Sunday, May 21, 2017

Weekend reading links

1. Even as rest of the world is slowly converting farmland for other purposes, Singapore is going in the opposite direction and trying to become a leader in agriculture technologies,
Singapore... is making land available for agriculture for the first time in decades for farmers to come up with high-tech ways of improving the city-state’s food security. About 60 hectares of land will be released for farming from August, in a move authorities describe as providing a “buffer” against disruption in the supply chain... The south-east Asian city-state imports more than 90 per cent of the food needed to cater for its 5.6m inhabitants, and outbreaks of disease among overseas suppliers are quickly felt in price rises at markets... Farmers bidding for the new land will be encouraged to outline their plans to raise productivity as well as competing on price. Land for agriculture is scarce in Singapore and it can be hard to find skilled agricultural labour but farmers benefit from government funding for research and development... Eugene Tan, a political analyst at Singapore Management University, likened the farming plan to Singapore’s efforts to achieve self-reliance in water through rainwater catchment, recycling and desalination. Water technology developed in Singapore has been exported globally... Singapore’s government recently supported a vast farming venture in north-east China, which is intended to provide a reliable supply of pork and other staples. The Jilin farming project, managed by a public-private partnership, covers 1,450 sq km — an area twice the size of Singapore.
2. Story of Amazon stock since listing in 1996,
Amazon shares, listed at $18 at the time, have produced a compound annual return of nearly 40 per cent since the initial public offering. That means, for example, that $100 invested in Amazon stock at the IPO would be nearly $64,000 today (the stock split three times, so one share at the IPO equals 12 shares now). Up 63,990 per cent since its IPO, Amazon ranks as the best-performing US-listed IPO since 1995, according to Dealogic. For context, Netflix is fourth and Yahoo is seventh and the total return on the S&P 500 in the same period is about 300 per cent. It was not always smooth sailing, though. Its shares sold off after the initial euphoria around the listing. Then, during the dotcom bust, adjusted for the stock splits, Amazon fell from a high of $113 in December of 1999 to a low of $5.51 in October of 2001.

3. From the FT, on rising manufacturing and general wages in China,
Average wages in China’s manufacturing sector have soared above those in countries such as Brazil and Mexico and are fast catching up with Greece and Portugal after a decade of breakneck growth that has seen Chinese pay packets treble. Across China’s labour force as a whole, hourly incomes now exceed those in every major Latin American state apart from Chile, and are at around 70 per cent of the level in weaker eurozone countries, according to data from Euromonitor International, a research group.
But for countries like India, even with stagnation in manfacturing wages since 2007 at $ 0.70 per hour, this does not automatically mean good news. In fact, other factors are likely to more than offset China's erosion of competitiveness due to rising wages. The rise in Chinese wages has been accompanied by an even faster rise in productivity. Further, the massive size of Chinese market, at around 20% for a range of sectors, makes local manufacturing even more commercially attractive. These are unlikely to reverse soon.

4. Ananth points to Steve Rattner's twitter graphic on corporate tax rates and revenues as percentage of GDP.
Apart from the US problem, there is one another interesting thing that stares out of this graph. Contrary to the conventional wisdom on lower tax rates leading to higher aggregate revenues, as reflected in the Laffer curve argument, we see a more or less linear relationship. If there is a plateauing of the revenues, it appears to happen at very high rates.

5. Fascinating profile of Emmanuel Macron, the new occupant of the Elysee Palace, by Simon Kuper,
His long, lone walk to the Louvre stage on the night he became president expresses a truth: Macron walks alone. Almost uniquely in modern western democratic history, he has won power without joining a party. He doesn’t seek enduring allies. Instead he seduces useful people, then drops and humiliates them... Le Pen derided Macron as a creature of “the system”. That was wrong: Macron is nobody’s creature, because he isn’t loyal to anyone. This makes him a born transgressor like Blair, keen to trample on sacred beliefs.
The good thing with not having an ideology for a politician is that it helps him or her view problems in the most prudent manner, especially useful when navigating complex public policy challenges. The bad thing is that in the absence of an ideology, without its moral underpinnings (either equity or efficiency, social or individual), the politician becomes an opportunist. Even in this world of convergent ideologies, it is rare to find a combination that marries the two parts in an acceptable proportion.

6. The always brilliant Lant Pritchett takes down RCTs by drawing attention to its X-axis (determinant variable) focus and agnosticism about a theory of the Y-axis (the variable of interest),
Rather than asking “what would help understand the observed (or achievable) variation in Y in ways that allow maximal benefit?” the research asks “who will let me do X or evaluate the impact of their doing X?”... But whether conditional cash transfers induce more attendance at school can be individuated, and sample sizes of individuals can generate statistical power, so we have dozens and dozens of RCT studies of CCTs. This is the case even though we know and knew, for sure, both before and after the CCT research that CCTs were not a very big determinant of even school attendance (across countries, over time, or across individuals), much less learning, much less overall development...

A casually identified estimate of the impact of X on Y does not answer the question of how big the impact is relative to the magnitude to be explained... While one might argue that having a complete model of Y is unnecessary to knowing whether X is a cost-effective intervention, without some sense of “how big” it is impossible to know whether there are other, potentially much bigger Xs left unexplored... That doing X had impact ΔY in a given context is, in the absence of theory, zero rigorous evidence about the likely impact of doing X in any other context. For that matter, it is only theory that can tell us what “context” even means... What a researcher can do in a “field experiment” or an NGO might be willing and able to do with an impact evaluation built in might have nothing to do with what is scalable, particularly scalable by the public sector.
7. I am inclined to argue that India's banking sector non-performing assets crisis has been caused by a combination of three factors - bad business decisions (by both corporates and banks), adverse contextual events and trends, and plain cronyism and corruption. And if and when someone documents the ongoing banking sector crisis in India, I shall not at all be surprised if the third factor played an important role in the build up of such bad assets.

Livemint has a nice graphic that points out that there were nearly 6800 wilful defaulters (or someone who has the ability to pay) who owed nearly Rs 1 trillion or a sixth of the NPAs.

8. Upending the caricature of farmers and agriculture being the problem in bank defaults, the graphic below shows that the problems is far more with corporate loans than with farm loans.
The picture does not change with stressed loans, where too corporate loans predominate.
9. An astonishing set of stats about men's tennis,
Since Roger Federer won his first Wimbledon in 2003, the big five have won every Grand Slam but five—Federer (18), Rafael Nadal (14), Novak Djokovic (12), and Andy Murray and Stan Wawrinka (three each). And the remaining five were won by five different men. That’s 50 of 55 titles taken by just five players—and, remember, Murray’s and Wawrinka’s titles only go back the last five years... Only one of the five one-time Grand Slam winners has come close to another major title... Only 24 men have reached the finals of the last 55 Grand Slams. Of these, 33 finals have been contests exclusively between the Big Five—all of them since the French Open in 2006; that’s 33 out of the last 44. Only 10 men have won these titles and the Big Five players are the only ones to have won more than one.
It is even more staggering if you leave out Murray and Wawrinka, who are clearly at far lower level than the Big Three. Incidentally, the Big Three are the only players other than Rod Laver and Andre Agassi to have completed a career Grand Slam of having won all the four majors. Should we celebrate a rare confluence of extraordinary talent or be concerned at the deficiency of competitors?

10. Finally, Laura Alfaro and Fabio Kanczuk examine the utility of fiscal rules and have this to say,
A simple debt rule that limits the maximum amount of debt is analyzed and compared to a simple deficit rule that limits the maximum amount of deficit per period. Whereas the deficit rule does not perform well, the debt rule yields welfare gains virtually equal to the optimal rule.
This converges with India's own recent shift of its fiscal accountability framework from a deficit-based one to a debt-based rule.


Karthik Dinne said...

As much as I admire and respect Prof. Pritchett, I think he got it wrong here. His criticisms against X-focused approach can be termed as originally the criticisms against Y-focused approach, that led (partially?) to the emergence of X-focused approach.

For instance, consider the isomorphic mimicry that Lant talks about. It's a byproduct of "Y-focused" approaches that frame questions as "what's the cause for Y in Z context?". Framing questions in such manner seeks to infer some "X" that led to "Y", so that it can be applied elsewhere. It ends up in isomorphic mimicry. X-focused approach reverses this to ask - does X cause Y in your context?

Instead of posing it as Y vs. X, it should be instead framed as "How can we integrate Y with X (lessons from Y guiding X focused studies), and X with Y (learnings from X focused approaches built into theory of Y)?"

Gulzar Natarajan said...

Thanks Karthik. Some thoughts that come to mind immdly.

I don't see the point you are making against what Lant is saying. He is using the X-Y frame to make five or six specific critiques of RCTs. I agree with all of them and I think even the randomistas acknowledge them. Maybe some of them were just as well applicable to Y-centric approaches. That does not in any way take away from the critiques of X-centric approaches.

And your last para is in some ways PDIA - start with the problem, its theories, and let the solution evolve iteratively...

Karthik Dinne said...

Oh, Okay! In detail... I have 4 specific concerns regarding Lant's arguments and 3 concerns on the alternative he proposes elsewhere. In brief, my point is that Lant's arguments in this article are incomplete and in some cases incorrect too. The alternative he proposes elsewhere is impractical, undesirable and infeasible.

On the arguments:

1. Almost all the criticisms cited can be easily applied to Y-centric approaches as well. The point being, the article portrays as if these issues are specific to x- centric approaches alone, arising from focusing on x, which otherwise aren't present in Y-centric approaches.

2. It is undeniable that there are aberrations in the RCT exuberance. One can easily quote examples for each category of criticism presented above. But that line of argument fails Hirschman's test (If you say that "A" is true, also ask if "the opposite of A" is true as well). ... meaning, for examples that fit into the mentioned categories, one can also quote examples which don't fit into these categories. Hence, his argument/analysis is incomplete.

3. The last argument in the article, on the x-focused approach not bettering the understanding of Y, goes against his own reasoning, argued elsewhere. Quoting Paul Krugman, Lant says that the value of economic research is like gravel, which is useful only as a collection. If one were to pick individual stones (papers), they would seem useless individually but their collection brings value.

Similarly, the value of x-focused approach is not in one single study. It's in the collection of studies. By pointing to lack of utility of individual studies, Lant is picking up individual stones from gravel and calling them useless, dismissing the value of the collective.

4. x-centric approaches arose precisely because of limitations of y-centric approaches. It definitely drifted in different directions later but I find it unfair that its value addition, how much ever it might be, is not acknowledged.

On Lant's alternative:

Frankly, I am not convinced by Lant's alternative (leaving aside PDIA). In the ongoing Blattman-Pritchett debate on Bill Gates's initiative to give free chickens to the poor in Nigeria, Lant says that Gates should instead use this money to fund Y-focused research that feeds good macro policy ideas into Nigerian government.

This to me is problematic on several counts - impractical, undesirable and infeasible, at least in short term.

1. World Bank already does this kind of work. I don't know if it impacted Nigerian government till now or if it will impact in future, as long as the political instability and ethnic conflicts continues. Not sure if Bill Gates can be of any value add here as of now, especially considering that WB already does it.

Even if one agrees for the sake of argument that Y-centric policies percolate and are successful, it doesn't do away the need for redistribution. Given that redistribution is here to stay, it is worth knowing the best form of redistribution in this context. Chicken vs. Cash RCT (x-centric approach) does add value to this knowledge base and complements other Y-based inferences.

2. I also don't think the binding constraint in Nigeria is the lack of knowledge on what to do. They have first order political issues of democracy, rule of law and addressing ethnic conflicts. An external private individual like Bill Gates should better not interfere in these political affairs. It's undesirable.

Of course, one can argue that Gates should work on the lines of Francis Fukuyama, training political activists of Uzbekistan and other countries, but there are different people for that. It's not the only way to contribute.

3. There is confusion between short term and longterm, and private individual vs. political activist. For an external private individual, to at least address the short-term concerns, it's still worth giving free chickens, till the big political issues settle.