This talk is run by The Institute for New Economic Thinking at the Oxford Martin School
Speakers: Francois Lafond & Doyne Farmer
Recently it has become clear that many technologies follow a generalized version of Moore's law, i.e. costs tend to drop exponentially, at different rates that depend on the technology. Here we formulate Moore's law as a time series model and apply it to historical data on 53 technologies. Under the simple assumption of a correlated geometric random walk we derive a closed form expression approximating the distribution of forecasting errors as a function of time. Based on hind-casting experiments we show that it is possible to collapse the forecasting errors for many different technologies at many time horizons onto the same universal distribution, which fits the data remarkably well. Our empirical results indicate that the rate of improvement of a technology is strongly positively correlated to the noisiness of the improvement process. As a practical demonstration we make distributional forecasts at different time horizons for solar photovoltaic modules, and show how our method can be used to estimate the probability that a given technology will outperform another technology at a given point in the future.
Tea & cake available from 3pm, talk starts at 3.15pm for 30 minutes to then be followed by discussion.