Happiness and the city an empirical study of

  • Slides: 29
Download presentation

Happiness and the city: an empirical study of the interaction between subjective well-being and

Happiness and the city: an empirical study of the interaction between subjective well-being and city satisfaction Irina Shafranskaya Dmitriy Potapov Anastasiya Bozhya-Volya Higher School of Economics, Perm September, 27 – 2016

Motivation and Purpose • City managers need to become “residents’ – centric”, as residents

Motivation and Purpose • City managers need to become “residents’ – centric”, as residents retention and resource attraction becomes the major problems of Russia cities; • City quality-of-life, which could be considered through the estimation of the residents’ satisfaction with the quality of different urban services, becomes an important performance indicator. Thus the issue of its effective measurement and monitoring becomes an urgent task. photo We propose and apply an assessment method designed to measure city satisfaction in relation with the subjective perception of individual wellbeing. It is designed to provide local policy-makers with a more refined tool for decision making in urban policy. photo ICARE - 2016

Theoretical Background • Overall satisfaction with a community can be decomposed into a variety

Theoretical Background • Overall satisfaction with a community can be decomposed into a variety of subdomains, each of which contributes to their overall feelings about the community (Sirgy et al. , 2000). • Residents’ satisfaction is largely determined by the variety of life domains, namely life satisfaction, happiness, job and income satisfaction (Diener et al. , 1999; Cummins and Cahill, 2001; Kelly, 2003). • City quality-of-life could be considered as the individual’s subjective experience of dealing with different urban services (Diener and Suh, 1997; Kahneman and photo Kruger, 2006) • Marketing perspective of residents satisfaction (Insch and Florek, 2008; Zenker et al. , 2013) • Subjective versus objective city quality-of-life approach (Tesfazghi, photo 2010; Obulicz – Kozaryn, 2013) ICARE - 2016

1. Linear Regression Model photo • Look for the effect of different urban services

1. Linear Regression Model photo • Look for the effect of different urban services satisfaction on the overall city satisfaction • From M 1 to M 4 include control variables to «purify» the effects • Measure only direct effects ICARE - 2016 photo

2. Path Analysis Model Measure direct and indirect effects photo XV April International Academic

2. Path Analysis Model Measure direct and indirect effects photo XV April International Academic Conference HSE, 2014

Data Collection • Door-to-door poll of more than 2000 inhabitants of Perm city (Russia).

Data Collection • Door-to-door poll of more than 2000 inhabitants of Perm city (Russia). City population is around 1 million people. • Sample is representative over • Gender • Age • City districts (7 areas) • Questionnaire contained 35 composite questions covering satisfaction and photo attitude to different aspects of life in the city (i. e. education, safety, etc. ) and overall city satisfaction, happiness and well-being. • Survey (sponsored by local authorities) was conducted in August-September, 2012 photo ICARE - 2016

Questionnaire Design Section 1 2 3 4 5 6 7 8 9 ICARE -

Questionnaire Design Section 1 2 3 4 5 6 7 8 9 ICARE - 2016 Model Parameter Culture (Cu_IND) section Education (Edu_IND) section Environment (Env_IND) section Healthcare (HC_IND) section Social security (SS_IND) section Safety (Saf_IND) section Sport (Sport_IND) section Subjective Well-Being Life Level Satisfaction Happiness Income Satisfaction City Satisfaction Number of statements (questions) 7 11 7 5 6 photo 6 6 1 1 photo 1 5

Data Preparation • Drop out observations with many (more than 2/3 in parcel) missings

Data Preparation • Drop out observations with many (more than 2/3 in parcel) missings (controlled for systematic bias). • Implemented parcel approach (Coffman, Mac. Callum, 2005) to convert different measures of a construct into one index. • Imputed missings in parcels as a prediction on the basis of linear regression, where dependent variables include other questions from the same parcel and social-demographic variables. photo • Generating parcel scores (indexes) as weighted sum of variables with equal weights. • 1 636 questionnaires fully and correctly filled in were included into analysis. photo ICARE - 2016

Sample Description Men Gender distirbution Women 0% 10% 20% 30% 40% 50% 60% 70%

Sample Description Men Gender distirbution Women 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Age distirbution 0% 10% 20% 14 - 17 30% 18 - 21 40% 22 - 35 50% 60% 36 - 49 70% 50 - 65 80% 90% 100% 65+ Civil Status 0% 10% 20% Married 30% Not married 40% 50% 60% Widowed / Divorced 70% photo 80% 90% 100% With partner Education 0% 10% 20% 30% Incomplete secondary ICARE - 2016 40% 50% Secondary 60% 70% Incomplete higher 80% 90% photo Higher 100%

Descriptive Statistics Variable Description # of Type, Scale # obs Mean S. d. indicators

Descriptive Statistics Variable Description # of Type, Scale # obs Mean S. d. indicators Min Max Personal happiness/satisfaction City. Sat_IND Life. Sat Happiness JIInc. Sat City satisfaction Life satisfaction Happiness Income satisfaction 5 1 1 1 Likert (1… 7) 1636 5. 1 3. 9 5. 6 3. 6 1. 4 1. 3 1. 5 1. 0 7 11 Likert (1… 7) 1636 5. 0 4. 3 1. 1 1. 0 photo 1. 0 7 Likert (1… 7) 1636 3. 6 1. 0 6. 9 5 Likert (1… 7) 1636 3. 5 1. 2 0. 9 7. 0 6 Likert (1… 7) 1636 3. 5 0. 9 1. 0 6. 5 6 6 Likert (1… 7) 1636 4. 0 3. 4 photo 1. 0 1. 1 0. 8 7. 0 5. 8 Urban services satisfaction indexes Cu_IND Edu_IND Env_IND HC_IND SS_IND Saf_IND Sport_IND ICARE - 2016 Culture satisfaction index Education satisfaction index Environment satisfaction index Health care satisfaction index Social security satisfaction index Safety satisfaction index Sport satisfaction index

Linear Regression Models – OLS Estimation (we report B - standardized β) ICARE -

Linear Regression Models – OLS Estimation (we report B - standardized β) ICARE - 2016 Factors Model 1 Model 2 Personal happiness/satisfaction 0. 12*** Life. Sat (0. 03) Happ JIInc. Sat Urban services satisfaction indexes 0. 17*** Cu_IND (0. 04) 0. 16*** 0. 15*** Edu_IND (0. 04) -0. 02 Env_IND (0. 04) 0. 13*** 0. 10*** HC_IND (0. 04) -0. 04 -0. 07 SS_IND (0. 05) 0. 22*** 0. 21*** Saf_IND (0. 04) 0. 05 0. 04 Sport_IND (0. 04) Model 3 Model 4 0. 02 (0. 03) 0. 24*** (0. 02) 0. 05 (0. 03) 0. 27*** (0. 02) 0. 05** (0. 03) 0. 13*** (0. 03) 0. 14*** (0. 04) 0. 01 (0. 04) 0. 08** (0. 03) -0. 03 (0. 04) 0. 20*** (0. 04) 0. 02 (0. 04) 0. 10*** (0. 03) 0. 11*** (0. 04) 0. 00 (0. 04) 0. 10*** (0. 03) -0. 02 (0. 04) 0. 16*** (0. 04) 0. 07** (0. 03) photo

Factors Model 1 Model 2 Social-demographic characteristics SD_Age. Gr SD_Civ. St_2 SD_Civ. St_3 SD_Civ.

Factors Model 1 Model 2 Social-demographic characteristics SD_Age. Gr SD_Civ. St_2 SD_Civ. St_3 SD_Civ. St_4 Linear Regression SD_Edu. Gr Models – OLS SD_Health_2 Estimation SD_Health_3 (continuation) SD_Health_4 SD_Gender SD_Work 2. 24*** 2. 12*** _cons (0. 20) Number of obs 1636 R-square 0. 15 * p<0. 1, ** p<0. 05, *** p<0. 01 XV April International Academic Conference Model 3 Model 4 1. 35*** (0. 21) 0. 27*** (0. 03) -0. 04 (0. 08) -0. 05 (0. 09) 0. 12 (0. 12) -0. 21*** (0. 02) -0. 03 (0. 07) 0. 12 (0. 10) -0. 01 (0. 13) photo -0. 01 (0. 06) -0. 15** (0. 07) 0. 86*** (0. 27) 1636 0. 21 photo 1636 0. 28

Path Analysis Model. DIRECT EFFECTS – ML Estimation (we report B - standardized β)

Path Analysis Model. DIRECT EFFECTS – ML Estimation (we report B - standardized β) ICARE - 2016 Factors City. Sat_IND Personal happiness/satisfaction 0. 05 Life. Sat (0. 03) 0. 27*** Happiness (0. 02) 0. 05** JIInc. Sat (0. 03) Urban services satisfaction indexes 0. 10*** Cu_IND (0. 03) 0. 11*** Edu_IND (0. 04) 0. 00 Env_IND (0. 04) 0. 10*** HC_IND (0. 03) -0. 02 SS_IND (0. 04) 0. 16*** Saf_IND (0. 04) 0. 07** Sport_IND (0. 03) Life. Sat Happiness 0. 19*** (0. 02) 0. 42*** (0. 02) 0. 17*** (0. 03) 0. 05** (0. 03) 0. 00 (0. 03) 0. 04 (0. 03) 0. 07*** (0. 03) 0. 17*** (0. 03) 0. 04 (0. 03) 0. 01 (0. 03) 0. 19*** (0. 04) 0. 07 (0. 04) -0. 09** (0. 04) 0. 08** (0. 04) -0. 10** (0. 05) 0. 10** (0. 04) 0. 06 (0. 04) photo

Path Analysis Model. DIRECT EFFECTS – ML Estimation (continuation) Factors City. Sat_IND Social-demographic characteristics

Path Analysis Model. DIRECT EFFECTS – ML Estimation (continuation) Factors City. Sat_IND Social-demographic characteristics 0. 27*** SD_Age. Gr (0. 03) -0. 04 SD_Civ. St_2 (0. 08) -0. 05 SD_Civ. St_3 (0. 09) 0. 12 SD_Civ. St_4 (0. 12) -0. 21*** SD_Edu. Gr (0. 02) -0. 03 SD_Health_2 (0. 07) 0. 12 SD_Health_3 (0. 10) -0. 01 SD_Health_4 (0. 13) -0. 01 SD_Gender (0. 06) -0. 15** SD_Work (0. 07) Number of obs 1636 R-squared 0. 29 * p<0. 1, ** p<0. 05, *** p<0. 01 XV April International Academic Conference Life. Sat Happiness -0. 05* (0. 02) 0. 09 (0. 06) -0. 12* (0. 07) -0. 08 (0. 09) 0. 05** (0. 02) -0. 05 (0. 06) -0. 23*** (0. 08) -0. 15 (0. 10) -0. 02 (0. 05) -0. 07 (0. 05) 1636 0. 52 -0. 18*** (0. 04) -0. 41*** (0. 09) -0. 32*** (0. 10) -0. 20 (0. 13) 0. 04 (0. 03) -0. 32*** (0. 08) -0. 31*** (0. 11) -0. 32** (0. 15) -0. 03 (0. 07) -0. 21*** (0. 08) 1636 0. 16

Path Analysis Model. City. Sat equation – ML Estimation (we report B - standardized

Path Analysis Model. City. Sat equation – ML Estimation (we report B - standardized β) ICARE - 2016 Factors DIRECT Personal happiness/satisfaction 0. 05 Life. Sat (0. 03) 0. 27*** Happiness (0. 02) 0. 05** JIInc. Sat (0. 03) Urban services satisfaction indexes 0. 10*** Cu_IND (0. 03) 0. 11*** Edu_IND (0. 04) 0. 00 Env_IND (0. 04) 0. 10*** HC_IND (0. 03) -0. 02 SS_IND (0. 04) 0. 16*** Saf_IND (0. 04) 0. 07** Sport_IND (0. 03) INDIRECT TOTAL 0. 01*** (0. 00) 0. 07*** (0. 02) 0. 05 (0. 03) 0. 28*** (0. 02) 0. 12*** (0. 02) 0. 05*** (0. 01) 0. 02 (0. 01) -0. 02* (0. 01) 0. 03** (0. 01) -0. 02 (0. 01) 0. 03** (0. 01) 0. 02 (0. 01) 0. 16*** (0. 04) 0. 13*** (0. 04) -0. 03 (0. 04) 0. 12*** (0. 03) -0. 04 (0. 04) 0. 19*** (0. 04) 0. 08** (0. 04) (no path)

Path Analysis Model. City. Sat equation – ML Estimation (continuation) Factors DIRECT Social-demographic characteristics

Path Analysis Model. City. Sat equation – ML Estimation (continuation) Factors DIRECT Social-demographic characteristics 0. 27*** SD_Age. Gr (0. 03) -0. 04 SD_Civ. St_2 (0. 08) -0. 05 SD_Civ. St_3 (0. 09) 0. 12 SD_Civ. St_4 (0. 12) -0. 21*** SD_Edu. Gr (0. 02) -0. 03 SD_Health_2 (0. 07) 0. 12 SD_Health_3 (0. 10) -0. 01 SD_Health_4 (0. 13) -0. 01 SD_Gender (0. 06) -0. 15** SD_Work (0. 07) Number of obs R-squared * p<0. 1, ** p<0. 05, *** p<0. 01 XV April International Academic Conference INDIRECT TOTAL -0. 05*** (0. 01) -0. 11*** (0. 03) -0. 10*** (0. 03) -0. 06 (0. 04) 0. 01* (0. 01) -0. 09*** (0. 02) -0. 10*** (0. 03) -0. 10** (0. 04) -0. 01 (0. 02) -0. 06*** (0. 02) 1636 0. 29 0. 21*** (0. 03) -0. 15* (0. 09) -0. 14 (0. 10) 0. 06 (0. 12) -0. 19*** (0. 03) -0. 12 (0. 08) 0. 02 (0. 10) -0. 10 (0. 14) -0. 02 (0. 07) -0. 21*** (0. 07)

Subjective Well-Being and City Satisfaction: Causality Identification • We suppose simultaneity between Happiness, Life.

Subjective Well-Being and City Satisfaction: Causality Identification • We suppose simultaneity between Happiness, Life. Sat and City. Sat (e. g. increasing the satisfaction with the city makes inhabitants feel happier) • We suggest instrumental variable strategy to manage this issue • Find instruments • Valid – not correlated with error term in the equation for City. Sat • Relevant – correlated with endogenous variables (Happiness and Life. Sat) photo • Socio-demographic characteristics Health and Civil (marital) Status are the candidates • Use general method of moments estimation • Test for validity and relevance • Test for weak instruments • Hausman specification test The 2 nd conference of HSE-Perm «Neighbors in Research» photo

OLS v. s. IV results Factors OLS Personal happiness/satisfaction 0. 05 Lif. Sat (0.

OLS v. s. IV results Factors OLS Personal happiness/satisfaction 0. 05 Lif. Sat (0. 03) 0. 27*** Happ (0. 02) 0. 05** JIInc. Sat (0. 03) City attributes indexes 0. 11*** Cu_IND (0. 03) 0. 11*** Edu_IND (0. 04) 0. 00 Env_IND (0. 04) 0. 10*** HC_IND (0. 03) -0. 02 SS_IND (0. 04) 0. 16*** Saf_IND (0. 04) 0. 07** Sport_IND (0. 03) IV -0. 39 (0. 35) 0. 46** (0. 18) 0. 22 (0. 14) 0. 11** (0. 05) 0. 10** (0. 04) 0. 02 (0. 05) 0. 12*** (0. 04) 0. 07 (0. 08) 0. 17*** (0. 04) 0. 07* (0. 04) The 2 nd conference of HSE-Perm «Neighbors in Research» Hausman test: Ho: difference in coefficients is not systematic chi 2(14) = 3. 16 Prob > chi 2 = 0. 9988 Can use more efficient OLS (path model ML estimates)

Parcel weighting – robustness check procedure To track the subjective weights choice problem in

Parcel weighting – robustness check procedure To track the subjective weights choice problem in the model we conducted the robustness check: 1. 500 times repeat the procedure: • for each indicator in each parcel take random independent draws from standard uniform distribution; • normalize this draws in each parcel to make their sum be equal to unity; • calculate new set of eight index variables with this weights; • estimate path model and stored total effects; photo 2. construct empirical distribution of this estimates. photo The 2 nd conference of HSE-Perm «Neighbors in Research»

Parcel weighting – robustness check results photo • Significant parameters keep appropriate sign in

Parcel weighting – robustness check results photo • Significant parameters keep appropriate sign in almost all 500 model replicas • Parameters distributions are unimodal with quite narrow support photo The 2 nd conference of HSE-Perm «Neighbors in Research»

Visualization of the Results photo ICARE - 2016

Visualization of the Results photo ICARE - 2016

Visualization of the Results photo ICARE - 2016

Visualization of the Results photo ICARE - 2016

Visualization of the Results photo Perm Workshop on Applied Economic Modeling

Visualization of the Results photo Perm Workshop on Applied Economic Modeling

Visualization of the Results photo ICARE - 2016

Visualization of the Results photo ICARE - 2016

Basic results 1. Considering that subjective well-being influence upon city satisfaction helps to increase

Basic results 1. Considering that subjective well-being influence upon city satisfaction helps to increase substantially the accuracy of the tool, which we use to measure city satisfaction (adjusted R-square grows from 0. 15 to 0. 28). 2. Priorities of city management could be clearly defined (as health care has the lowest satisfaction estimation (3. 5) and significantly influences city satisfaction, one could develop the health-care focused city policy, which could meet residents’ needs and increase city satisfaction). 3. City satisfaction has both structural and cumulative ‘nature’. photo 4. Path Analysis Model is more appropriate to identify influence of urban services on overall city satisfaction than Linear Regression Model, because we observe significant indirect effects. 5. There is an influence of Happiness on City Satisfaction but not vice versa. photo 6. Perm residents are happy people Perm Workshop on Applied Economic Modeling

Limitations and Further Research Limitations • One city – one case • Only urban

Limitations and Further Research Limitations • One city – one case • Only urban services managed by local administration are taken into account • Parcel approach Further research • Carry out group analysis (split sample on the basis of observed socialphoto demographic variables and estimate the model for this groups) • Identify homogeneous clusters and estimate the model for this clusters photo ICARE - 2016

Questions and comments are welcome! Higher School of Economics , 2016 www. hse. ru

Questions and comments are welcome! Higher School of Economics , 2016 www. hse. ru