Methods for Imputed Rents Estimation and Their Impact on Income Distribution

Rivista di Statistica Ufficiale. Vol. 1, pp. 17-38, 2009

22 Pages Posted: 25 Sep 2010

See all articles by Claudio Ceccarelli

Claudio Ceccarelli

National Institute of Statistics (Istat)

Andrea Cutillo

National Institute of Statistics (Istat)

Davide Di Laurea

National Institute of Statistics (Istat)

Date Written: 2009

Abstract

In the relevant literature it is well established that imputed rents - a non-monetary income source for all households whose living accommodation is not rented at a market price - has to be accounted for in order to provide a more accurate and comparable measure of the total disposable household income. A more controversial issue is about the choice among distinct estimation approaches and how to operationalise them.

This paper aims to analyse the impact of different estimation methods for imputed rents on the distribution of disposable income as well as on inequality measures, using data from EUSILC. Subjective rents - self-assessed values by the respondents - are compared with predicted values coming from econometric techniques, i.e. an hedonic regression model, being both methods coherent with the ‘rental equivalence value’ approach.

Three main issues are assessed: i) pros and cons of each method; ii) if the tenure discount (the length of tenure), which turns to be a relevant factor in determining the observed values of the market rents, has to be taken into explicit account in predicting imputed rents; iii) how to prevent biases on model predictions eventually arising if a selection process between tenants at a market price and households living with different accommodation occurs.

Sommario Questo lavoro presenta un’analisi sull’affitto figurativo condotta sui dati dell’indagine EUSILC; questa posta figurativa, stimata per chi non vive in abitazioni in affitto a prezzi di mercato, va considerata per ottenere una nozione di reddito disponibile più accurata nelle analisi su distribuzione del reddito, disuguaglianza e povertà. Vengono messe a confronto diverse metodologie, in particolare le stime derivanti dall’autovalutazione dei rispondenti con quelle derivanti da tecniche di regressione econometrica secondo la teoria della domanda edonica. Tre sono le principali questioni affrontate: quale sia la metodologia più corretta; se lo sconto temporale che si osserva tra gli affittuari debba essere considerato anche nel calcolo dell’affitto figurativo; se esiste un processo di selezione sulla condizione abitativa che può portare a stime distorte degli affitti figurativi. Vengono infine presentati alcuni risultati relativi all’incidenza del rischio di povertà e alla concentrazione dei redditi una volta che si considera questa posta come parte del reddito disponibile.

Keywords: Affitti figurativi, distribuzione del reddito; imputed rent, sample selection bias, income distribution

JEL Classification: C24, D31, D33

Suggested Citation

Ceccarelli, Claudio and Cutillo, Andrea and Di Laurea, Davide, Methods for Imputed Rents Estimation and Their Impact on Income Distribution (2009). Rivista di Statistica Ufficiale. Vol. 1, pp. 17-38, 2009, Available at SSRN: https://ssrn.com/abstract=1681911

Claudio Ceccarelli

National Institute of Statistics (Istat) ( email )

Via Cesare Balbo 16
00184 Rome, 0185
Italy

Andrea Cutillo

National Institute of Statistics (Istat)

Via Cesare Balbo 16
00184 Rome, 0185
Italy

Davide Di Laurea (Contact Author)

National Institute of Statistics (Istat) ( email )

Via Cesare Balbo 16
00184 Rome, 0185
Italy