In the video below (link provided) Hans Rosling puts on an amazing show discussing — of all topics! — the statistics of global economic growth. He does it in four minutes. Below is a link to this video. Watch the video. Be entertained and amazed.
I love this topic as much as the next economist, but I would never have expected it to become such an impressive piece of entertainment. It is only polite to bring up this topic in a dinner table conversation among relatives if they need a sleeping aid. <warning: sarcasm alert> The statistics of global economic growth is a certifiable snoozer. Try the topic on the nieces and nephews sometime. It will motivate them to write their college research thesis on something other than economics. <end of alert>
Look, I am just saying. Not everybody loves the central statistical topic of each profession.
More to the point, this topic would not seem to lend itself to showmanship. Yet, Rosling makes it entertaining. Neh, he does it while discussing — of all things! — the statistics for two hundred years of economic growth for two hundred countries. In less than four minutes. It has to be seen to be believed.
How does Rosling manage to be so entertaining with statistics? Hey, he is a good speaker, so give the guy some credit. However, star power still needs substance to stick. Rosling does something subtle to his substance. To pull off this presentation Rosling uses a number of statistical tricks. Rosling does not hide these tricks, and he is upfront about a few of them, but they are also easy to miss. That is the point of this post: to point out some of the consequences from these tricks, especially the ones that massage the main message.
A great video
Here is the post:
That was great, wasn’t it? Quite impressive for a such a short presentation. Rosling and his producers packed a tremendous amount of information into four minutes.
And, hey, there is a large bit of truth in what Rosling says. Things are markedly better for the human conditions across the globe in comparison to two hundred years ago, when life was — quoting the famous Hobbe’s phrase — nasty, brutish, and short. Indeed, Rosling’s talk was positively rousing, perhaps even inspiring. <warning: euphoria alert> That presentation was so inspiring it makes me want to get up and shout a cheer for the recent accomplishments of the human species. It even makes me stop worrying as a parent and feel better about the future prospects for my children. <end of alert>
*Sigh*. Ah, I hate to burst a bubble, but I cannot help but see behind the presentation and into the substance. Rosling’s rousing presentation became overly rosy.
Look, it is one thing to advertise how good things have gotten in the last two hundred years, but the story is somewhat different if most of the progress happened only in the last fifty. The last fifty have been quite amazing, but it is also rather fragile. And not that great for many people.
Here is what I mean.
First off, let me grossly over-simplify a bit. The movement in the first half of the graph — the first one hundred years — is largely about the experience of a few countries. Most of these are European, with honorable mentions to a few places with unexploited resource endowments, such as North America and Australia/NZ. A few other places outside Europe really do not begin to get close until the second half of the graph.
Rosling does admit as much. Britain is first out of the gate due to the industrial revolution. Then there is little progress in addressing global inequality in the next hundred years. Those hundred years are largely about a few countries breaking out of their misery. Indeed, the US and Canada start to make their ascent to wealth sooner than other places, to be sure, but do not let that distract from the bigger point. Until one hundred years ago only a few European countries — and a couple exceptions, such as the US and Canada — had figured out how to (minimally) escape Malthusian misery.
In short, and this is my second observation, the story does not get very interesting until a bit more than one generation ago — say, fifty year ago. And even then, there was still substantial uncertainty about whether the path to wealth discovered by some European countries could spread to any other part of the globe other than North America and Australia/NZ. Almost no other country had gotten as rich as the European countries, except those with oil reserves.
<warning: digression with a point> And, in case you do not get the point, exploiting other unexploited resource endowments is great while it lasts, but not a path to sustainable growth. Pennsylvania once had oil, for example, and it did not last. Other things had to happen to Pennsylvania to make the growth stick. Let that be a warning to all countries who live high on the hog today on oil reserves. <end of digression>
Third, and here is where it gets interesting, something wonderful started to happen after World War II. Forty years ago Taiwan, Singapore, Japan, Hong Kong, and Korea appeared to have an opportunity to reach some level of wealth on par with the European countries.
To be sure, nobody was quite sure. Initially the gains were spotty, and each case seemed to be the exception in one respect or another. It was not clear there was a pattern, and even if there was one, if it could be repeated in new locations. Latin American growth had made some progress and then stagnated. Many countries had no positive prospects at all, especially in much of Africa. Still, the emergence of the Asian tigers, as they sometimes are called, was something positive.
That leads to my next point. Back up this graph thirty years from the present. Then it would not have looked nearly as good. Both India and China (the two biggest circles) were considered economic laggards only three decades ago. Hong Kong was becoming a rich city, and so was Mumbia (it was called Bombay in the English speaking world back then…), but each of those cities was also a special case. It was not possible to identify positive general patterns across either China or India.
In short, China’s growth only started to take off in the 80s, and India’s in the 90s. The last twenty years helps the euphoric story because the biggest circles are not at the bottom of the graph any longer.
It makes me wonder what Hobbes would have said about the tigers.
Rosling uses a few other tricks. First, he visually cheats a bit with his statistics. He uses the log scale on the x axis, namely, on per capita income. I know this is a statistically valid thing to do, but I also think it is visually deceptive.
Here is why. This gives the appearance that the very rich are close to the very poor. The human eye naturally projects linear scales on graphical distances. Yet, in a log scale graph, the same distance at the right means a lot less than the same distance at the left. As far as I know, only professional mathematicians, economists, and engineers make the right adjustment in their head quickly enough. Log scale does not come naturally to the other 85% of the educated human population.
In short, a video aimed at the mass audience — which is what this is, after all — would have been more honest to use a linear scale. Ah, but that would not have looked as nice. After all, most of the circles are on the left, and in a linear scale, they would not move very much to the right in the animation if the x axis was a linear scale (where the middle represented 20,000 of income, not 4,000, as it presently does.) Also, most of the circles would end up further from the left than they appear in the present graph — to the left of the middle.
Professional economists might actually disagree with what I just said. After all, an additional thousand dollars of yearly income matters a great deal more to someone who previously made a thousand dollars of year income compared to someone who made, say, twenty thousand dollars of income. (Yes, I understand that point, and have no issue with the considerable research, must of it recent, arguing for using log of income to compare welfare across countries.) That observation should be paramount in a professional context.
While it is more entertaining to see the circles move a long distance, it is not necessarily more informative about the substance of the message for this video. Using a log scale reduces the facts that most economists take for granted — that most countries still live in poverty or not far above it, and most of the human population lives in poverty or not far above it. That sense does not come across in this graph.
Second, and more subtly, Rosling uses per capita income to move circles to the left and right, as well as up and down. He only hints at the issues of omitting the variance within a county, which he does at the end of the lecture. His illustration uses one example from China, and it hints at the problem. To be fair, that is a gesture towards coming clean, and perhaps that is the most one can expect in four minutes. <warning: criticism alert> Yet, I would not call it truth in advertising either. Indeed, for the purposes of this talk it strikes me as deceptive. <end>
Here is the big and blunt problem: using big and little circles hides the poverty. Once again, I realize that using small and big circles to represent population size is a nice visual trick, and it is statistically valid for some professional purposes. Nonetheless, it hides poverty within many countries by averaging it out. Moreover, for purposes of this video — recall: mass audience — hiding the poverty contributes to communicating too positive a message.
What the graph actually says is this: today in countries that have unlocked economic growth, there are two sorts of experiences. In one type of country, many people still are poor while some aspire to be rich, and even obtain riches. In another type of country, few people are poor, and most are reasonably well off, while some are downright rich. Many newly industrialized countries are the former (with big splits between rich and poor) and many long-time industrialized countries are the latter (with small percentages of poor).
More concretely, many parts of India and China still are VERY poor — and that poverty characterizes the economic experience of substantial parts of their populations. That poverty brings down the averages a lot. In contrast, in rich countries, only a small fraction are very poor (been to Appalachia recently?), so the average does not come down very much.
That is a big point. If the big circles did not hide the variance of poverty within a few countries with very big populations, such as China and India and Indonesia and Pakistan, then the picture would appear different. If every province of China and every state of India and Pakistan were given its own circles, then the graph would look much less positive. To include variance in the graph would give the appearance that things had not improved nearly as much in nearly as many places. It would be impossible to walk away from the presentation with quite as euphoric and optimistic assessment of the recent experience.
Visually speaking, a more honest graph would have made more effort to organize the data so each of the dots represented roughly the same number of people, especially for the largest countries with the largest variance, such as China and India. Use of large circles for a few countries — then projecting an average number into a graph in log scale — obscures too much.
No parting thoughts
Ah, I have said enough. There you have my two cents. What did you think?