if you mean economic valuation, then I think the fundamental problem is that we ascribe value to the final services, because these are the secosystem ervices that directly influence human well-being. If you were to try and value supporting services, how would you do so without reference to the provisioning, regulating and cultural ecosystem services they enable?
That is not to say we should not consider the values associated with supporting services, I discuss some of these issues in this paper
Abson, David J., and Mette Termansen. "Valuing ecosystem services in terms of ecological risks and returns." Conservation Biology 25.2 (2011): 250-258.
You are essentially asking how we value inputs to production of goods and services we value. How we value ecosystem services that support ecosystem goods we directly use can be done using an Marginal revenue product approach, just as we value labor and capital. However, since the ecosystem services we directly value are non-market goods, it becomes a bit more difficult to find the final prices that are used in the MRP. using non-market methods to find the final 'price' you can then use ecological models models to find marginal product...then you have MRP. I believe there is a paper by Tom Crocker out there that does this type of thing. It was published in the 1980s I think. Hope this helps.
Yes, the question of valuing ecosystem services has intrigued me for many years. I am not sure this has been done empirically, but the theory for it is solid.
It is indeed quite a relevant question. As other colleagues have pointed out, if we are interested in the market value of services derived from an ecosystem we should then pay attention to provisioning services, as market prices for those are mainly found in the marketplace. However, that does not lead to a whole story about value (please note the difference I make between prices and values, as noted by many other colleagues in out area). The problem of double counting mainly happens within forms of exchange value, only market prices or cultural values or energetic value, and so on. Congruity and avoiding double counting then can be applied within each exchange value, not across.For instance, Fisher et al2008, address values at the marketplace (final services) and proposes a classification system to avoid double counting on that exchange value dimension. Other authors have contributed to other dimensions.