When a Forecast Shouldn't be Charted

Silly chart of the day comes from Morgan Stanley, whose Jonathan Garner thinks that emerging-market infrastructure spending is likely to grow by about 12% per year over the next decade. That's an interesting forecast, but it's basically one number. Charts, by contrast, are great at displaying whole tables full of numbers, to help you see how things change over time. What happens when you try to stretch one number like 12% to the point at which it fills a whole chart? You get something like this:

Nothing is happening in this chart. The bars simply march up in lockstep, increasing at a steady pace, with all the different regions keeping exactly the same share of infrastructure spending as they have today. It's a way of taking an educated estimate - that 12% figure - and giving it some spurious graphical authority. That authority is then bolstered by that list of sources at the bottom, which includes the World Bank. (Which I'm sure was used just to find out present-day ratios, not for its forecasts.)

In general, it's OK to put forecasts in bar charts like this one, but only as the final bar or two of a long series of actual figures. If the chart had shown infrastructure spending from say 1997 to 2010, that would be one thing. It would have lots of actual data, and then a bit of forecasting at the end. But an entire chart which is nothing but forecasts, and only one forecast at that? Junk.