The question that I'm interested in is whether questions about WTP make any sense in a scenario where we are dealing with a paid service.

To give it a little bit of context: consider Uber, for example. Suppose that our empirical goal as a researcher is to conduct a survey among Uber users which should somehow reveal how much they value the presence and the ability to use Uber. Ideally, I want to elicit the average consumer surplus via such survey.

I'm fully aware that WTP can be properly elicited only in a controlled experiment, and that surveying people about their WTP might be not a very good approach. However, it's sometimes used when you need some figures, even if they are not very reliable (e.g. in a setting of public healthcare, frequently used in health economics).

Now, if we were dealing with some free-to-use service, then asking WTP questions (in a proper manner) seems proper in terms that we understand what this reported WTP means. However, in the case of Uber I have some doubts -- consumers already pay money for using such service as Uber, and if they are asked about their WTP for continuing to use Uber, it's not that clear and trivial to me what will their reported valuation consist of and what is the alternative in their mind (e.g. the baseline that they compare with the current situation might be the world without Uber but with the same taxi prices (so then WTP would represent only convenience of using the service), or the world without Uber with prices that were in place until Uber launched (thus implying presence of the whole effect of Uber on the market in WTP).

What is your take on this? At first glance, I didn't find a lot of literature on the topic. Please feel free to share your thoughts and maybe relevant literature.

Thanks a lot in advance to everyone!

  • 1
    $\begingroup$ What do you mean by willingness to pay “approach”? Willingness to pay is by definition the maximum price customer is willing to pay for a good or service. That is not approach but some parameter that has to be empirically estimated. Also surveys are typically the worst way to try to estimate willingness to pay. I know they are sometimes used in health economics because there are no other good ways of estimating value of life and things which public healthcare needs to make decision, so some numbers have to be produced even if with bad method, but it’s not good approach in general $\endgroup$
    – 1muflon1
    Oct 21, 2023 at 14:37
  • $\begingroup$ @1muflon1 yes, I understand clearly what WTP is, I just may have worded it a bit oddly. And I clearly understand that WTP can indeed be properly elicited only in an experimental setting. However, my setting is alike the one with Uber described above, and my case is exactly when I have to produce a number even though it's not the best approach to try and determine the WTP via a survey. So really my question is -- what are the potential pitfalls and aspects of thinking of WTP when dealing with paid services. $\endgroup$ Oct 21, 2023 at 15:06
  • $\begingroup$ Could you then please edit your question to clarify it for everyone? $\endgroup$
    – 1muflon1
    Oct 21, 2023 at 15:08
  • $\begingroup$ @1muflon1 I've tried, thanks $\endgroup$ Oct 21, 2023 at 15:23
  • 1
    $\begingroup$ Shouldn't the survey tell the participants the scenario they should be considering? The scenario can be varied to get a more complete picture of how WTP depends on various factors. The demand for services like Uber is a "derived demand", and so the WTP will differ because of all sorts of things (in addition to the prices and other characteristics of substitutes). Also, the quality of Uber's service is not uniform; it depends on the driver and the levels of congestion etc. $\endgroup$
    – smcc
    Oct 21, 2023 at 16:54

1 Answer 1


I do not believe it is a problem as long as you word your questionnaire correctly.

Willingness to pay can depend on available substitutes, as it is well known that the availability of substitutes changes the demand curve. So, if you want to know their willingness to pay for Uber, conditional on substitutes such as a taxi, just ask them to imagine that scenario.

If you want them to report willingness to pay in a world without any alternative to Uber, simply instruct them to imagine that scenario.

I don't think there is any problem with such an approach, other than what was already mentioned in the comments about the reliability of surveys. You can also inform readers in your report about these potential issues and provide an explanation.

To improve the results you can collect data on some background characteristics of the surveyed people (like age) and report results conditional these characteristics.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.