Stated preference methods and empirical analyses of equity in health economics
Abstract: This dissertation considers two different aspects of health economics; (i) stated preference methods and (ii) empirical analyses of equity. The first essay deals with the issue of the choice of econometric models when analysing data from a closed-ended contingent valuation survey in health economics. The features of provision, either compulsory or voluntary, and the fact that certain individuals may not participate in the programme ought to affect the choice of econometric specification are argued to be important in the choice of econometric models. This paper suggests that, in general, a two-part model should be applied because it considers the most common case, namely that of voluntary participation where certain individuals do not participate in the offered programme and this is also indicated in a Monte Carlo study. The second essay compares two different methods of identifying non-participants which are performed either before or after the valuation question. When using a two-part model, the results indicate a starting point bias in the participation model when non-participants are identified after the valuation question in a hand eczema primary prevention programme. The third essay analyses an open-ended contingent valuation survey of individuals' willingness to pay for reductions in the local level of air pollution in Sweden using a multilevel approach in order to allow for hierarchically structured data. The results indicate that the majority of variations are between individuals. However, our results also indicate that there are variations at higher levels and this may be explained by homogenous preferences for a reduction in air pollution among individuals living in the same household and/or region with a similar level of air pollution. In the fourth essay a choice experiment is employed to examine the feasibility of using this technique to elicit individuals' preferences for efficiency-equity trade-offs in the health care sector. In this experiment each respondent made pair-wise choices between two identical asthma prevention programmes. However, the programmes differed in efficiency as each programme included different numbers of recipients. Equity was defined by concern for whom health was improved and this was operationalised by varying the socio-economic characteristics of the recipients. Our findings indicate that respondents are willing to forgo efficiency if individuals with a low general health status or low after-tax income benefited from the programme. Internal tests of validity indicate stable and transitive preferences. The primary purpose of the fifth essay is the presentation of a method which encompasses a wide range of normative judgements when measuring inequity in the health care sector by employing concentration curves. This approach is operationalised by ethical weights given by a beta distribution. In an empirical application, we analyse income-related inequity concerning whether an individual has visited his or her GP in Great Britain in the fortnight before the survey during the period 1984-1995. The results indicate that while changes of inequity over time are not statistically significant, inequity favouring the poor is statistically significant for half of the years analysed. The degree of inequity is, to some extent, sensitive to the normative judgement imposed, especially if greater ethical weights are put on poorer recipients.
This dissertation MIGHT be available in PDF-format. Check this page to see if it is available for download.