Trillions gained and lost. Estimating the magnitude of growth episodes
Working Paper 26
Lant Pritchett, Kunal Sen, Sabyasachi Kar and Selim Raihan
We propose and implement a new technique for measuring the total magnitude of a growth episode: the change in output per capita resulting from one structural break in the trend growth of output (acceleration or deceleration) to the next. The magnitude of the gain or loss from a growth episode combines (a) the difference between the post-break growth rate versus a counter-factual “no break” growth rate and (b) the duration of the episode to estimate the difference in output per capita at the end of an episode relative to what it would have been in the “no break” scenario. We use three “counter-factual” growth rates that allow for differing degrees of regression to global average growth: “no change” (zero regression to the mean), “world episode average” (full regression to the mean) and “unconditional predicted growth” (which uses a regression for each growth episode to predict future growth based only on past growth and episode initial level). We can also calculate the net present value at the start of an episode of the gain or loss in output comparing the actual evolution of output per capita versus a counter-factual. This method allows us to place dollar figures on growth episodes. The top 20 growth accelerations have Net Present Value (NPV) magnitude of 30 trillion dollars – twice US GDP. Conversely, the collapse in output in Iran between 1976 and 1988 produced an NPV loss of $143,000 per person. The top 20 growth decelerations account for 35 trillion less in NPV of output. Paraphrasing Lucas, once one begins to think about what determines growth events that cause the appearance or disappearance of output value equal to the total US economy, it is hard to think about anything else.