Compensating Wage Differentials for Living in Sardinia.

Participants complete an online survey in which they evaluate hypothetical job opportunities that differ in location and salary. In each scenario, respondents choose between two job offers: one corresponding to their current or baseline location and another located in an alternative region.

The comparison depends on the respondent’s current place of residence. For individuals living outside Sardinia, the alternative job is located in Sardinia. For individuals living in Sardinia, the alternative job is located outside Sardinia. The salary associated with the alternative job is adjusted across a sequence of choices in order to identify the salary differential required for the respondent to prefer one location over the other.

After completing an initial baseline set of job-choice scenarios, respondents are randomly assigned to an additional condition. In most cases, the job opportunity located in Sardinia is described together with one of three pre-specified institutional improvements that could make living and working in Sardinia more attractive: better air connections with more frequent flights and lower fares to and from Sardinia; shorter waiting times for specialist visits and diagnostic exams; or a program facilitating the job relocation of a partner when moving to Sardinia.

In a smaller share of cases, as a placebo, respondents instead complete additional scenarios without institutional improvements. This placebo condition allows the study to distinguish the effect of institutional improvements from the effect of receiving additional information or completing additional choice tasks.

By comparing responses across baseline, policy, and placebo scenarios, the study estimates individuals’ willingness to live and work in Sardinia and how this willingness changes when institutional and service conditions improve.

2026-03-10

2026-12-31

Baseline willingness to pay (WTP) for Sardinia
The compensating wage differential required for a respondent to choose a job located in Sardinia relative to an alternative location in the baseline job-choice scenarios.

Policy willingness to pay (WTP) for Sardinia
The compensating wage differential required for a respondent to choose a job located in Sardinia relative to an alternative location when the Sardinian job is presented together with an institutional improvement.

The primary outcomes measure the compensating wage differential associated with choosing Sardinia relative to an alternative location.

These outcomes are constructed from a sequence of binary job-choice scenarios. In each scenario, respondents choose between two hypothetical job offers that differ only in location and net monthly salary. One option corresponds to the respondent’s current or baseline location and is associated with a respondent-specific reference salary. This reference salary is constructed as the midpoint of the respondent’s self-reported current net monthly income bracket. The other option corresponds to an alternative location and is associated with a salary that is updated across successive choice tasks.

The relevant alternative location depends on the respondent’s current place of residence. For respondents living outside Sardinia, the alternative job is located in Sardinia. For respondents living in Sardinia, the alternative job is located outside Sardinia. In both cases, the salary attached to the alternative option is adjusted adaptively on the basis of the respondent’s previous choices in order to identify the reservation wage differential between the two locations.

The adaptive procedure produces a lower bound and an upper bound for the respondent’s reservation wage interval. When the final interval is not censored, the main outcome variable is constructed as the midpoint of this final interval. This midpoint is interpreted as the estimated compensating wage differential associated with choosing Sardinia relative to the alternative location. The survey also stores the final lower and upper bounds of the interval and indicators for censoring.

Two primary outcomes are constructed. The first is baseline willingness to pay for Sardinia, derived from the baseline choice sequence. The second is policy willingness to pay for Sardinia, derived from the choice sequence in which the Sardinian option is paired with one of the pre-specified institutional improvements. These outcomes allow us to estimate the baseline attractiveness of Sardinia relative to alternative locations and how this changes when institutional and service conditions improve.

Beliefs about own potential salary in Sardinia
Respondents’ expected net monthly salary if they were employed full-time in Sardinia.

Beliefs about typical salaries in Sardinia
Respondents’ beliefs about the net monthly salary of a typical full-time employee in Sardinia.

Confidence in salary beliefs
Self-reported confidence in the salary expectations provided in the belief questions.

Preference for living in Sardinia
Self-reported preference for living in Sardinia measured on a numerical scale.

Heterogeneity in willingness to pay
Differences in the main outcomes across key individual characteristics, including gender, age, current place of residence, previous experience living in Sardinia, current income, and employment characteristics.

Secondary outcomes capture respondents’ beliefs about economic conditions in Sardinia and allow the analysis of heterogeneity in relocation preferences.

First, the survey measures beliefs about potential earnings in Sardinia, asking respondents to report the net monthly salary they expect to earn if they were employed full-time in Sardinia. This variable captures subjective expectations about individual labor market returns in the region.

Second, respondents report their beliefs about the net monthly salary of a typical full-time worker in Sardinia, providing a measure of perceived wage levels in the Sardinian labor market. For these belief questions, respondents also report their confidence in their answers, which provides additional information on the strength of their expectations.

The survey also includes a measure of general preference for living in Sardinia, which captures how attractive respondents perceive the region as a place to live.

Finally, the survey collects detailed demographic and labor market information that allows us to study heterogeneity in willingness to accept jobs in Sardinia. These characteristics include gender, age, current place of residence, previous experience living in Sardinia, current salary, employment status, and other socioeconomic characteristics. These variables will be used to analyze how relocation preferences and willingness-to-pay measures vary across different groups of respondents.

The study is implemented as an online survey experiment based on repeated hypothetical job-choice scenarios. Participants are asked to choose between two job offers that differ in location and net monthly salary. One option corresponds to the respondent’s current or baseline location, while the other corresponds to an alternative location drawn from a pre-defined set conditional on the respondent’s current place of residence.

For respondents living outside Sardinia, the alternative job is located in Sardinia. For respondents living in Sardinia, the alternative job is located outside Sardinia. Across a sequence of binary choices, the salary associated with the alternative location is adjusted in order to identify the salary differential required for the respondent to prefer one location over the other.

All respondents first complete a baseline set of job-choice scenarios. After the baseline module, respondents are randomly assigned at the individual level to one additional module. In most cases, respondents are assigned to a policy condition in which the Sardinian option is paired with one of three pre-specified institutional improvements: better air connections with more frequent flights and lower fares to and from Sardinia; shorter waiting times for specialist visits and diagnostic exams; or a program facilitating the job relocation of a partner when moving to Sardinia. In a smaller share of cases, respondents are assigned to a placebo condition in which they complete an additional sequence of scenarios without institutional improvements.

The design allows us to estimate the compensating wage differential associated with choosing Sardinia relative to an alternative location under baseline conditions and to assess how this differential changes when institutional and service conditions improve.

Not available

Randomization is implemented automatically within the online survey platform (Qualtrics) using the platform’s built-in randomization tools combined with embedded JavaScript. Respondents are randomly assigned at the individual level to the experimental conditions by the survey software at the moment they reach the relevant block in the survey flow.

After completing the baseline elicitation module, the survey platform randomly assigns respondents to one additional module. In approximately 80 percent of cases respondents are assigned to a policy scenario, while in approximately 20 percent of cases they are assigned to a placebo scenario. Within the policy condition, the specific institutional improvement attached to the Sardinian option is randomly drawn from a predefined set of policy interventions.

The unit of randomization is the individual respondent.

Randomization occurs at the respondent level within the online survey platform. After completing the baseline elicitation module, each respondent is individually assigned by the survey software to one additional experimental module. In approximately 80 percent of cases respondents are assigned to the policy condition, while in approximately 20 percent of cases they are assigned to the placebo condition.

Within the policy condition, the specific institutional improvement attached to the Sardinian job offer is also randomly assigned at the individual level. No clustering or group-level randomization is used in the experiment.

No

3.000

3.000

All respondents complete the baseline elicitation module. After the baseline module, respondents are randomly assigned at the individual level to one additional module.

Approximately 80 percent of respondents will be assigned to the policy condition, in which the Sardinian job option is paired with a randomly assigned institutional improvement.

Approximately 20 percent of respondents will be assigned to the placebo condition, in which respondents complete an additional elicitation sequence without institutional improvements.

Randomization occurs at the individual level, and the design is not clustered.