SOAS Research Online

Abstract

We argue that gamma discounting (Weitzman, 2001) can be understood as a veridical approach to combining experts' forecasts, in which experts are treated as either right or wrong and are weighted equally in pursuit of the 'true' forecast. More appropriate is the optimal-seeking approach in which experts are wrong to varying degrees and forecasts are combined to minimise forecast errors. With uncertainty in the true mean rate characterised in this way the decline in the Certainty Equivalent Discount Rate (CER) is minimal compared to gamma discounting. Using the asymptotic distribution leads to virtually no decline in the CER, even in the deep future. The same is true when small samples of experts are used or when experts are influenced by schools of thought and hence are no longer independent. Risk adjustment is further shown to have minimal impact on our results, with obvious implications for policy