A Clark Award for economic history

by on April 27, 2017 at 1:45 pm in Economics, History | Permalink

I refer you to the excellent post by A Fine Theorem on David Donaldson, here is one excerpt:

Donaldson’s CV is a testament to how difficult this style of work is. He spent eight years at LSE before getting his PhD, and published only one paper in a peer reviewed journal in the 13 years following the start of his graduate work. “Railroads of the Raj” has been forthcoming at the AER for literally half a decade, despite the fact that this work is the core of what got Donaldson a junior position at MIT and a tenured position at Stanford. Is it any wonder that so few young economists want to pursue a style of research that is so challenging and so difficult to publish? Let us hope that Donaldson’s award encourages more of us to fully exploit both the incredible data we all now have access to, but also the beautiful body of theory that induces deep insights from that data.

The post is superb, yet A Fine Theorem remains underrated.

1 Ray Lopez April 27, 2017 at 2:26 pm

Just to be a cynic, perhaps Donaldson’s work was 99% perspiration and 1% inspiration, since the conclusion was so obvious. GM Rogoff in his book “This Time” complained getting long-term data on economics is difficult.

2 Nodnarb the Nasty April 27, 2017 at 3:44 pm

It’s not that hard, though. Vincent Geloso (PhD, LSE) and Mark Koyama (PhD, Oxford) are both telling in this regard, even being able to blog about their findings.

Maybe there’s something about doing economic history in the UK that brings out the best in econ historians…

3 Troll Me April 27, 2017 at 5:44 pm

Old datasets are rarely comparable, and who even knows where they are.

For comparison, check data.worldbank.org, where there are hundreds of datasets on nearly every country in the world. But they do not go back so far that you can really do “economic history” with them.

If I understand correctly, it’s difficult to figure out which of limited data is suitable to research questions (contrarily, which research questions might be answered with existing data), and then there’s the hassle of getting the data. I don’t think many sources just send the data to anyone who tweets them a request, but I doubt there are serious impediments to those who can demonstrate that they are actually doing research.

Aside from that, if there is proprietary data, it may not be affordable for the researcher.

Point is that access is theoretically available, but practically difficult. (Right?)

4 Troll Me April 27, 2017 at 5:56 pm

CAN’T DO not “can do”

5 Ray Lopez April 27, 2017 at 10:58 pm

@Troll Me – Eye, eye, eye do anything that you want me to do…but I can’t go for that, no, no, no can do…. (Hauling Oats song)

Ponder this: given a limited set of years, you cannot do a rigorous study of the data since N (sample size) is too small. But to get around this, you can ‘bootstrap’ the data by simply running lots of “Monte Carlo” type simulations, pretending that you are ‘rewriting history’. Simple example: you know that US stocks return about 7% a year (Siegal’s constant, though, I digress, John C. Bogle today says the future constant should be closer to 4%). However, instead of simply relying on 100 years of stock market data as the data actually occurred historically, you assume that the years are independent and randomly mix up and replace the years. So, if stocks when up 11% in 1950, and 12% in 1951, then down -5% in 1952, you would mix up this years and assume anything can happen within the range of these variables (-5,11,12 etc). This would change the final return and numbers, and give a more “true”, “alternative history” (but not alternative facts!) scenario. Bootstrapping. However, you are making the assumption–which is generally valid–that there’s a Markov chain process, a fancy term meaning last year does not influence this year’s data, the data are independent, and there’s no ‘path dependence’ of the data. Generally this is true. But it’s a bit of a dodge. For this reason, Knick Taleb and others say that N is always too small for any economic analysis, and that we just don’t know what will happen due to those pesky Canadian white crows or black swans.

Thanks for the chat. Or monologue.

6 Amigo April 28, 2017 at 2:21 am

Pretty sure a Markov chain process is built upon idea that the probability of what comes next depends upon what comes before. At least that’s how I understood and used it in a musical algo.

7 Ricardo April 28, 2017 at 2:38 am

Bootstrapping is a technique that allows one to construct confidence intervals around estimates without making the assumption that errors are normally distributed. It won’t solve the problem of small N, though.

8 Ray Lopez April 28, 2017 at 8:41 am

@Amigo – it’s related, but technically I think I’m right (but it’s one of those negative sign things that’s just the flip side of the same coin) Wikipedia: “Loosely speaking, a process satisfies the Markov property if one can make predictions for the future of the process based solely on its present state just as well as one could knowing the process’s full history, hence independently from such history; i.e., conditional on the present state of the system, its future and past states are independent”

@Ricardo – I agree N being a small number won’t help much in bootstrapping, but they I think you are implicitly bootstrapping when doing Monte Carlo simulation of finance data, which is what I was referring to, though I guess just using the term “Monte Carlo” is more accurate.

9 Troll Me April 28, 2017 at 10:30 am

I’m more comfortable with missing variables and the “unknown unknowns” that this integrates into the analysis, as compared to pretending to know what you don’t know and then attaching a bunch of decimal places onto your p-hacked results.

The USA and UK have good historical economic data for many purposes, and as interest in economic history returns, papers pointing to the data available in the 1920s or 1850s will circulate more, and people will know where to find what data is available or suitable for which kinds of questions.

10 Troll Me April 28, 2017 at 10:32 am

Ricardo – Monte Carlo simulations with small N datasets can be done.

In any instance, the first question should be “why the hell did you know do it any other way?”

Probably any alternatives are not obviously better when you see someone do that. I don’t find the method convincing in the least, but nor do I know better ways to tease anything out of such limited data.

11 Ricardo April 28, 2017 at 11:44 am

“Ricardo – Monte Carlo simulations with small N datasets can be done.”

Sure, my point is that small N is not the reason you bootstrap confidence intervals. You do it in cases where you want to avoid assuming that error terms are normally distributed. T-stats assume that error terms are normally distributed but bootstrapped confidence intervals only require the more general assumption of finite variance.

12 Troll Me April 28, 2017 at 7:35 pm

Gotcha. Thanks.

13 carlospln April 28, 2017 at 12:00 am

It didn’t seem to be a problem for David Hackett Fischer.

14 rayward April 27, 2017 at 2:40 pm

Historians use narrative to recount the past. Economists use data to recount the past. Never the twain shall meet. Charles Morris has written an account of the great depression. https://www.amazon.com/Rabble-Dead-Money-Depression-1929%C2%961939/dp/1610395344/ref=asap_bc? Morris has been described as “a writer of popular but rigorous works of economic history”. What Morris does in this book is provide a narrative of the causes of and responses to the great depression, including the causes he prefers and leaving out those he doesn’t. Is that economic history or a polemic? And that encapsulates the inherent conflict of economic history.

15 Ricardo April 28, 2017 at 1:37 am

There are exceptions, though. Niall Ferguson uses data and even regressions in some of his work while Christina Romer has used narrative sources to identify “natural experiments” in the historical record.

16 dearieme April 27, 2017 at 2:40 pm

If huge sums were spent on those railways and yet the Indians didn’t make good economic use of them, then India ought to pay Britain reparations. Obvs.

17 spudist April 27, 2017 at 4:31 pm

The railways when constructed, were never meant to be controlled by the Indians. As the article states, it was primarily built for military use by the British raj with the economic benefit being seen as a collateral benefit because they could tax the profits. I hardly think India is the one that must pay reparations. I think that Dr. Tharoor makes a strong case for the contrary in his book (https://www.amazon.com/Inglorious-Empire-What-British-India/dp/1849048088)

18 Believe it! April 27, 2017 at 5:11 pm

I love how a country of people who literally walk around crapping in public should tell anyone what to do

19 Thiago Ribeiro April 27, 2017 at 6:26 pm

If the greedy Englishmen had taught and civilized them as we, Brazilians, surely would have done instead of oppressing them and exploring them, they would not be living this way. But plundering and preying are the British way.

20 Mark Thorson April 27, 2017 at 10:20 pm

Brazil, the land where everything works. Until it doesn’t, of course. Like tomorrow.


21 Anon April 28, 2017 at 1:36 am

I thought History was always a case more of plundering and praying.

22 Thiago Ribeiro April 28, 2017 at 5:27 am

We are not slaves, fearing the whip of merciless corporate master, this is tge difference between Brazilians and Americans.

23 Thiago Ribeiro April 28, 2017 at 5:28 am

“I thought History was always a case more of plundering and praying.”
Only when those in charge are merciless plunders.

24 dearieme April 27, 2017 at 6:50 pm

Do you really think the comment was about India and Britain?

25 wiki April 27, 2017 at 3:42 pm

A Fine Theorem exaggerates the importance of the model to Donaldson’s insights. It is only a tour de force because of the ridiculous strictures economics places on theoretically based empirical work and what are laughingly called structural estimations.

26 polyglot April 27, 2017 at 9:56 pm

Love ‘a Fine Theorem’ but there is another side to the story- one which calls into question the value of this type of analysis in the social sciences. In the case of the Indian Railways, we have data-sets on cost of transport and also some information on why different groups had differential access to transport at the stated price. It is folly to pretend that this information is consistent with Donaldson’s work. The truth is some groups could and can utilise reliable rail transport at a certain price while other groups could not do so with any degree of certainty or without facing legal or extra-legal sanctions. Those groups who could use the railways did not generate surpluses for the hinterland. Thus Donaldson’s work, as summarised, is wrong headed. He lacked an emic perspective and his work is worthless. There was an old Indian Prince, mentioned by Art Buchwald, who was convinced that the President of the US was married to the Queen of England. He would drink the brandy he had himself distilled the previous week and hold forth on this subject with great loquacity and every appearance of adherence to the rules of Buddhist logic. Yet, his ratocinations on American affairs- e.g. his view that the Queen had got jealous of Jackie Kennedy and arranged a sniper to get rid of her- were worthless.
Sadly, the same can be said for Donaldson’s work as cited.

27 anonymous as usual April 27, 2017 at 10:02 pm

The “Settlers of Catan” seems like a fun game, I bet when guys like Donaldson play it they are the way I am when (with years of small squad military leadership training), um, when I watch a movie about, well, small squads facing military problems. To change the subject slightly, but only slightly: Something nice I saw today at the local NoVa commercial district : a smiling pregnant woman with an unattractive face (not unattractive to me, for the record, but objectively not what is called an attractive face – symmetry off, complexion a little too much on the dry but sweaty/greasy side, surprisingly small and lusterless eyes, all framed by shortish pulled-back curly (curly is a good thing unless it is pulled back like that) and dry (never a good thing) and half-gray hair – pulled back, as I said, in an uncomfortable looking way – (well that being said, she will never read this so that was not rude in any way, and besides she is without a doubt more attractive to the opposite sex than I am or ever will be again….). For some randomish reason, that (seeing that random pregnant-lady smile) cheered me up more than winning at “Settlers of Catan” would have cheered me up, unless, like, I was playing “Settlers of Catan” with just the right mix of Competitors. I am older than 40 by a lot so will never win a Clark Medal: I did get a trophy once at the bowling alley, when I was a kid, for my first game over 100, but the guy who engraved the little brass plaque on the trophy spelled my name wrong and I rolled a 124, not a 104, like the plaque says (I lost the trophy halfway through the 70s, and might be a little off on the 124 and the 104). (I also will not ever – never ever – get a Clark Medal because I am not remotely an economist. I am completely fine with that) . (On the other hand I do write colloquial English in the way one would expect from someone who grew up speaking colloquial English – so there’s that). (That’s not nothing.)

28 anonymous as usual April 27, 2017 at 10:37 pm

If by some extreme chance you are not a native English speaker and nevertheless read my 10:02 comment – NoVa is Northern Virginia, commercial district is a nice word for strip mall (I like this particular strip mall because it has a Trader Joe’s where the management seems to treat the employees better than the local Giant or the local Whole Foods, and it has a cool pet supplies store and a music store where I bought a violin for a wonderful friend – who plays Mozart on it from time to time, which is nice), “um” is almost never spelled out but I put it there for continuity purposes if anybody was reading that as if it were the spoken word (otherwise the repeated ‘when’ would seem bulky). Towards the end of the comment, “never ever” is one of Taylor Swift’s catchphrases and I like quoting Taylor Swift because lots of my pals in the real world (well, three or four) like to talk with me (again, in the real world) about the commonalities between Taylor Swift and Dylan. Plus, the misengraving actually happened to Charlie Brown a couple years before me – well he was a fictional character, and I did read the Charlie Brown bad trophy strips even before it happened to me but thought that people who engrave trophies care about reality, that what happened to Charlie Brown was just fiction. Well they (the people who engrave trophies) don’t (care about reality) , as I found out (and I think it was actually 134 – I clearly remember 4 strikes in a row beginning in the sixth frame and, absent a really bad performance in the earlier frames, which I don’t remember at all because it didn’t happen, 134 is the likeliest score). Congrats to Professor Donaldson by the way.

29 Larry Siegel April 28, 2017 at 2:02 am

If it’s that hard to get an excellent economic history paper published, why doesn’t somebody (say, the Mercatus Center) start a new journal of economic history?

30 Ricardo April 28, 2017 at 3:31 am

Isn’t it a signalling problem? If you are a young academic, you want to publish in the best journal around, even if the time from submission to publication is outrageous. If Mercatus were to start a new journal of economic history, how could it avoid the stigma of being a dumping ground for papers that weren’t good enough for AER, Journal of Economic History, etc.?

I suppose the economics profession could encourage people (especially professors who make tenure or hiring decisions) to spend more time reading papers and have informed opinions on them instead of looking at where the paper was published. But that is hard to do in a field like economics where economic historians won’t necessarily have informed opinions on theory papers and vice versa.

31 Larry Siegel April 29, 2017 at 9:33 pm

The problems are analogous to those of any new product introduction; the new product has to have a distinctive niche or an edge over existing products, and must be well capitalized and well advertised. In addition there is the signaling problem you mention. But new journals are started and some succeed, so one could figure out what the successful ones did and imitate them.

32 Troll Me April 28, 2017 at 10:35 am

A centre dedicated to promoting and maintaining access to aggregations of economic data for macroeconomic analytical purposes?

In practice, that could just be a website with a bunch of links, or maybe even hosting a mirror site for downloads of the data.

33 Ricky Tylor April 29, 2017 at 1:00 pm

You got to be extremely watchful for these things especially long term, as these things can really turn over the head if we are not sensible in our approach. I always work with careful way and trade through plan, it’s easier with broker like OctaFX which is top class having small spreads at 0.1 pips, high leverage up to 1.500 while there is day to day market updates for free, it’s all incredibly effective and allows me to work nicely.

Comments on this entry are closed.

Previous post:

Next post: