Robert N. Stavins and Lawrence H. Goulder explain the importance and utility of the concept of discounting to evaluate the net present (or future) values (NPV) of social and environmental policies in an excellent article in Nature. Discounting helps to convert the costs and benefits from various periods into their (quantitative) equivalents at a given time, and helps policy makers choose from among a set of alternatives.
They argue that we can go ahead with the investments, if there is a potential Pareto improvement - the winners from the given policy could compensate the losers and still be better off. The problem with this approach is two-fold
1. What should be the discount rate?
2. How do we identify and quantify the present and (especially) future benefits and costs of these interventions?
But as Prof Stavins writes, "The uncertainties are substantial and unavoidable, but they do not invalidate the use of discounting (or benefit-cost analysis). They do oblige analysts, however, to assess and acknowledge those uncertainties in their policy assessments."