« www.ideastorm.com | Main | Assorted links »
Noah Stoffman asks me two questions
I have two pretty random questions about which I would love to see a discussion on MR:
What will applied economics research be like in 50 years? I spend a huge amount of time gathering, cleaning, and organizing data. I spend a lot of time writing code to do analysis. Will this become unnecessary? Will I just be able to say to my computer "Check if this relationship exists in the data"? If that happens, what will be special about people with PhDs?
Suppose you were given a large amount of money (say $10 million) and you wanted to make sure that you would remain (relatively) wealthy in as many future states of the world as possible. Where would you invest it? Remote arable land? Organizing a cult of followers?
The lesson, of course, is that "pretty random" questions rarely are. Usually it is someone asking the same question twice.
I believe with p = 0.6 that the world is in for a "great disruption." It has come to MSM first but it will not end there. In the longer run I am optimistic about the results of this change -- computers will free up lots of human labor -- but in the meantime it will have drastic implications for income redistribution, across both individuals and across economic sectors. For a core metaphor, the internet displacing paid journalism and classified ads is a good place to start. The value of newspapers has been sucked into Google.
Later, we'll be much better at measuring which research Ph.d's are contributing value and which ones are not, or at least we'll think we are. Since academic achievements follow a Power Law, that will mean a huge ouch for many would-be academicians. The new professor will need to be skilled in assembling collages of information, raising money, and communicating to broader public audiences. Either that or his research should be very obviously of the top order. The distribution of income across professors will become radically less equal as indeed the trend has been for well over a decade now.
If you have $10 million, the safest thing to do is to diversify across currencies, buy government securities of various kinds, hold $1.5 million in gold, and otherwise not invest at all. Oh yes, invest in some cheap hobbies. In a real crunch remote land is worthless -- transport costs -- and your cult followers are as likely to betray you as not. Trying to become a professor is no longer such a safe path.
Once The Great Disruption becomes more evident, entertainment will be very very cheap. Medical treatments will become either much more expensive or again very cheap. If you get the wrong ailment, you're going to need the $10 million.
Robin Hanson believes we are headed back toward a Malthusian equilibrium; in contrast I believe that machines will never outcompete humans across the board and so the scenario will more closely resemble Baumol and Bowen's "cost disease." The variance of real wages will continue to rise.
I predict the equity premium will go up.
Posted by Tyler Cowen on October 27, 2009 at 10:57 AM in Economics | Permalink
Comments
The answer to the second question is 95% in gold. Gold will have value as long as people still trade- this isn't true of paper money or government securities. Take the other 5% and buy a farm somewhere relatively remote from dense population centers, buy a Toyota Landrover along with the gas and accessories to run it for 10 years, and buy yourself 5 years worth of canned goods.
Posted by: Yancey Ward at Oct 27, 2009 11:12:29 AM
If I understand Prof. Cowen correctly, by a "great disruption" he means major economic change. A historical analogy, which I have been studying recently, would be the effect of opening the American plains and the Russian steppes to agriculture in the later 19th century, combined with the railroads and steamships to transport grain to Western Europe. The effect of these changes was to destroy the value of cropland in Western Europe, since locally grown grain was no longer competitive. This change turned the supremely self-satisfied gentry of Jane Austen into the rather pathetic nebbishes of Nancy Mitford.
Anyway, the point is, a disruption of this kind isn't a collapse or a catastrophe, merely an event which upends traditional economic activities and creates social upheaval. There isn't any reason to go for gold or canned goods. Basically, a diversified portfolio will do fine.
Posted by: y81 at Oct 27, 2009 11:23:49 AM
yancey ward, would you spend the other 5% on a police force to guard the gold?
and while you're at it, you might need a second police force to defend the gold against the first one.
Posted by: babar at Oct 27, 2009 11:29:22 AM
"in contrast I believe that machines will never outcompete humans across the board and so the scenario will more closely resemble Baumol and Bowen's "cost disease."
What probability conditional on no civilization-wrecking catastrophe?
Posted by: Curious at Oct 27, 2009 11:30:27 AM
If I were an economist and I only had special knowledge about an obscure coding language to the point where I couldn't speculate better than on a diversified portfolio, I'd invest in rectifying that situation. I might start by writing a pop-econ book about how the market is inefficient and what that really means. Things are less efficient when change happens fast.
Increasing income inequality is good if you are good. My bet is that the money will flow away from the gatekeepers. It's going to be easier to do without their expertise. We never really liked them anyway. You also won't have to spend as much time teaching basics, but then, you won't get to spend as much time teaching the basics.
People are not just going to decide that division of labor is passé. On the other hand, I think we may still be coming to grips with the fact that any building over 10 stories isn't a sure thing.
Posted by: Andrew at Oct 27, 2009 11:35:38 AM
Babar,
Nothing is guaranteed- ever. The question was really what would one do to maximize the probability of remaining wealthy over all possible future states of society given that you had $10 million today. Gold is the store of value that survives the maximal number of such future states.
In any case, gold has one other significant advantage over other hard assets- it is the easiest to conceal, and concealment is the only real answer to your concern.
Posted by: Yancey Ward at Oct 27, 2009 11:37:32 AM
Step away from the ledge...way too much fear in these comments. 95% gold? Give me a break!
Posted by: crispy&cole at Oct 27, 2009 11:40:28 AM
If I were investing 10 million to preserve wealth for a maximum of possible states, I'd rather have the resources and knowledge to make drinkable alcohol to just gold, as that's always been in demand.
Posted by: nelsonal at Oct 27, 2009 11:47:28 AM
Yes, it's possible that civilization will completely collapse... but at least everyone will still want gold more than ever!
Posted by: Barbar at Oct 27, 2009 11:47:53 AM
Crispy,
It isn't fear. I was just answering the question as it was asked. He didn't ask how to increase his wealth in the maximal number of scenarios, just how to preserve it.
Posted by: Yancey Ward at Oct 27, 2009 11:49:39 AM
My best shot might be:
50% CEDARS
30% Index Funds
20% A House
This should provide a nice income stream as well as the ability to keep pace with broad economic growth.
Posted by: Tuttle at Oct 27, 2009 11:50:48 AM
As someone who analyzes data for my job, I have to say I don't think the first question is well formulated. The relatively small set of fixed rules that can be encoded into software are the "gimmes" of data analysis - you aren't creating much value just plugging and chugging calculations and finding "relationships" in the data, and that is largely the case even now with decent statistical software. The real value creation lies in asking the right (novel) questions, organizing a specialized data set that answers those questions effectively, and then understanding the difference between a useful answer and a trivial one (lots of academics seem not to understand the last one). The real value-added work consists mostly of improving understanding of societal relationships, human behavior, and relative priorities that can always be refined beyond the current state of software.
Posted by: An Onyx Mousse at Oct 27, 2009 12:01:21 PM
Just invest in the skills of survival, basic agricultural/foraging/hunting skills, mechanical skills, and water treatment. Learn to enjoy free things. Get a decent personal library (just large classic books). You don't need $10 m to do that. You don't need large ingots of gold. You'll be fine.
Every disaster scenario implies a return to solitary man (who never really existed anyway). We are forced back to our primitive state and to be segregated from our fellow man. Here is my question: Is there a disastrous state of the world where we are forced to be more *interdependent* than we would prefer?
Posted by: Reed at Oct 27, 2009 12:02:54 PM
But what if an asteroid containing $1 trillion worth of gold hits Earth? That would be a catastrophe for gold bugs. Better diversify and hold 50% silver.
Posted by: anon at Oct 27, 2009 12:04:15 PM
Best insurance against "the future" is to be part of a cohesive community away from an urban center that is as self-sustaining as possible.
In other words, give a look to the Amish or Mennonites.
Brian
Posted by: Brian Timoney at Oct 27, 2009 12:14:14 PM
@nelsonal I think that's too easy. All you need is a sugar, water, and yeast. People in prison manage to make alcohol.
Posted by: bob at Oct 27, 2009 12:15:46 PM
I think the point of gold is that if you have to have some value in nearly all possibilities, you are left with gold and guns.
I don't believe there is a high probability of worst case, but I did just buy some more guns. They will hold value as well. When you get a good bit of money, you start to realize how cheap it is.
Posted by: Andrew at Oct 27, 2009 12:19:34 PM
Robin Hanson's Mathusian scenario occurs at the tailing edge of the far future, even after the rapacious hardscrapple frontier scenario plays out, not something we need to hedge against within our own lifetimes.
Posted by: Sam Wilson at Oct 27, 2009 12:26:07 PM
Since academic achievements follow a Power Law...
You mean you think that there will eventually be achievements in any human endeavor that don't follow a power law? The more you jack up productivity in any given area, the steeper the power distribution curve gets.
I think Onyx Mousse is on the right track. You can always crowd-source the data analysis, but somebody has to ask the right question. I doubt that asking the right question will require academic credentials, though...
Posted by: TheRadicalModerate at Oct 27, 2009 12:29:27 PM
As for investment, I would say normal stuff like equity and bonds should be around 50%. 1 million will buy you a TON of survivalist equipment and training. The rest can go into real estate and campbell's soup and all that countercyclical stuff.
This "great disruption" is going to be centred around the lowering of marginal costs far below what people are used to seeing. Anything that can be reduced to 1s and 0s will become practically free. Unlike Tyler, I think this will lead to a dramatic increase in global equality. Sure, people's incomes will diverge, but so what? If you can get all the entertainment and education and information services you want for free, you will be better off. I don't see a possible world where basic necessities will increase in real cost. The price of food, basic shelter and clothing has been dropping in terms of labour required to get it for a long time. It is entirely plausible to me that there will be an invention that makes food unscarce. Likewise, increasing robotics will make manufactured goods increasingly cheaper.
Keep in mind that anyone can maintain older styles of living simply by not trading with the outside world. Think of the Amish and the logic of comparative advantage. Dystopian futures are implausible on this basis alone.
Posted by: azmyth at Oct 27, 2009 12:35:11 PM
@An_Onyx_Mousse and anyone else: aren't there data-mining methods that can do the job of finding questions to ask? Principal components analysis, self-organizing maps, etc?
Posted by: Silas Barta at Oct 27, 2009 12:42:19 PM
To question 2, Hedge the things that you think you will want. Energy and commodities can be hedged. Medical care is a little tougher, you could invest in drug companies. (Maybe you could send your kids to nursing/medical school or go yourself). Invest in people family and close friends. Them certain index funds VTI, PID, etc. Pay off your home and add insulation and do other improvements that have payback.
Posted by: Floccina at Oct 27, 2009 12:49:13 PM
Will I just be able to say to my computer ...?No. Computers will make PhDs less valuable by making them more productive but not this way. Natural language processing and "common sense" reasoning are notoriously difficult for computers, not because computers differ so fundamentally from brains but because brains incorporate far more information than we imagine. PhDs themselves don't realize how much diverse information they bring to bear on analytical problems, and computers don't possess all of this information. Computers can't easily pass a Turing Test for the same reason.
Related question: Shouldn't a more productive PhD become more valuable? No. Lords of the productivity enhancing capital become more entitled to value added. If PhDs themselves are lords of this capital, their enhanced productivity can become more valuable to them, but that often doesn't happen in practice. The "valuable" factors will be lords of propriety, like the lords of feudal Europe, as they are now.
Where would you invest it?The answer depends heavily on the current state of the world. If I give you $10 million in Federal Reserve Notes for title to 10,000 acres of remote, arable land, I've only exchanged one stack of paper for another. Do I trust the armed men securing the value of the title more than the armed men securing the value of the FRNs? If future states are highly uncertain, why prefer one over the other? What is my title worth when the established protection racket peddling it disintegrates?
If I were really worried, I'd buy a dense, valuable commodity (dense in value, not necessarily in mass) and bury it in diverse places. I don't need to own these places. I only need to know that most of them will be accessible to me. I'm not worried about trespassing in this scenario, but I'd bury a few small arms along with other commodities. I'm not sure gold is one of these commodities. Maybe they're seed for very high yield, genetically modified foods.
If you have $10 million, the safest thing to do is to diversify across currencies, buy government securities of various kinds, hold $1.5 million in gold, and otherwise not invest at all.You assume that most established states remain viable. It's probably a safe bet, so "investment" in entitlement to tax revenue (government securities) remains attractive, but with this assumption, I might buy title to land instead. If entitlement to tax revenue becomes too attractive, established states stagnate.
Posted by: Martin Brock at Oct 27, 2009 1:31:50 PM
People used to sit on their front porches or at the bazaar chattering. Now they sit in their blogs or forums on the internet chattering.
I predict in the future people will pretty much be doing the same: chattering.
(And in both cases, most of the chattering will be (a) wrong, (b) somewhat pointless, but still (c) entertaining--passing the time--because that's pretty much what people are doing--passing the time. And while the chattering on the front porch is lost in the noise of the wind, and the chattering on the internet seems more permanent, in fact the noise of the internet wipes out the internet chatter just as well as that of the wind--all things considered. Hence no change.)
Posted by: T.R. Elliott at Oct 27, 2009 1:33:22 PM
About data analysis and PhD's. Data analysis is essentially organization of data. Statistical methods help you do that, but in the end of the day your data analysis will only be any good if you know your data really well. Cleaning the data is really making a judgement call about what data-points are worth including in the analysis and which ones are not. You can't do that well just using mechanical methods.
Also: you can always do more complicated things. 40 years ago you could write a paper with a single linear regression. Now a days you do bayesian estimates of structural models.
Last, the ability to process data will increase, but so will the amount of data available will be much larger and you will need a lot of skilled people to sort through that. The trend that I see right now is empirical work becoming more and more important, particularly in macro where it was second class. And this is all good.
But I agree with Cowen that it will be harder to make a living in academia just from technical skill. Or at least, I hope so.
Posted by: Felipe at Oct 27, 2009 1:48:31 PM
