Regional Studies

ISSN: 0034-3404 (Print) 1360-0591 (Online) Journal homepage: http://www.tandfonline.com/loi/cres20

La Dolce Vita: Hedonic Estimates of Quality of Life in Italian Cities Emilio Colombo, Alessandra Michelangeli & Luca Stanca To cite this article: Emilio Colombo, Alessandra Michelangeli & Luca Stanca (2014) La Dolce Vita: Hedonic Estimates of Quality of Life in Italian Cities, Regional Studies, 48:8, 1404-1418, DOI: 10.1080/00343404.2012.712206 To link to this article: http://dx.doi.org/10.1080/00343404.2012.712206

Published online: 30 Aug 2012.

Submit your article to this journal

Article views: 430

View related articles

View Crossmark data

Citing articles: 11 View citing articles

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=cres20 Download by: [37.44.207.36]

Date: 26 January 2017, At: 08:06

Regional Studies, 2014 Vol. 48, No. 8, 1404–1418, http://dx.doi.org/10.1080/00343404.2012.712206

La Dolce Vita: Hedonic Estimates of Quality of Life in Italian Cities EMILIO COLOMBO*, ALESSANDRA MICHELANGELI† and LUCA STANCA*

*Department of Economics, University of Milan Bicocca, Piazza dell’Ateneo Nuovo, 1, I-20126 Milan, Italy. Emails: [email protected] and [email protected] †Department of Economic and Legal Systems, University of Milan Bicocca, Piazza dell’Ateneo Nuovo, 1, I-20126 Milan, Italy. Email: [email protected] (Received February 2011: in revised form June 2012) COLOMBO E., MICHELANGELI A. and STANCA L. La dolce vita: hedonic estimates of quality of life in Italian cities, Regional Studies. This paper investigates quality of life in Italian cities using the hedonic approach. It analyses micro-level data for housing and labour markets to estimate compensating differentials for local amenities within four domains: weather, environment, services and society. Large compensating differentials in housing markets are found, whereas the effects on wages are relatively small. Quality of life varies substantially across space and is generally better in large and medium-sized cities of the Centre–North. Services and social conditions are strongly related to overall quality of life. It is also found that, across cities, quality of life is positively and significantly related to subjective well-being. Quality of life

Hedonic prices

Housing markets

Well-being

COLOMBO E., MICHELANGELI A. and STANCA L. 甜蜜的生活：意大利城市生活质量的享乐(特征)评价，区域研究。本文 运用享乐(特征)方法，探讨意大利的城市生活质量。本文分析住宅与劳动市场的微观层级资料，评价地方舒适性在 气候、环境、服务与社会四大面向中的补偿性差异。我们在住宅市场中发现大幅的补偿性差异，对于工资的影响则 相对薄弱。不同地方的生活质量差异甚大，普遍来说，中、北部的中、大型城市具有较好的生活质量。服务与社会 条件强烈关乎总体生活质量。本研究亦发现，在各城市中，生活质量明确且显著地与主观幸福有关。 生活质量

享乐(特征)价格

住宅市场

幸福

COLOMBO E., MICHELANGELI A. et STANCA L. La vie en rose: des estimations hédonistiques de la qualité de la vie en Italie, Regional Studies. À partir de l’approche hédonistique, cet article cherche à examiner la qualité de la vie des grandes villes italiennes. On analyse des données micro-économiques sur les marchés du logement et du travail afin d’estimer les différentiels compensateurs des équipements collectifs pour quatre domaines: à savoir, le temps, l’environnement, les services et la société. Il s’avère d’importants différentiels compensateurs pour ce qui est des marchés du logement, alors que l’effet sur les salaires est relativement modéré. La qualité de la vie varie énormément à travers l’espace et s’annonce mieux en règle générale dans les grandes villes et les villes moyennes situées dans la partie Centre-Nord du pays. Les services et les conditions sociales sont étroitement liés à la qualité de la vie dans son ensemble. Il s’avère aussi que la qualité de la vie est positivement et fortement corrélée au bien-être subjectif à travers les grandes villes. Qualité de la vie

Prix hédonistiques

Marchés du logement

Bien-être

COLOMBO E., MICHELANGELI A. und STANCA L. La dolce vita: hedonische Schätzungen der Lebensqualität in italienischen Städten, Regional Studies. In diesem Beitrag untersuchen wir mit Hilfe des hedonischen Ansatzes die Lebensqualität in italienischen Städten. Wir analysieren Daten auf Mikroebene für die Wohnungs- und Arbeitsmärkte, um die kompensierenden Differentiale für lokale Merkmale innerhalb von vier Bereichen zu schätzen: Wetter, Umwelt, Dienstleistungen und Gesellschaft. Auf dem Wohnungsmarkt finden wir große kompensierende Differentiale, während die Auswirkungen auf die Gehälter relativ gering ausfallen. Die Lebensqualität schwankt innerhalb des Raums erheblich und fällt in den großen und mittelgroßen Städten in der Landesmitte und im Norden generell besser aus. Dienstleistungen und soziale Bedingungen stehen in einem engen Zusammenhang mit der generellen Lebensqualität. Ebenso stellen wir für sämtliche Städte fest, dass die Lebensqualität positiv und signifikant mit dem subjektiven Wohlbefinden verknüpft ist. Lebensqualität

Hedonische Preise

Wohnungsmärkte

Wohlbefinden

COLOMBO E., MICHELANGELI A. y STANCA L. La dolce vita: cálculos hedónicos de la calidad de vida en las ciudades italianas, Regional Studies. Mediante un enfoque hedónico, en este artículo investigamos la calidad de vida en las ciudades italianas. © 2012 Regional Studies Association http://www.regionalstudies.org

Hedonic Estimates of Quality of Life in Italian Cities

1405

Analizamos datos a nivel micro para los mercados laboral y de la vivienda con el fin de calcular los diferenciales compensatorios en las prestaciones locales en cuatro sectores: clima, medio ambiente, sociedad y servicios. Observamos que en los mercados de la vivienda existen importantes diferenciales compensatorios, mientras que los efectos en los salarios son relativamente pequeños. La calidad de vida varía considerablemente en el espacio y, en general, es mejor en las ciudades de tamaño grande y medio del centro-norte. En general, los servicios y las condiciones sociales están estrechamente relacionados con la calidad de vida general. También comprobamos que en todas las ciudades, la calidad de vida tiene un vínculo positivo y significativo con el bienestar subjetivo. Calidad de vida

Precios hedónicos

Mercados de la vivienda

Bienestar

JEL classifications: C4, D5, H4, J3, J6, Q2, R2

INTRODUCTION In recent years, as fiscal federalism has come to the forefront of the policy debate in several countries, the measurement and comparison of quality of life (QoL) across regions and urban areas has become a key issue for policy-makers and the general public. As a consequence, QoL and its determinants have received increasing attention well beyond the academic debate (RAPPAPORT , 2009). A large body of literature has proposed several methods for measuring QoL in regions and cities on the basis of their observable characteristics (for recent reviews, see, for example, BLOMQUIST , 2007; and LAMBIRI et al., 2007).1 Within this literature, QoL is generally defined as the weighted average of a set of local amenities. Therefore, one of the key issues is how to weigh the different amenities appropriately. Following the theoretical approach proposed by ROSEN (1979) and extended by ROBACK (1982), several versions of the hedonic price method have been used to value amenities and construct QoL indicators. Within this framework, households reveal their preferences for the bundle of attributes that characterize urban areas through their location decisions. The monetary value of a local amenity can be determined on the basis of the housing prices households are willing to pay and the wages they are willing to accept to locate in a given area. The basic intuition is that, in equilibrium, the absence of spatial arbitrage implies that households pay higher rents, or accept lower wages, in order to live in areas with better amenities. QoL can therefore be measured, and compared across areas, by weighting local amenities with the implicit prices derived from compensating differentials in housing and labour markets. Over the last decades, several studies have followed this approach, differing in terms of scope, selection of amenities and spatial disaggregation level. While several applications of the hedonic approach to the measurement of QoL across urban areas exist for the United States (for example, ROBACK , 1982; BLOMQUIST et al., 1988; KAHN , 1995; SHULTZ and KING , 2001; COSTA and KAHN , 2003; EZZET -LOFSTROM , 2004; GABRIEL and ROSENTHAL , 2004; SHAPIRO , 2006; ALBOUY , 2008; RAPPAPORT , 2008; WINTERS , 2010), there are relatively

fewer studies that compare QoL across cities outside the United States (for example, CHESHIRE and SHEPPARD , 1995; GIANNIAS , 1998; BERGER et al., 2008; SRINIVASAN and STEWART , 2004; BUETTNER and EBERTZ , 2009). The present study is, to the best of the authors’ knowledge, the first application of the hedonic approach to micro-level data to measure and compare QoL in Italy.2 The analysis focuses on the determinants and the distribution of QoL across Italian cities (province capitals). This paper uses individual-level data for housing prices and wages, together with city-level data on local amenities, in order to estimate compensating differentials in housing and labour markets. Implicit prices for local amenities are obtained within four main domains: climate, environment, services and society. The estimated implicit prices are used to rank the 103 Italian province capitals on the basis of overall and domain-specific QoL. The relationship between QoL and subjective well-being across cities is also examined. The results indicate that the presence of amenities results in large compensating differentials for the housing market, whereas the effects on wage differentials are relatively small. Overall, all the estimated full implicit prices have the expected sign. Substantial geographic variation in QoL was found, with the overall index reflecting different classes of amenities across cities. QoL is generally better in medium-sized cities of the Centre–North. Northern cities fare better for services and social conditions, while relatively worse for climate and environmental conditions. The domainspecific QoL indicators are related to the overall index in various degrees. While the weather component is negatively and weakly related to overall QoL, the environment component is positively but weakly related to overall QoL. The services and society components are instead positively and strongly related to overall QoL. Finally, it is shown that QoL across Italian cities is positively related to subjective wellbeing, with social conditions (crime, unemployment and urban density) playing a key role. The remainder of the paper is structured as follows. The second section briefly reviews the theoretical

Emilio Colombo et al.

1406

framework. The third section describes the data (details on the data sets and definition of variables used for the analysis are provided in Appendix A). The fourth section discusses the methodology. The fifth section presents the results. The sixth section concludes.

THEORETICAL FRAMEWORK Following ROSEN (1979) and ROBACK (1982), this paper considers a spatial equilibrium model where households and firms compete to locate in areas characterized by different bundles of amenities. Households derive utility from consuming a composite consumption good, housing and local amenities. Access to local amenities is obtained by living in a given location. Labour income allows the purchase of both the composite consumption good and housing. In city j, a household’s indirect utility is: v j = v(wj − rj , Aj )

(1)

where v(·) is the maximum level of utility that the household can obtain with wage wj, housing rent rj, and the vector of amenities Aj, with ∂v/∂w . 0, ∂v/∂r , 0 and ∂v/∂aij > < 0 depending on whether ai is a consumption amenity or disamenity. The price of the composite consumption good (x) is normalized to 1, so that:3 xj = wj − rj The composite consumption good is produced by firms that use a constant returns-to-scale technology using labour and land as inputs. For simplicity, the land rent is assumed to be equal to the housing rent rj. The consumption good is tradable and its price is fixed by international competition. The unit production cost in city j is: c j = c(wj , rj , Aj )

(2)

with ∂c/∂w . 0, ∂c/∂r . 0 and ∂c/∂aij > < 0 depending on whether ai is a production amenity or disamenity. Equilibrium requires the absence of spatial arbitrage, so that household utility and production costs are equal across cities: ∗

u = v(wj − rj , Aj )

(3)

1 = c(wj , rj , Aj )

(4)

In a spatial equilibrium, differences in wages and housing costs should compensate individuals and firms for differences in location-specific characteristics. Fig. 1 illustrates the equilibrium described by equations (3) and (4). Better amenities cause the iso-utility curve to shift upwards, resulting in higher housing costs and

Fig. 1. Spatial equilibrium with rents and wages lower wages, under the assumption that amenities do not have productivity effects. However, if local amenities also affect firms’ productivity, the net effect on wages is ambiguous. A higher level of a production amenity would result in an upward shift of the isocost curve. Note that while there is no ambiguity in the effect of amenities on rents, the presence of an amenity can produce higher wages in equilibrium if the effect on firms’ labour demand dominates the effect on households’ labour supply. Wages and housing costs can be used to obtain implicit prices for amenities. Taking the total differential of (3), and rearranging, the following is obtained: ∂v fi = ∂aij




drj dwj ∂v = − ∂xj ∂aij ∂aij

(5)

drj is the equilibrium compensating differential for ∂aij dwj is the equilibrium-compensating housing costs; and ∂aij differential for wages. The valuation of an amenity can therefore be obtained from the marginal responses of housing costs and wages to different levels of amenities. Given the estimates of the implicit prices fi, an index of QoL for city j can be constructed as the weighted sum of each amenity i, with weights given by the implicit prices fi that reflect households’ preferences: where

QoLj =



fi aij

(6)

i

Urban QoL indices thus constructed can be interpreted as the monetary value that the representative household attributes to the bundle of amenities available in each city. DATA The empirical analysis relies on three different data sets covering a period between 2001 and 2010. Two data

Hedonic Estimates of Quality of Life in Italian Cities sets provide individual-level information on the housing market and the labour market, respectively. The third data set provides city-level information on amenities. A detailed description of the variables and sources is provided in Appendix A. This paper focuses on cities defined as the municipalities of province capitals. The unit of analysis is therefore the municipal area of province capitals, rather than the whole province territory (this definition is important for the interpretation of city rankings and maps). Overall, 103 cities are considered. The size distribution is rather skewed, with only four cities near or above 1 million inhabitants (Roma, Milano, Napoli and Torino) and more than half with fewer than 100 000 inhabitants. Housing market data are from the Real Estate Observatory of the Agenzia del Territorio (AT), and refer to individual house transactions in Italian cities (province capitals) between 2004 and 2010 at semi-annual frequency. In addition to housing transaction prices, the data set provides a detailed description of structural characteristics, such as surface area, age of building, number of bathrooms, floor level, number of garages or car parks, level of maintenance, location (centre, semi-centre, suburb), quality of building (good, average, bad), and neighbourhood characteristics, such as quality of the area and distance from public services, commercial services and transportation system. Two features of the AT data set should be noticed. First, rental data are not available. However, the rental market has a relatively limited importance in Italy (less than 20% of existing houses) and the available data are unreliable due to high tax evasion. Second, housing prices are based on actual sales, and are therefore potentially not representative of the total housing stock. A market measure of the value of all properties is not available for Italian cities. The housing price data were crossvalidated with data for the Land Registry imputed income (‘Rendite Catastali’), available for all properties in different cities. Across provinces there is a high correlation between housing prices from the two sources, ranging between 0.55 and 0.63 depending on the type of properties considered. This suggests that restricting the analysis to housing prices based only on sales data is not likely to produce a substantial bias relative to the underlying population of property prices. Labour market data are repeated cross-sections of individual workers in the private sector for years 2001 and 2002 from the Istituto Nazionale Previdenza Sociale (INPS – Italian National Social Security Institute). The data set provides information on annual earnings, type of occupation, full-time or part-time work status, contract length, and province of work. The employee’s longitudinal records were linked to the demographic and firms archives in order to obtain information on worker characteristics (gender, age, nationality, province of residence, etc.) and firm characteristics (size and sector of activity). The sample was restricted to all employees aged between sixteen and

1407

seventy-five years who live in the same city where they work for at least thirty weeks in a year. Seasonal workers were not included in the sample. Wages of part-time workers were converted to a full-time equivalent using a 1.4 multiplicative factor.4 Annual earnings were total yearly wages net of social contributions paid by firms, but gross of social contributions and income taxes paid by workers. The labour market data set presents two main limitations. First, information on workers’ education levels is not available. The occupation level (executive, manager and white collar, blue collar, apprentice) was therefore used as a proxy for educational attainment. Second, there is a substantial difference in the timing of the data between labour (2001–2002) and housing market data (2004–2010). This feature is due to data availability, as there are no analogous micro-level data sets, with information on job location, available for more recent years. However, given that this paper is using data for employees only (entrepreneurs and self-employed are not included), the distribution of wages and workers characteristics across cities is relatively stable over time. The cross-sections for 2001 and 2002 can therefore be considered a good approximation of those matching the timing of housing market and local amenities data. Information on local amenities and characteristics for the municipalities of the 103 Italian provinces was collected for the period 2001–2008 from the Istituto Nazionale di Statistica (ISTAT – Italian National Statistical Office) and other sources (for details, see Table A1 in Appendix A). Summary statistics are provided in Table 1. This paper considers twelve city-level amenities within four different domains: climate, environment, services and society. Climate is measured with three indicators: temperature (maximum in January), precipitation (monthly average) and humidity (maximum in July). The environmental domain is based on both physical features of the territory (percentage of green areas in the city and a dummy variable indicating a coastal city) and pollution (number of polluting agents present in the air). Indicators for the quality of services focus on education (teacher-to-pupil ratio), culture (index of cultural infrastructure, measuring several dimensions of the city’s cultural offerings, such as museums, cinemas, theatres, etc.), and transport infrastructure (a multi-modal indicator of accessibility by air, train and car). The society domain refers to socio-economic conditions of cities, as measured by crime rate, unemployment rate and population density. Population density is included in the list of amenities in order to account for the large body of literature on agglomeration effects that emphasizes several channels through which city density affects QoL.5 Although some of these channels, such as pollution and accessibility, are accounted for by the set of amenities, density can be associated with additional forces that are not captured in the specification. The unemployment rate is included among the amenities since it is an important determinant of social conditions and individual well-being,

Emilio Colombo et al.

1408

Table 1. Descriptive statistics Variable Precipitation (mm/month) Maximum temperature in January Humidity in July (%) Coast (dummy) Green areas (%) Air pollution (number of agents) Teacher-to-pupil ratio (%) Transport infrastructure (index) Cultural infrastructure (index) Violent crime (per 1000) Unemployment rate (%) Urban density (1000/km2) Real housing price Real wage Subjective well-being

Mean

Standard deviation

Minimum

Maximum

68.6 9.1 67.3 0.3 6.9 7.7 9.7 91.6 87.3 4.1 11.1 1.2 220 963.6 18 653.2 6.8

22.4 3.7 5.2 0.5 11.2 2.6 0.9 24 77.6 1.5 7.5 1.4 147 019.4 8184.2 0.66

28.9 4 51 0.0 0.1 1.3 8.3 47 18.9 1.1 2.8 0.1 34 542.6 6929.7 3.33

139.7 16 77 1.0 71.9 15.4 13.3 161 579.2 9.9 31.4 8.3 1 100 000 63 498.4 9.25

Note: Number of observations =103 (Italian province capitals) for local amenities and subjective well-being; 158 512 for real housing prices; 167 908 for real wages. Sources: ISTITUTO NAZIONALE DI STATISTICA (ISTAT) (2009) and other sources, as detailed in Table A1 in Appendix A.

over and above its pure economic costs (for example, WINKELMANN and WINKELMANN , 1998), and is commonly used in the hedonic literature (for example, ROBACK , 1982; HERZOG and SCHLOTTMANN , 1993; DELLER et al., 2001; BUETTNER and EBERTZ , 2009). It should be noted that both unemployment and population density can be considered endogenous outcomes. Higher QoL may contribute to higher unemployment, which may serve as a negative compensation for living in a nicer place. However, several studies show that regional migration flows within Italy are mainly driven by employment opportunities (for example, FURCERI , 2006; ETZO , 2011), suggesting that unemployment can be considered a cause, rather than a consequence, of households’ location decisions. A better QoL may also lead to higher density (RAPPAPORT , 2008). However, Italian cities are generally constrained by borders inherited from the medieval period that rigidly define the size of the city, while strict urban regulations pose severe limitations to the vertical growth of buildings. This suggests that the endogeneity of population density is likely to play a minor role in the case of Italian cities. Overall, the list of city attributes that are considered here as local amenities is rather selective. This is due to the relatively small number of observations for province capitals (103) and the high degree of collinearity among the several possible indicators of QoL determinants. The relevance of local fiscal conditions for QoL, for example, has been examined by GYOURKO and TRACY (1989, 1991), whose results indicated that local taxes should be included in the set of amenities for QoL measurement (also ALBOUY , 2009). The present analysis does not include local taxes in the list of amenities since the centralization of the Italian fiscal system leaves limited room for local authorities to determine tax rates.

Differences in non-housing prices across cities may also significantly affect the measurement of QoL. Failing to account for non-housing price differentials may lead to overestimate significantly QoL in cities with higher living costs (ALBOUY , 2008). However, there are no available estimates of non-housing price differentials across Italian cities. As a validity check, the non-housing price differences across the twenty main Italian cities reported in a recent study by ISTAT (2009) were examined. The comparison with the present data set shows that there is a high correlation across cities between housing and non-housing price differentials. More importantly, the size of nonhousing price differentials is very limited relative to housing price differentials.6 METHODS The implicit price of amenities is measured by estimating two separate equations for the log of housing prices and wages: ln phjt = bo + b1 Xhjt + b2 Ajt + 1hjt

(7)

ln wkjt = go + g1 Zkjt + g2 Ajt + hkjt

(8)

where lnphjt is the real price of housing unit h in city j at time t; Xhjt is a vector of housing characteristics; Ajt is a vector of local amenities in city j; wkjt is the real wage of individual k in city j at time t; Zkjt is a vector of individual characteristics; and 1hjt  N(0, s21 ) and ηkjt  N(0, s2h ). Housing prices and wages are measured at constant 2004 prices. The application of the hedonic approach is based on the assumption that there are no unobserved characteristics for housing units, workers and cities that are correlated with observable local amenities. The detailed information on individual characteristics (Xhjt and Zkjt)

Hedonic Estimates of Quality of Life in Italian Cities is used to control for the heterogeneity of houses and workers. The structural characteristics in Xhjt include property size, age of the building, number of bathrooms, floor level, number of floors, number of lifts, number of garages or car parks, housing type, unit conditions, housing features, value type, location, and quality of building. Neighbourhood characteristics include position and quality of the area. Worker and firm characteristics in Zkjt include gender, age, nationality, province of residence, type of occupation, contract length, size of the firm, and sector of activity. Year dummies are included to account for time-fixed effects. Cities’ unobserved heterogeneity are also controlled for by including population size and a dummy for region capitals. Equations (7) and (8) are estimated by ordinary least squares (OLS) using approximately 151 000 and 158 000 observations, respectively. Robust standard errors are used with clustering at the city level in order to allow for within-city correlation. In order to obtain the full implicit price of each amenity the estimated coefficients β2 and γ2 in equations (7) and (8) are converted into annual household expenditures. The coefficients for the housing price equation, evaluated at the mean housing price, are converted into imputed annual rents by applying a 7.85% discount rate, as in BLOMQUIST et al. (1988). The coefficients for the wage equation, evaluated at the mean wage, are multiplied by 1.64, the average number of workers per household (BANK OF ITALY , 2008), in order to obtain household annual wages comparable with household housing expenditures. This allows the full price fi for each amenity to be computed. As in equation (6), full implicit prices are multiplied by the value of each amenity in each city j, relative to the overall mean, in order to obtain QoL indices for the 103 Italian provinces.

RESULTS This section presents the results of the empirical analysis. It starts by describing the implicit prices estimated from the housing price and wage hedonic equations. It then considers QoL indices for the 103 province capitals, both overall and by individual domain. Finally, it examines the relationship between QoL and subjective wellbeing across cities. Implicit prices of amenities

Table 2 reports estimation results for equations (7) and (8). The model explains more than 70% of the variability of housing prices (column 1), while the marginal explanatory power of local amenities is about 8%. The amenity variables are jointly significant (F = 10.51, p < 0.00), and two-thirds of the amenities are individually statistically significant. All the amenity coefficients have the expected sign. Controlling for structural and neighbourhood characteristics, housing prices are

1409

Table 2. Parameter estimates for amenities

Precipitation (mm/month) Maximum temperature in January Humidity in July (%) Coast (dummy) Green areas (%) Air pollution (agents) Teacher-to-pupil ratio (%) Transport infrastructure (index) Cultural infrastructure (index) Violent crime (per 1000) Unemployment rate (%) Urban density (per km) R2 R2 (controls only) Number of observations

Housing prices

Wages

−0.002 (−1.218) 0.025** (2.159) −0.009** (−2.170) 0.143** (2.105) 0.004* (1.948) −0.010 (−1.546) 0.018 (0.527) 0.004*** (3.147) 0.000 (1.439) −0.037** (−2.312) −0.028*** (−4.560) 0.045*** (2.704)

−0.000 (−0.690) −0.003 (−1.602) −0.002*** (−3.136) 0.002 (0.258) 0.000 (0.861) 0.001 (1.392) 0.015*** (3.113) 0.001*** (4.536) −0.000 (−0.085) 0.001 (0.316) −0.003*** (−3.511) 0.004* (1.939)

0.71 0.62 151 493

0.60 0.59 157 717

Note: Dependent variable: log house prices (column 1) and log wages (column 2). Ordinary least squares (OLS) estimates, t-statistics are reported in parentheses (clustered standard errors allowing for correlation within cities). *Significance at the 0.10 level (** at 0.05, *** at 0.01). The set of regressors at city level also includes population size and a regional capital dummy variable. The housing equation also includes structural and neighbourhood characteristics, as described in the fourth section. The wage equation also includes worker and firm characteristics, as described in the fourth section.

higher in warmer cities, with lower humidity and lower average precipitations.7 As regards environmental conditions, housing prices are higher in cities located on the coast, with less pollution and more green areas. Focusing on services, positive housing price differentials are observed in cities with a higher teacher-to-pupil ratio, better transport and more cultural infrastructure. Regarding social conditions, housing prices are lower in cities with higher crime and unemployment rates and lower population density. The model explains 60% of the variability of wages (column 2), while the marginal explanatory power of local amenities is only 1%. The amenity variables are jointly significant (F = 43.70, p < 0.00), and five coefficients are individually significant. Two-thirds of the estimated coefficients in the wage equation have the same sign as in the housing equation, thus implying implicit prices with opposite signs.8 As discussed in the second section, the finding of counter-intuitive implicit prices in the wage equation can be interpreted as an indication that local amenities affect the location

Emilio Colombo et al.

1410

Table 3. Implicit prices of amenities (€/year)

Precipitation Temperature Humidity Coast Green areas Air pollution Teacher-to-pupil ratio Transport Cultural infrastructure Violent crime Unemployment rate Urban density

Housing

Wage

Total

QoL change

−27.8 385.4 −118.1 2357.3 43.3 −169.7 486.8 67.0 8.5 −673.9 −421.1 870.6

7.1 67.9 67.5 −53.2 −5.6 −36.9 −428.7 −21.4 0.4 −16.9 99.5 −147.4

−20.6 453.2 −50.6 2304.1 37.7 −206.7 58.2 45.6 9.0 −690.8 −321.6 723.2

−461.3 1662.1 −264.4 1096.7 423.0 −540.6 52.7 1096.7 694.9 −1004.0 −2404.1 998.7

Note: Figures reported are the compensating differentials, expressed in euros at constant 2004 prices, of a one unit change in the corresponding amenity, based on the estimates in Table 2. Quality of life (QoL) changes (column 4) are the changes in QoL associated with a 1 SD (standard deviation) in the corresponding amenity.

decisions of both households and firms, so that the net effect on wages of the presence of local amenities is ambiguous. For example, to the extent that the quality of education represents an amenity not only for households but also for firms, higher teacher-topupil ratios in a given city will result in both higher labour supply by households and higher labour demand by firms. Another possible explanation of the relatively small effects of amenities on wages is the high degree of centralization of the Italian labour market. The data set refers to wages for employees, regulated by nationwide contracts that impose strong limitations to regional wage differences for a given occupation. This implies that wages may not fully adjust to compensate for differences in amenities across cities. The choice of including only employees in the sample, while excluding selfemployed workers, was made in order to obtain higher reliability of statistical information concerning declared wages. The empirical evidence indicates that tax evasion is low for employees and much higher for

Table 4. Quality-of-life (QoL) index ranking, overall Rank 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

City

QoL index

Rank

City

QoL index

Rank

City

QoL index

Pisa Firenze Ancona Trieste Imperia Bologna Milano Pesaro Massa Venezia Prato Bergamo Livorno Treviso Lucca Grosseto Siena La Spezia Genova Lodi Cagliari Chieti Vicenza Reggio Emilia Lecco Padova Varese Pescara Forli Bari Napoli Macerata Roma Latina Pavia

8441 6401 5685 4943 4711 4261 3916 3626 3576 3537 3119 3009 2802 2595 2500 2376 2291 2284 2267 2064 2005 1996 1950 1877 1855 1667 1657 1531 1499 1464 1451 1329 1197 1181 1132

36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70

Como Savona Lecce Arezzo Parma Pistoia Ragusa Ferrara Cremona Brescia Gorizia Ravenna Oristano Mantova Bolzano Rimini Torino Biella Salerno Belluno Avellino Piacenza Aosta Sassari Cuneo Verona Teramo Brindisi Trento Modena Pordenone Sondrio Viterbo Udine Caserta

1096 1022 1021 837 645 511 496 428 413 412 397 333 12 −18 −113 −134 −137 −137 −188 −214 −227 −254 −323 −402 −408 −489 −578 −585 −804 −810 −850 −958 −977 −989 −1004

71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103

Palermo Catanzaro Catania L’Aquila Rovigo Perugia Agrigento Novara Reggio Calabria Siracusa Ascoli Piceno Messina Vibo Valentia Matera Frosinone Benevento Taranto Vercelli Alessandria Trapani Terni Rieti Nuoro Asti Verbania Cosenza Isernia Campobasso Crotone Foggia Caltanissetta Potenza Enna

−1114 −1325 −1363 −1376 −1526 −1584 −1743 −1799 −1833 −1841 −1842 −1956 −1986 −1992 −2118 −2369 −2415 −2499 −2570 −2614 −2933 −3103 −3129 −3423 −3715 −3973 −4267 −4684 −5124 −5201 −5518 −5593 −6693

Note: Figures refer to the municipal area of province capitals, rather than the whole provincial territory. Sources: Istituto Nazionale di Statistica (ISTAT), Istituto Nazionale Previdenza Sociale (INPS) and Agenzia del Territorio.

Hedonic Estimates of Quality of Life in Italian Cities the self-employed (for example, BORDIGNON and ZANARDI , 1997). As a consequence, the wage equation would not be informative for the latter category of workers. Table 3 reports the estimated full implicit prices of local amenities. These can be interpreted as the monetary amounts, expressed in euros at constant 2004 prices, that households would be willing to pay annually for a one unit change in the corresponding amenity. For example, the housing price component for temperature (column 1) indicates that households are willing to pay €385.4 per year for an additional temperature of 1°C in January. The wage component (column 2) is also positive (67.9), and the full implicit price (column 3) is therefore €453.2.9 This indicates that the representative household is willing to pay €453.2 per year, through higher housing costs and lower wages, for living in a city with an additional temperature of 1°C in January. The comparison between columns 1 and 2 indicates that the implicit prices from the housing equation are generally larger than those from the wage equation. As a result, all the full implicit prices have the expected sign: beneficial amenities have positive prices (temperature in January, coastal location, green areas, teacher-topupil ratio, transport and culture infrastructure, urban density), while the opposite holds for disamenities (precipitation, humidity, air pollution, crime and unemployment). The estimated implicit prices reported in Table 3 are not directly comparable across amenities, as they are measured in different units. In order to compare the relative size of the effects of different amenities, Table 3 (column 4) also reports the change in QoL associated with 1 SD (standard deviation) in the corresponding amenity using the full implicit prices. The results indicate that, among the disamenities, unemployment has the largest effect on QoL, followed by violent crime and air pollution. Among the amenities, temperature in January, transport infrastructure and coastal location have the largest effects on QoL. Quality of life in Italian cities

Table 4 reports the city-ranking for the QoL index based on the full implicit prices. The index is normalized with respect to the country average and it can be interpreted as the amount, in 2004 euros, that a representative household would be willing to pay to live in a city with a given bundle of amenities, relative to a city with the average set of amenities. The results indicate that amenities account for substantial variation in QoL. As shown in Table 4, the city with the highest QoL is Pisa with a score of 8441. This indicates that, on average, households would be willing to pay €8441 for living in a city with a corresponding bundle of amenities, relative to a city with average levels of amenities. This is a considerable compensating differential, when compared with households’ average annual

1411

Fig. 2. Quality-of-life (QoL) index, overall real wage of approximately €30 000 in the sample. Negative values can be interpreted as the monetary price that households would be willing to receive to be compensated for living in a city with a ‘bad’ bundle of amenities. At the bottom of the ranking is Enna in Sicily, with an overall QoL index of –6693. This indicates that the representative household would need to receive approximately 20% of its annual income to be compensated for living in a city with a corresponding bundle of amenities. Fig. 2 displays the geographic distribution of QoL across Italian cities. Overall, QoL is higher in cities located in the Centre–North. A clustering of high scores can be observed for cities in the Tuscany and Liguria regions. Cities in the South generally display a relatively worse QoL, with clustering of low scores in the cities of the Basilicata, Calabria and Sicily regions. QoL is generally better in large cities (Bologna, Firenze, Venezia) or medium-sized cities (for example, Pisa, Trieste, Imperia, Ancona, Siena, Pesaro and Parma) in the Centre–North. The largest cities display relatively high scores, with Milano and Roma ranking seven and thirty-three, respectively. The overall index presented in Table 4 is constructed using the full implicit prices to weigh all the twelve amenities considered in the analysis. Fig. 3 reports the geographical distribution of the four domain-specific QoL indices. A clear North–South divide can be observed. Cities in the North generally fare better with respect to services and social conditions, while relatively worse with respect to climatic and environmental

1412

Emilio Colombo et al.

Fig. 3. Quality-of-life (QoL) indices, by domain conditions. The opposite applies to the South. Overall, cities located in the Centre–North are generally characterized by relatively high scores in all the domains considered.

Table 5 displays pair-wise correlations between overall and domain-specific QoL indices, with the corresponding significance levels given in parentheses. Climatic and environmental conditions are positively

Hedonic Estimates of Quality of Life in Italian Cities Table 5. Cross-correlations among quality-of-life (QoL) overall and domain indices Variables

Weather

Environment

Services

Environment

0.51 (0.00) −0.55 (0.00) −0.83 (0.00) −0.15 (−0.12)

−0.24 (−0.01) −0.49 (0.00) 0.27 (−0.01)

0.55 (0.00) 0.64 (0.00)

Services Society Overall

1413

regions. Several recent papers use subjective wellbeing data to estimate the value of local amenities (for example, LUECHINGER , 2009; FREY et al., 2010; FERREIRA and MORO , 2010). This subsection examines the relationship between the hedonic QoL indices described above and subjective well-being across the 103 Italian provinces. Two specific questions are asked.

Society

. .

0.51 (0.00)

Is QoL related to well-being? What domains of QoL (types of local amenities) are more relevant for individual well-being?

The source for the subjective well-being data is the Survey of Household Income and Wealth (SHIW), conducted on a biennial basis by the Bank of Italy on a representative sample of about 8000 Italian households. The question about the evaluation of happiness is contained in a section that was only present in the 2004, 2006 and 2008 waves. The question on happiness was formulated as follows: ‘Considering every aspect of your life, how happy do you feel?’. Answers were on a scale between 1 (extremely unhappy) and 10 (exceptionally happy). This paper considers average scores at province level for the three available waves. Table 6 presents estimation results for cross-city regressions of subjective well-being on QoL, overall and by individual domain. Overall (column 1), QoL is positively related to subjective well-being, and the relationship is significant at the 10% level. The estimated coefficient indicates that a €1000 improvement in QoL is associated with a 0.02 increase in selfreported happiness, on a scale between 1 and 10. The explanatory power of overall QoL for subjective well-being is, however, very low (R2 = 0.01). Columns (2) to (5) report coefficient estimates obtained separately for individual QoL domains. The four estimated coefficients are all individually strongly significant. The coefficients for the weather and environment indices have a negative sign, while the opposite holds for the services and society indices.

Note: Figures reported are pair-wise correlations between QoL indices across the municipal areas of the 103 Italian province capitals (p-values in parentheses). Sources: Istituto Nazionale di Statistica (ISTAT), Istituto Nazionale Previdenza Sociale (INPS) and Agenzia del Territorio.

correlated. Similarly, services and social conditions are strongly positively related. However, the climatic and environmental indices are negatively related to both services and social conditions. As a result, the domainspecific indicators are related to the overall index in various degrees. While climatic conditions are negatively correlated with overall QoL (although not significantly), environmental conditions are positively but weakly related to QoL. Services and social conditions are positively and strongly related to overall QoL.

Quality of life and subjective well-being

The relationship between local amenities and subjective well-being has recently received much attention in the literature. OSWALD and WU (2010) found a strong and significant relationship between QoL, as measured by hedonic indices, and subjective well-being across states in the United States. MORO et al. (2008) proposed a new approach to construct QoL indices that relies on subjective well-being data, with an application to Irish

Table 6. Quality of life (QoL) and subjective well-being across Italian cities (1) Overall

0.02* (1.77)

Weather

(2)

−0.05*** (−2.70)

Environment

(3)

(4)

(6)

0.05*** (3.28) 0.10

0.01 (0.22) −0.03 (−1.05) 0.00 (0.05) 0.04** (2.02) 0.11

−0.07** (−2.31)

Services

0.05** (2.04)

Society R2

(5)

0.01

0.07

0.06

0.03

Note: Dependent variable: subjective well-being. QoL indicators are divided by 1000, so that reported figures should be interpreted as the change in self-reported happiness, on a scale between 1 and 10, associated with a €1000 change in QoL. Ordinary least squares (OLS) estimates are based on 103 observations; t-statistics are reported in parentheses based on heteroskedasticity-robust standard errors. *Significance at the 0.10 level (** at 0.05, *** at 0.01).

Emilio Colombo et al.

1414

This reflects the fact that subjective well-being is generally lower in southern cities, characterized by better climatic and environmental conditions and worse services and social conditions. Indeed, when controlling for all dimensions of QoL jointly (Table 6, column 6), there is no significant association between weather, or environment, and subjective well-being. The society index is the only QoL component significantly related to subjective well-being in the multivariate regression. This reflects the fact that crime and unemployment have the strongest negative effects on QoL among the (dis-)amenities considered.

CONCLUSIONS This paper has presented the first application of the hedonic approach to measure and compare QoL across Italian cities. It used micro-level data for housing and labour markets, together with city-level data on local amenities, to estimate compensating differentials for local amenities. It obtained full implicit prices for local amenities and used them to construct QoL indices for the municipalities of the 103 Italian provinces. The analysis focused on amenities, and QoL indices, within four domains: weather, environment, services and society. It was found that the presence of local amenities results in large compensating differentials in the housing market. The effects of amenities on wage differentials are, instead, relatively small. This is likely to reflect not only the productivity effects of amenities, but also the relatively high degree of centralization of the Italian labour market. Nevertheless, all the estimated full implicit prices have the expected sign. This is an important result, as it indicates that, despite the relative rigidity of prices and wages in the Italian economy, compensating differentials in housing and labour markets can be effectively used to price local amenities and construct QoL indices based on revealed preferences. Overall, local amenities account for substantial variation in QoL, with differences across cities that are quantitatively relevant. The representative household would be willing to pay about onequarter of its average annual income in order to live in the city with the best bundle of amenities. QoL is better in large and medium-sized cities of the Centre–North, while cities located in the South generally display a worse QoL. Focusing on QoL domains, cities in the North fare better with respect to services and social conditions, while relatively worse for climatic and environmental conditions. The opposite pattern applies to cities located in the South. Cities in the Centre–North are generally characterized by relatively high scores in all the domains considered. Finally, it was found that QoL is positively and significantly related to subjective well-being across Italian cities. Overall, the comparison of QoL across cities on the basis of revealed preferences provides objective

information that is particularly relevant to inform the debate on fiscal federalism, while also indicating specific directions for economic, urban and environmental policy. More generally, the analysis highlights the importance for the municipal, regional and central governments to establish information systems for monitoring the determinants of urban QoL. This would significantly improve one’s ability to detect disparities in QoL across cities and identify appropriate policy actions. Acknowledgements – The authors gratefully acknowledge the Agenzia del Territorio (Osservatorio del Mercato Immobiliare) for housing market data; the Fondazione Rodolfo De Benedetti for labour market data; and the Bank of Italy for subjective well-being data. Financial support by the Italian Ministry of University and Research is also gratefully acknowledged. The authors thank Luciano Canova for his contribution to this project; and participants at the 2010 meetings of the American Real Estate and Urban Economics Association (AREUEA); the International Society for Quality of Life Studies (ISQOLS); and the Italian Association of Regional Sciences (AISRE) for helpful comments.

APPENDIX A: DATA Information on city characteristics and amenities was collected from several sources, as detailed in Table A1. Data for weather conditions were obtained from ISTAT and other specific sources (http://www. ilmeteo.it). The variables considered referred to maximum temperature in January, monthly millimetres of precipitation and average humidity in July. Indicators of environmental conditions were from ISTAT and included the share of green areas for the total city area and the number of air-polluting agents. A dummy variable identified cities bordering the sea (the dummy is coded 1 if the centre of the city is fewer than 10 km from the coast). Services included education, transport and culture. For education the teacher-to-pupil ratio (the average of primary and secondary schools), from the Italian Ministry of Education, was used. For transport an accessibility measure (a multimodal index that considers accessibility by air, train and car, equal to 100 for the European average; source: European Observation Network for Territorial Development and Cohesion (ESPON) project; http://www.espon.eu) was used. Finally, cultural services with an index of the cultural infrastructure of the city (accounting for museums, theatres, cinemas and libraries) were measured. The index was set to 100 for the Italian average (source: Istituto Tagliacarne; http://www. tagliacarne.it). The number of violent crime acts per capita was from the Ministry of Justice. The unemployment rate and urban density (1000 inhabitants/km2) were from ISTAT. Also included among the control variables were population size and a dummy for regional capitals (ISTAT).

Hedonic Estimates of Quality of Life in Italian Cities Housing market data were from the Real Estate Observatory of the Agenzia del Territorio (OMI-AT), a public agency within the Ministry of the Economy. The data on individual house transactions in Italian cities (municipalities of province capitals) from 2004 to 2010 were selected. In addition to sale prices, the data set provides a detailed description of housing characteristics, such as surface area, number of bathrooms, floor level, number of garages, location (centre, semicentre, suburb), quality of building (good, average, bad), quality of the area, distance from transportation and services. Table A2 provides descriptive statistics for housing characteristics. Labour market data were from the Istituto Nazionale Previdenza Sociale (INPS – Italian National Social Security Institute). The Employees’ archive, containing information on workers employed in the private sector who are insured with the INPS, was used. Wages refer to private sector workers’ annual earnings. In addition, the data set provides information on the level of occupation, whether the job is full-time or part-time, contract length, province of work, and sector of economic activity. Personal and demographic characteristics include gender, age, nationality and province of residence. Table A3 reports descriptive statistics for workers and firm characteristics.

Table A1. Description and sources of variables Variable Precipitation

Temperature Humidity Coast Green areas Air pollution Education

Transport Culture Crime Unemployment Urban density

Description Millimetres of rain per month, average over twelve months. Sources: http://www.ilmeteo.it and ISTAT Average temperature over the year. Sources: http://www.ilmeteo.it and ISTAT Air humidity (%), yearly average. Sources: http:// www.ilmeteo.it and ISTAT Coast dummy equal to 1 if a city is within 10 km from the coast. Source: authors’ calculation Percentage of urban green over urban area. Source: ISTAT Number of polluting agents in the air. Source: ISTAT Teacher-to-pupil ratio (%), average of primary, secondary and upper secondary schools. Source: Italian Ministry of Education Multimodal accessibility index (train, air, car), ESPON space = 100. Source: ESPON Index of cultural infrastructure, Italian average = 100. Source: Istituto Tagliacarne Number of violent crimes per 1000 inhabitants. Source: ISTAT Percentage rate. Source: ISTAT Percentage rate. Source: ISTAT

Note: ESPON, European Observation Network for Territorial Development and Cohesion; ISTAT, Istituto Nazionale di Statistica.

1415

Table A2. Housing and neighbourhood characteristics: descriptive statistics Variable

Mean

Standard deviation

Minimum

Maximum

Total surface Age of building Floor Number of floors Number of bathrooms Penthouse Elevator Housing type: flat Housing type: low-cost Housing type: luxury Housing type: detached house Conditions: normal Conditions: good Conditions: poor Features: exclusive area Features: garage Features: balcony Features: attic Features: basement Location: central Location: noncentral Location: rural Location: semicentral Location: suburban Location quality: very poor Location quality: normal Location quality: very good Public transport: absent Public transport: distant Public transport: near Services: absent Services: distant Services: near Commercial services: absent Commercial services: distant Commercial services: near

4.61 38.41 2.10 4.64 3.20

0.39 33.91 1.74 2.27 2.65

2.56 0 0 0 1

6.21 505 27 29 7

0.01 0.55 0.74 0.22

0.11 0.50 0.44 0.41

0 0 0 0

1 1 1 1

0.01

0.09

0

1

0.03

0.17

0

1

0.87

0.34

0

1

0.11 0.02 0.07

0.31 0.15 0.26

0 0 0

1 1 1

0.17 0.11 0.04 0.23

0.37 0.31 0.19 0.42

0 0 0 0

1 1 1 1

0.24 0.31

0.43 0.46

0 0

1 1

0.01 0.28

0.04 0.45

0 0

1 1

0.18

0.38

0

1

0.01

0.10

0

1

0.90

0.30

0

1

0.09

0.29

0

1

0.01

0.12

0

1

0.10

0.30

0

1

0.88

0.32

0

1

0.02 0.22 0.76 0.01

0.15 0.41 0.43 0.10

0 0 0 0

1 1 1 1

0.13

0.34

0

1

0.86

0.35

0

1

Note: Number of observations = 158 512. Source: Real Estate Observatory of the Agenzia del Territorio (AT).

Emilio Colombo et al.

1416

Table A3. Worker–firm characteristics: descriptive statistics Variable Male Age Age squared Number of paid days Nationality: Asia Nationality: Africa Nationality: Latin America Executive Manager and white collar Blue collar Apprentice Temporary contract Small firm Medium firm Large firm Agriculture Electricity Chemistry Metalworking Food, textile, wood Building materials Commerce and services Transport and communications Credit, insurance Public administration

Mean

Standard deviation

Minimum

Maximum

0.66 37.65 1520.61 291.32 0.01 0.02 0.01 0.01 0.39 0.56 0.04 0.06 0.27 0.44 0.26 0.01 0.02 0.06 0.20 0.17 0.08 0.19 0.06 0.12 0.10

0.47 10.15 801.41 37.27 0.09 0.14 0.06 0.06 0.49 0.50 0.21 0.23 0.44 0.50 0.44 0.06 0.13 0.24 0.40 0.37 0.27 0.39 0.23 0.33 0.30

0 16 256 151 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 75 5625 365 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Note: Number of observations = 168 255. Source: Istituto Nazionale Previdenza Sociale (INPS).

NOTES 1. For earlier reviews of alternative approaches to the measurement of QoL, see also LUGER (1996), DIENER and SUH (1997), and GYOURKO et al. (1999). 2. MADDISON and BIGANO (2003) estimated the marginal willingness to pay for climate variables in Italian cities. SCHIFINI D’ ANDREA (1998) relied on socio-economic indicators to assess QoL in Italy in a comparative perspective. CICERCHIA (1996) proposed a set of objective and subjective indicators for the measurement of QoL in urban areas. NUVOLATI (2003) applies the social indicators approach to analyse the evolution of QoL in the 103 Italian provinces from 1989 to 2001. 3. Note that equation (1) assumes inelastic demand for housing. This assumption is supported by the empirical evidence (for instance, HANUSHEK and QUIGLEY , 1980), although recent studies for the United States estimate that the after-tax price elasticity of housing demand is about –0.5, and the income elasticity about 0.25 (SINAI , 2008). 4. This conversion is based on the average number of hours worked in a part-time job that generally ranges between four and six (about two-thirds of the daily number of hours worked in a full-time job). 5. As suggested by ROTEMBERG and SALONER (2000), higher city density results in a labour market pooling where workers face many prospective employers, thus obtaining stronger bargaining power. More generally,

6. 7.

8.

9.

GLAESER (1999) argued that urban density increases the rate of interaction between people, who can thus learn and accumulate human capital faster. On the negative side, BENABOU (1993) argued that social interactions in dense areas could lead to social and occupational segregation. GLAESER (1998) emphasized that people living in more dense cities are less likely to trust others (urban anonymity) and this may result in lower social capital. The coefficient of variation is 0.29 and 0.008 for housing and non-housing price differentials, respectively. The authors considered the use of a quadratic specification for the three weather indicators, as by RAPPAPORT (2007). The explanatory power of the models is virtually unchanged. Adjusted R2 is 0.72, as opposed to 0.71, for the housing price equation and unchanged at 0.60 for the wage equation. Similar results are obtained by BLOMQUIST et al. (1988) and GYOURKO and TRACY (1991) for US urban areas and, more recently, by BUETTNER and EBERTZ (2009) for German counties. Note that the wage component of the implicit price in dwj column (2) is obtained as − (see equation 5). This daij explains the sign reversal with respect to the coefficient estimates reported in Table 2, column 2. For more details on the calculation of the full implicit prices, see the fourth section.

REFERENCES ALBOUY D. (2008) Are Big Cities Really Bad Places to Live? Improving Quality of Life Estimates Across Cities. NBER Working Paper Number 14472. National Bureau of Economic Research (NBER), Cambridge, MA.

Hedonic Estimates of Quality of Life in Italian Cities

1417

ALBOUY D. (2009) The unequal geographic burden of federal taxation, Journal of Political Economy 117(4), 635–667. BANK OF ITALY (2008) Survey of Household Income and Wealth, 2008. Bank of Italy, Rome. BENABOU R. (1993) Working of a city: location, education and production, Quarterly Journal of Economics 108(3), 619–652. BERGER M., BLOMQUIST G. and SABIRIANOVA K. (2008) Compensating differentials in emerging labor and housing markets: Estimates of quality of life in Russian cities, Journal of Urban Economics 63, 25–55. BLOMQUIST G. (2007) Measuring quality of life, in ARNOTT R. and MCMILLEN D. (Eds) A Companion to Urban Economics, pp. 479–501. Blackwell-Synergy, Oxford. BLOMQUIST G., BERGER M. and HOEHN J. (1988) New estimates of quality of life in urban areas, American Economic Review 78(1), 89–107. BORDIGNON M. and ZANARDI A. (1997) Tax evasion in Italy, Giornale degli Economisti e Annali di Economia 56, 169–210. BUETTNER T. and EBERTZ A. (2009) Quality of life in the regions: results for German counties, Annals of Regional Science 43, 89–112. CHESHIRE P. and SHEPPARD S. (1995) On the price of land and the value of amenities, Economica 62, 247–267. CICERCHIA A. (1996) Indicators for the measurement of the quality of urban life, Social Indicators Research 39(3), 321–358. COSTA D. and KAHN M. (2003) The rising price of nonmarket goods, American Economic Review 93(2), 227–232. DELLER S., TSAI T., MARCOUILLER D. and ENGLISH D. (2001) The role of amenities and quality of life in rural economic growth, American Journal of Agricultural Economics 83(2), 352–365. DIENER E. and SUH E. (1997) Measuring quality of life: economic, social and subjective indicators, Social Indicators Research 40, 189–216. ETZO I. (2011) The determinants of recent interregional migration flows in Italy: a panel data analysis, Journal of Regional Science 51(5), 863–1032. EZZET-LOFSTROM R. (2004) Valuation of metropolitan quality of life in wages and rents, Cityscape: A Journal of Policy Development and Research 7(1), 23–37. FERREIRA S. and MORO M. (2010) On the use of subjective well-being data for environmental valuation, Environmental and Resource Economics 46(3), 249–273. FREY B., LUECHINGER S. and STUTZER A. (2010) The life satisfaction approach to environmental valuation, Annual Review of Resource Economics 2(1), 139–160. FURCERI D. (2006) Does labour respond to cyclical fluctuations? The case of Italy, Applied Economic Letters 13, 135–139. GABRIEL S. and ROSENTHAL S. (2004) Quality of the business environment versus the quality of life: do firms and households like the same cities?, Review of Economics and Statistics 86, 438–444. GIANNIAS D. (1998) A quality of life based ranking of Canadian cities, Urban Studies 35, 2241–2251. GLAESER E. (1998) Are cities dying?, Journal of Economic Perspectives 12(2), 139–160. GLAESER E. (1999) Learning in cities, Journal of Urban Economics 46(2), 254–277. GYOURKO J. and TRACY J. (1989) The importance of local fiscal conditions in analyzing local labor markets, Journal of Political Economy 97, 1208–1231. GYOURKO J. and TRACY J. (1991) The structure of local public finance and the quality of life, Journal of Political Economy 99, 774–806. GYOURKO J., KAHN M. and TRACY J. (1999) Quality of life and environmental comparisons, in MILLS E. S. and CHESHIRE P. (Eds) The Handbook of Applied Urban Economics, pp. 1413–1454. North-Holland, Amsterdam. HANUSHEK E. and QUIGLEY J. (1980) What is the price elasticity of housing demand?, Review of Economics and Statistics 62(3), 449–454. HERZOG H. and SCHLOTTMANN A. (1993) Valuing amenities and disamenities of urban scale: can bigger be better?, Journal of Regional Science 33(2), 145–166. ISTITUTO NAZIONALE DI STATISTICA (ISTAT) (2009) Le differenze nel livello dei prezzi al consumo tra i capoluoghi delle regioni italiane – Anno 2009. ISTAT, Rome. KAHN M. (1995) A revealed preference approach to ranking city quality of life, Journal of Urban Economics 38, 221–235. LAMBIRI D., BIAGI B. and ROYUELA V. (2007) Quality of life in the economic and urban economic literature, Social Indicators Research 84, 1–25. LUECHINGER S. (2009) Valuing air quality using the life satisfaction approach, Economic Journal 119(536), 482–515. LUGER M. (1996) Quality-of-life differences and urban and regional outcomes: a review, Housing Policy Debate 7, 749–771. MADDISON D. and BIGANO A. (2003) The amenity value of the Italian climate, Journal of Environmental Economics and Management 45(2), 319–332. MORO M., BRERETON F., FERREIRA S. and CLINCH J. (2008) Ranking quality of life using subjective well-being data, Ecological Economics 65(3), 448–460. NUVOLATI G. (2003) Socioeconomic development and quality of life in Italy, in SIRGY J., RAHTZ D. and COSKUN SAMLI A. Advances in Quality-of-Life Theory and Research, pp. 81–98. Kluwer, Dordrecht. OSWALD A. and WU S. (2010) Objective confirmation of subjective measures of human well-being: evidence from the USA, Science 327(5965), 576–579. RAPPAPORT J. (2007) Moving to nice weather, Regional Science and Urban Economics 37(3), 375–398. RAPPAPORT J. (2008) Consumption amenities and city population density, Regional Science and Urban Economics 38(6), 533–552. RAPPAPORT J. (2009) The increasing importance of quality of life, Journal of Economic Geography 9(6), 779–804. ROBACK J. (1982) Wages, rents, and the quality of life, Journal of Political Economy 90(6), 1257–1278.

1418

Emilio Colombo et al.

ROSEN S. (1979) Wage-based indexes of urban quality of life, in MIESZKOWSI P. and STRATZHEIM M. (Eds) Current Issues in Urban Economics, pp. 74–104. John Hopkins Press, Baltimore, MD. ROTEMBERG J. and SALONER G. (2000) Competition and human capital accumulation: a theory of interregional specialization and trade, Regional Science and Urban Economics 30(4), 373–404. SCHIFINI D’ANDREA S. (1998) Italian quality of life, Social Indicators Research 44, 5–39. SHAPIRO J. (2006) Smart cities: quality of life, productivity, and the growth effects of human capital, Review of Economics and Statistics 88(2), 324–335. SHULTZ S. and KING D. (2001) The use of census data for hedonic price estimates of open-space amenities and land use, Journal of Real Estate Finance and Economics 22, 239–252. SINAI T. (2008) Urban housing demand, in DURLAUF S. and BLUME L. (Eds) New Palgrave Dictionary of Economics, 2nd Edn. Macmillan, London. SRINIVASAN S. and STEWART G. (2004) The quality of life in England and Wales, Oxford Bulletin of Economics and Statistics 66(1), 1–22. WINKELMANN L. and WINKELMANN R. (1998) Why are the unemployed so unhappy? Evidence from panel data, Economica 65(257), 1–15. WINTERS J. (2010) Differences in Quality of Life Estimates Using Rents and Home Values. IZA Discussion Papers Number 6703. Institute for the Study of Labor (IZA), Bonn.