While lots of ink and bits and bytes were spilled yesterday regarding the deficit projections in the President’s new budget, it’s very important to consider the uncertainty around such estimates, especially when you’re talking ten years out in the future. A bump in growth, interest rates, inflation, productivity, health costs, not to mention policy changes that will occur between now and then can all have major impacts.
Below you see that there’s a 90% chance that by 2025 the deficit will range from of a surplus of 2.8% percent of GDP and a deficit of -7.7%, with central forecast of -2.5%.
The lines in the figure are drawn under the assumption that future forecast errors are normally distributed based on past forecast errors. But were we to get a little Bayesian, we might ask: based on some underlying trends in some of the key movers noted above, can we make an educated guess about which end of the range might be more likely?
I think so, but running out now, so more to come!
I wish all the graphs showed the uncertainty. Being a scientist (well, at least I was one for a decade or so) I know that all real numbers are noisy and most have non-linear feedback so when I see these projections that just show a single line extrapolated from a handful of points I know someone is lying with statistics again. Too bad we don’t live in a rational society…