Micro and Macro Resilience Measures of an Economic Crisis

Abstract

Using Italian Households Budget Survey data over the period 1997-2013, a Cragg model in a life-cycle context is specified to compare the consumption behaviour in the pre- and post-crisis time and develop different micro and macro measures of resilience against crisis shocks. Cohort profiles for participation in and for consumption of tourism services in the pre- and post-crisis time are determined so as to explore the households’ resilience by generations. Next, the households’ resilience according to socio-demographic characteristics is addressed. As for the macro-resilience, combining individual expenditure elasticity, we examine whether and how Italian regions have responded to the recent economic crisis.

Appendices

Appendix 1: The Zero-Expenditure Approach to the Demand Model

Tourism is a censored good, that is, not all individuals participate in the tourism market. Consequently, the model in Eq. (2) should be revisited to account for a large proportion of observations with a value of tourism expenditure equal to zero (i.e., a censored variable). In the applied model analysis, we adopt the Cragg approach to the zero-expenditure model (Cragg 1971) because of its flexibility. The ‘double-hurdle’ model assumes that (i) an individual has to desire a positive amount of goods or services (first hurdle: the participation decision), and (ii) there must be favorable circumstances for a positive expenditure to occur (second hurdle: the consumption decision). Formally, our approach integrates the probit model to determine the probability of y^* > 0 and the truncated normal model for given positive values of y^*, as follows:participation decision:

$$ {d}_i^{\ast }={\alpha}^{\prime }{z}_i+{v}_i,\mathrm{where}\ {v}_i\to N\left(0,1\right) $$

(5)

$$ {d}_i=\left\{\begin{array}{cc}1& \mathrm{if}\ {d}_i^{\ast }>0\\ {}0& \mathrm{otherwise},\end{array}\right. $$

(6)

and consumption decision:

$$ {y}_i^{\ast }={\upbeta}^{\prime }{x}_i+{u}_i,\mathrm{where}\ {u}_i\to N\left(0,{\upsigma}^2\right) $$

(7)

$$ {y}_i=\left\{\begin{array}{cc}{y}_i^{\ast }& \mathrm{if}\ {d}_i=1\ \mathrm{and}\ {y}_i^{\ast }>0\\ {}0& \mathrm{otherwise},\end{array}\right. $$

(8)

where x and z are a different set of variables affecting the two decision stages, while the variables are assumed to be uncorrelated with their respective error terms; ${d}_i^{\ast }$ is a latent variable that denotes binary censoring and ${y}_i^{\ast }$ the unobserved latent value of expenditure. Eq. (8) indicates that the observed expenditure y_i is zero, either when there is censoring at zero (${y}_i^{\ast}\le 0$) or when there is faulty reporting, due to some random circumstance. In other words, a positive level of tourism consumption (${y}_i={y}_i^{\ast }$) is observed only if the individual is a potential tourist (d_i = 1) and actually consumes tourism services (${y}_i^{\ast }>0$).

An interesting feature of the Cragg model is that different sets of determinants affect the two hurdles. Separating the two decision stages is particularly relevant in modeling tourism consumption, since the decision to travel can be assumed to be mainly related to social factors, while the decision about how much to spend on a holiday depends on individual budget constraints.¹⁰

Appendix 2: Data Description and Cohort Definition

The empirical analysis has been performed on the data from the Households Budget Survey (HBF), carried out by the Italian Office of Statistics (ISTAT). A sample of 400,473 households was collected over the period 1997 to 2013. With regard to tourism, HBF observes the monthly total amount of expenditure of the household on trips for personal purposes. Following international standards, HBF records expenditure, both for national and international trips on holidays, leisure and recreation, visiting friends and relatives, health treatment, religious activity and pilgrimages, etc. The survey does not provide information about the reason for the trip, so it is impossible to investigate how it affects tourism expenditure. We use ‘tourism expenditure’ below to refer to expenditure on trips for personal purposes. The respondent is a member of the household who reports the tourism expenditure of all members of the family. The survey does not provide information about the number of household members making a trip; but it does indicate if Italian families travel abroad or in the country and if they report an expenditure other than zero. Data on expenditure is supplemented by a rich set of economic, demographic, and sociological variables on Italian households. The HBF survey is performed every year and involves a random sample of the population. A pool of time series of cross sectional observations can thus be set up and groups of households can be followed over time by means of cohort techniques (Deaton 1985). The cohorts have been identified by using the age of the head of the household.

Following Attanasio and Weber (1994), Browning et al. (1985) and Deaton (1985), we group households on the basis of the age of the head of the household, using five-year age band cohorts, and we track the cohorts over time. The age of each household head (i.e. “a”) is defined as the mid-range age of the age-class¹¹ which the household head belongs to, while cohort “c” is defined as c = t - a, where “t” is the year in which the household was interviewed. In Table 7, the definition of cohorts, the head of the household’s age in the first and last year of observation, and the size of the cohorts are reported.

Table 7 Cohort Definition

Appendix 3: Model Estimates

Table 8 Participation and Consumption functions estimates