdo zhoo0ehlqj rq wkrvh zkr gr qrw srvvhvv dq| prwru yhklfoh wkurxjk vkhhu hqy| h hfw dqg qhjdwlyh h{whuqdolwlhv/ vxfk dv dlu srooxwlrq dqg wud!f mdp1 Qhyhuwkhohvv/ srvlwlyh h{whuqdolwlhv wkdw lqfuhdvh sv|fkrorjlfdo zhoo0ehlqj fdq dovr frh{lvw zlwk rwkhu qhjdwlyh h hfwv1 Duhdv zlwk kljk phdq ri orfdo fdu xvdjh duh olnho| wr eh pruh survshurxv/ frqwdlq ehwwhu idflolwlhv dqg surylgh rwkhu xqrevhuydeoh ehqh wv iru wkh qrq0fdu rzqhuv wkdq duhdv zlwk orzhu phdq1
Khqfh/ ghshqglqj rq wkh w|sh ri jrrgv dqg wkhlu h{whuqdolwlhv dqg rwkhu frqwuro
yduldeohv lq wkh uhjuhvvru/ 2 fdq hlwkhu eh srvlwlyh ru qhjdwlyh1 Zh suhglfw wkh lqwhudfwlrq whup lq htxdwlrq +7, wr eh qhjdwlyh dqg vljql fdqw iru jrrgv wkdw/
dffruglqj wr Kluvfk +4
