Tuesday, May 31, 2016

More on Noah Smith’s Low, Low, Standards

To review, Noah Smith asserted that poor countries “incredible progress” in the last 30 years. Challenged, he erroneously tried to make examples of Mexico and Brazil– both of which have suffered pretty substantial growth failures in the period. (He presented data which was not adjusted for inflation which made gains look much, much larger than an honest defense would permit.) He then pointed to the faster growth of 2000s. Indeed, the 2000s were in the last 30 years. But if he meant ten years then why say 30? This makes no sense, especially as growth again slowed over the last five years.

As a consequence of this spat, Smith has tried to peg me as a poverty-decline-denier, writing
[O]n Twitter, David Rosnick strongly challenged the very existence of a rapid recent drop in poverty. At first he declared that the poor-country boom was purely a China phenomenon.
Yes, I first argued that the “poor-country” boom of the last 30 years was largely China. All evidence points to just this. I am hardly the only person to point out that China can bias your thinking about worldwide trends. Branko Milanović wrote of the “ambivalent role of China in global income distribution” several months back.

But I never argued that, say, \$1/day poverty rates have not actually fallen rapidly. Smith knows this very well and is simply trolling (successfully, to judge by my reaction.) The only difference between us is that I am far less convinced that the rapid fall is as “incredible” as Smith insists.

How could I possibly think that Smith is hyperbolic when he writes things like
This is incredible— nothing short of a miracle. Nothing like this has ever happened before in recorded history.
In truth, this is a ridiculous statement. Of course it has never happened before. So long as poor countries grow and that poor people in those countries share in that growth, then it would be inevitable that fewer and fewer people would live below any threshold of absolute poverty. Believe me, I am thrilled that the percentage of people living in such poverty has come down rapidly. But would I call it a miracle? Please.

Are we supposed to be impressed that someone with income of \$0.99 per day one year now has an income of \$1.01 per day? Starvation is not determined by a bright line. Be happy that such poor people are starving a little less than before, but let us not pretend that their situations have so very much improved.

But poverty rates are falling more rapidly these days, right? Yes, the numbers do bear this out. But it is a mistake simply to infer that they are falling more rapidly because of incredible progress among poor countries. If we found incredible shared growth in poor countries (such as China) then the economy will rapidly pull large numbers of people out of poverty. But there is another important factor: if an incredible number of people live just below the poverty line, then little growth will also pull large numbers of people out of poverty. Such progress in poverty reduction would be best described as “credible.”

That is, even if China had not grown so rapidly, the fact that so large a proportion of people there lived near the poverty line meant that large reductions in poverty rates were all but a matter of time. The fact that China grew so fast for so long meant that large numbers quickly approached and then passed the threshold. So which is it? Faster growth or more people happening to live just short of the poverty line? Consider two countries with exactly the same (high) poverty rate, but one country (dark blue) happens to have more of its population near the poverty line as in Figure 1.

Figure 1: Distribution Matters for Poverty Reduction

Holding constant each country’s level of inequality (measured by the Gini), if both countries grow at the same rate, then the more equal country will pull about 75 percent more people across the poverty line. Put another way, if the more equal country grows at 2 percent, the less-equal country must grow at 3.5 percent in order to keep its poverty rate from rising above that of its neighbor. The important point here is that it takes a lot of growth to make up for not having a population already near the poverty line.

Given any reasonable distribution of world income, an unprecedented share of people would approach the poverty line. And in fact this is exactly what happened in the late 1980s.

(source: Our World In Data)

Of course, fast-growing China pulled its poor across the line much faster than other countries. No matter what Smith thinks about India’s rate of growth, there is a vast chasm between China’s reduction in its poverty rate and what India managed in the same period.

(source: Our World In Data)

So the “miracle” of accelerated poverty reduction is the result of a combination of fast-growing China and the non-China world’s modal income happening to lie close to the poverty line. If Noah Smith wants to sell it as a miracle, I guess that is fine so long as he understands it took China to turn the coincidence into a miracle. Coincidence might be the kind of stuff he considers heady, but not me.

Monday, May 30, 2016

Noah Smith has very low standards

Noah Smith made a ridiculous claim and insists that he is correct. But he does not stop there. For some reason, he chooses to blow it all up by exaggerating the differences in our positions.

To review, Smith asserted “incredible progress in poor countries” over last 30 years. Doubtless, there has been progress over the last three decades. After all, countries do generally grow. But incredible? Progress was generally more slow in comparison to, say, the 1960s and 70s. To be sure, some countries have done well. India– a Smith favorite— grew at an average annualized rate of 4.6 percent per capita between 1985 and 2015. And sure, there are a lot of poor people in India. But I do not consider this “incredibly” fast growth in comparison to 1960-80. The Penn World Tables show countries growing at least 4.6 percent per-capita per year between 1960 and 1980: Spain, Malawi, Tunisia, Portugal, Malaysia, Brazil, Greece, Thailand, Cyprus, Korea, Hong Kong, Japan, Gabon, Taiwan, Singapore, Malta, Botswana, and Romania.

Maybe every single one of these countries simply caught up following a period of poor growth yet no such claim may be made regarding India. Maybe I just have higher standards for incredulity. I find Botswana growing 8 percent per year a bit more impressive.

Smith did look at papers I suggested— which indicates that since 1980 there has slowdown in many indicators of progress across all income levels. But then he moves the goalposts, writing about per-capita GDP growth
By [CEPR’s] measure, the 2000-2010 decade exceeds or ties the supposed golden age of the 60s and 70s, for all but the top income quintile.
[T]he graph clearly shows that Rosnick is wrong, and the recent unprecedented progress of global poor countries is not just a China story. Case closed.
No doubt, growth has resumed for a period. And had Smith originally asserted that poor countries enjoyed a decade of growth close to that of middle-income countries in the 1960s and 1970s, I could hardly argue. But again, we are looking for “incredible” growth over three decades— not one. Twenty years of growth one percentage point below is not made-up for by a decade half a percentage point of growth above. Hardly “unprecedented” is growth in recent decades.

I argued that truly “incredible” growth over the last few decades comes from China, rather than “poor countries” in general. Yes, China has a large share of the world’s poor, but it is still not kosher to attribute to “poor countries” what may be specific to one country. Haiti’s poor can take little comfort in China’s success. Still, between 1985 and 2015, China sustained 8.6 percent per-capita growth and now produce 12 times per person than 30 years ago. Assuming the poor in China benefitted anywhere near as well as per-capita output might suggest, we are talking about tremendous gains to very large numbers of people with who had precious little. I will return to the question of poverty later.

First, however, I would like to address Smith’s complaint that CEPR’s several reports are deficient because they averaged countries rather than people. We use countries because that is the unit of observation, and we do not want to say countries generally improved performance if Chinese policymakers did something right while the IMF pushed terrible policies on the most vulnerable. Even Smith used “countries” and not “people.” Perhaps this was just a miscommunication on his part. Twitter is not always the clearest of communication channels.

But I am willing to humor Smith on this point. What happens if we weight real per-capita GDP growth by population and create a worldwide index?

Figure 1: World Growth in per-capita GDP
Sources: Penn World Tables (1950-2011), IMF World Economic Outlook (2011-) including IMF projections

Indeed! Population-weighted growth has accelerated rather than declined. However, my argument is that this is driven by growth in China, which obviously does weight strongly. Has population-weighted growth outside China accelerated?

Figure 2: World (less China) Growth in per-capita GDP
Sources: Penn World Tables (1950-2011), IMF World Economic Outlook (2011-) including IMF projections

Oops. No evidence for “incredible” growth here.

Figure 3: Without China, A Shortfall in World Growth Relative to Trend
Sources: Penn World Tables (1950-2011), IMF World Economic Outlook (2011-) including IMF projections

Yes, we see some catch-up in the 2010s– catch-up which has not as of yet continued past the period examined in CEPR’s paper on the subject. Of course, this mixes higher and lower-income countries. So maybe this is a golden age of poor-country development. Or not. But it is clear that population weights matter because of China rather than because of poor people. Which is nothing new.

It seems to me that not only is Smith willing to stand by data which in no way supports his case but he is easily impressed as well. More on that in the next post.

Noah Smith Shuts His Ears and Cries “La La La”

Recently, Noah Smith made reference to “the incredible progress in poor countries in the last 30 years.” Clearly, Noah has a low bar for incredulity. Based on a variety of measures, countries generally progressed more slowly in recent decades than they had previously. (pdf) (pdf) I challenged Noah on this point, suggesting that he was unduly influenced by China‘s truly extraordinary growth. For example, average annual per-capita GDP growth in Latin America over 2000-10 was little more than half that of 1960-80– clearly better than the 1980’s, which saw growth one-tenth that of the previous two decades.
Nor did growth accelerate in the last five years— averaging 1.1 percent. Progress, to be sure, but the last 30 years saw 1.2 percent annual average growth in per-capita GDP. Smith’s reference to 30 years of incredible growth certainly does not apply there.

So it was extra provocative when Smith cited Mexico and Brazil as examples of places where “living standards have improved a lot.” Mexico grew 38 percent per-capita in the 30 years between 1985 and 2015— a whopping 1.1 percent per year. Brazil grew even more slowly at 0.8 percent per year. When called on it, Smith wrote Doubled! Now, the data here shows 85 percent growth from 1999-2013, which seemingly would have resulted in a 16-year doubling from 1999 to 2015 if the trend in this data continued.

How can we reconcile Mexico‘s slow growth 1985-2015 and much more rapid growth between 1999-2013? Does Smith believe Mexico collapsed by a quarter between 1985 and 1999? Actual growth was 6.6 percent— slow, but not that bad. The answer was pretty obvious to those of us who have looked at the data in any depth. Smith’s numbers are in current dollars. Adjusting for inflation, Mexico grew 28 percent from 2000-15. That is a far, far cry from Smith’s 100 percent.

Worse, Smith cited Mexico in support of progress over 30 years— not 15. Over the previous 15, Mexico grew all of 8 percent— a mere half of one percent per year. Brazil grew even more slowly, averaging only 0.8 percent per year from 1985-2015. Smith simply does not know what he is talking about yet insists he did not err. What is it called when one reiterates a wrong position once confronted with the facts? Derp, perhaps?