The impact of the introduction of Italian property tax on urban development: a regional regression model

Abstract

Building on and empirically enlarging a previous study on aggregate Italian national data, this article tests whether the introduction of the property tax into the Italian system has dampened construction activity, as proxied by building permits. The latter are also good indicator of regulatory policy and local government behaviour. The heuristic hypothesis put forward in this article is that, because of concomitant favourable market conditions and the devolution process that began in the 1990s, the introduction of the property tax (ICI) induced municipalities to adopt less tight urban policies to offset budgetary needs and compensate for the reduction in central government transfers. To this end, it estimates an econometric model to verify the impact of the main economic variables on new housing supply. Unlike other studies, this article run an analysis at regional level and test for fixed effects and structural break. Our estimates support our hypothesis, evidencing a time effect. They also confirm that introduction of the ICI tax did not affect the construction sector. Careful attention, therefore, should be given to the issue of whether leaving urban planning and the power to levy property taxes under the same jurisdiction.

Keywords:

• Urban development
• land use
• property tax
• building permits

JEL Classification:

• H71
• R31
• Q24
• R38
• C23

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Salvatore Bimonte, PhD in Agricultural Economics, is Full Professor of Environmental Economics at the Department of Economics and Statistics, University of Siena, Siena, Italy. His main research interests are in the field of Environmental Economics, Sustainable Development, Tourism Economics, Economics of Happiness.

Arsenio Stabile, Msc in Economics, works as technician and research assistant at the Department of Business and Law, University of Siena, Siena, Italy. His main interests are in the field of Econometrics, Mathematics, Data Analysis.

Notes

1 It has been estimated that in 2008 the average market value of a dwelling was about 3.7 times the corresponding cadastral value (Agenzia del Territorio, 2010).

2 The refusal rate of planning applications is normally used as a measure of regulatory restrictiveness (Bramley, 1998; Hilber & Vermeulen, 2012).

3 We refer the reader to the paper of Bimonte & Stabile (2105) for a more in-depth description of the reform and municipalities’ budgetary situation.

4 In 2002 ICI accounted for the 68% of total local tax revenues (Guerrieri, 2003).

5 The reform of cadastre is a critical political issue in Italy. Over the last 20 years, various Italian governments tried to revise it, but all legislative initiatives failed.

6 Property tax based on market value of houses may cause municipal fiscal stress. Indeed, a negative cycle in the real estate market causes a decrease in the property values. For cities, this means less revenue or higher tax effort (Kim, 2017).

7 However, it is worth noting that in Italy there have been various building amnesties, notably in 1985, 1994 and 2003.

8 In a recent (February 2014) article, The Guardian denounced the ‘scandal’ of empty properties in Europe; http://www.theguardian.com/society/2014/feb/23/europe-11m-empty-properties-enough-house-homeless-continent-twice.

9 See Bernardi & Poggio (2004) and (ISTAT, various years).

10 For a more in-depth explanation of the reform, see Bimonte & Stabile (2015).

11 Eraslan (2016) evidenced that expectation of future increase in housing prices fuels housing demand and prices. Coupled with other factors, it fosters housing market fluctuations.

12 In her real-time bubble detection method for the housing market, Shi (2017) distinguishes between changes induced by fundamentals dynamics and a housing price bubble.

13 Our model is expressed in terms of changes in lagged variables rather than their level. This specification is more appropriate for flow models and more consistent with the time series characteristics of the data that are more traditional specification in levels (Mayer & Somerville, 2000b). It makes it possible to address the econometric problem of non-stationarity of series.

14 To take in due consideration an anonymous referee’s suggestion, we also estimated a model without considering K. In fact, since K_t (stock variable) is the cumulative sum of BP_t (flow variable), this could give rise to causality problem (Tables A1b and A2b).

15 It is worth noting that the usual Hausman test could not be computed because data are heteroskedastic. In this case, under the homoscedasticity hypothesis, the differential variance-covariance matrix is not positive definite and test is invalid (Greene, 2008). We let the reader note that when estimating the model without K the Hausman test confirms our choice to use an RE model (Table A3).

16 When θ is close to 0, RE and OLS estimates tend to coincide. This happens when ${\sigma }_{\mu }^2<{\sigma }_{\epsilon }^2,$ that is, when there is homogeneity.

17 Our choice was based on a research hypothesis. However, in order to avoid any bias, we also tested for different timing. We considered different years (from 1990). 1995 turned out to be the only significant dummy.

18 The effect is even greater (4.55) when we do not consider the dampening induced by the stock effect (Table A4).

19 Just to have an idea, setting 1990 = 100, local tax revenue grew to 258 in 2002 while government transfers passed to 63% (Guerrieri, 2003).

20 The INVIM was progressive with rate varying from 5% to 30% and the tax base depended on the initial value of the property and number of years it was held by the seller.

21 This is a kind of reciprocal agreement whereby local authority grants an adaptation of the urban plan and developers realize their investment, building public infrastructure in exchange.

22 Eraslan (2016) found evidence of this issue in London and the USA, especially in time of uncertainty.