Study on air travel tax

You might be thinking to yourself: "Why do we need another study on an air travel tax? Wasn't one even published just a couple of weeks ago?" It is true, there have been studies done on the question of taxing air travel before, the latest one by gfs-zürich on behalf of umverkehR, published in June. The disadvantage of these surveys, however, is that they are not able to capture the multidimensionality of a air travel tax. They do not show why someone is in favor of or opposed to such a tax. Furthermore, no insights are provided into the trade-offs an individual makes when deciding to approve/oppose the tax. As a consequence, political actors find themselves on a rather weak scientific foundation when attempting to draw up an air travel tax policy. This is where the Conjoint Analysis method comes into play.

Source image: magazin.nzz.ch

In a Conjoint Analysis, participants are repeatedly presented with two randomly generated air travel tax policies. Here, the so-called "attributes" and "levels" play a central role. Attributes are characteristics of a tax, such as "tax rate short-haul flight", "tax rate long-haul flight", "use of earnings" etc. The levels are the values that an attribute can take on. For the attribute "tax rate short-haul flight", the levels could be "CHF 5", "CHF 30", "CHF 60" or "CHF 90". Also, the levels do not have to be numeric, as the attribute "use of earnings" shows: Here, levels such as "redistribution to population", "investments in climate projects" or "investments in railway network" are thinkable.
Based on the attributes and levels, hypothetical versions of an air travel tax are generated, i.e., for each attribute a level is randomly selected. Then, the two randomly generated versions are displayed side-by-side. At this point, the participant must decide which of the two versions he/she prefers (see example in table below). This choice task is repeated multiple times, each time with two newly generated versions.
In the end, through statistical analysis, we can identify which of the attributes (and its levels) were important to the participant, and which attributes were ascribed little to no importance. For instance, it seems likey that a higher level in the attribute "tax rate short-haul flight" will decrease acceptance of a tax, all else being equal. At the same time, it might well be that a majority of the pariticipants does not care about the attribute "use of earnings", i.e., their acceptance/opposition does not change when changing from "redistribution to population" to "investment in railway network", all else being equal. All this can be identified thanks to the Conjoint Analysis method.
This is where the added value lies compared to "traditional" surveys: With the help of these findings, political actors working on an air travel tax policy in the future will understand, with regard to which attributes special caution is recommended - and for which attributes they enjoy discretion.