David R. Henderson  

Krugman on "Unproductive Finance"

Why My Billion-Dollar Plan Won... NSA Surveillance: A Cost/Benef...

In the last couple of weeks, Paul Krugman's blog has been a target-rich environment. In the next few days, I'll have one or two more posts on recent Krugman posts, but one this morning caught me eye.

I'll quote almost the whole short post, title "Unproductive Finance." Krugman writes:

More than half a century ago, in his classic paper on the economics of speculation, Paul Samuelson noted the perverse rewards to knowing stuff just slightly before everyone else. He asked readers to imagine someone who, somehow, consistently received crucial information one second before everyone else. As he pointed out, the social value of that extra second would be minimal; but the private rewards could be huge.

Today, Kevin Drum points us to a real-world example quite close to Samuelson's thought experiment. It turns out that Thomson-Reuters pays the University of Michigan a million dollars a year to provide selected clients with the results of the latest survey of consumer confidence 5 minutes before the rest of the world sees them-and to provide super-special clients with this information 2 seconds before the rest.

This is a trivial example; still, what we see here, as Drum notes, are real resources being devoted to the socially useless task of getting an economic number slightly before the hoi polloi.

On the first paragraph, I have no objection. Krugman states Samuelson's classic accurately and Samuelson made an excellent point.

The second and third paragraphs are the problem. The key issue here is whether the Thomson-Reuters case is an instance of what Samuelson talked about. The basic idea is that the resources spent on getting information early are wasted. But are they?

There are three possibilities, but before getting to them, let me point something out. Remember that someone needs an incentive to get the information that the University of Michigan gathers. If the information on consumer sentiment is worth getting--and neither Krugman nor Kevin Drum, whom he cites, argues that it's not--then the University of Michigan needs to make money from gathering it.

Now to the three possibilities:
1. If the payments made by Thomson-Reuters and others who get the information earlier are needed to give U. of M. the appropriate incentives to gather quality information, then the payments are not wasted.
2. If the payments made by Thomson-Reuters and others who get the information earlier are not needed to give U. of M. the appropriate incentives to gather quality information, then the payments are producer surplus to the U. of M. and there is no social loss from the payments--it's just a transfer.
3. If in case #2 above, the producer surplus is used for low-value uses at U. of M.--this is both a non-profit university and a government university, after all--then there is a waste.

So only in case 3 above is it "unproductive finance." I'm pretty sure Krugman isn't going with case 3.

Comments and Sharing

COMMENTS (37 to date)
Daniel Kuehn writes:

Doesn't the surplus transfer you discuss fall under the "private rewards" that Krugman mentions? Isn't the whole point that Thomson-Reuters's ability to profit from the early information - while a private gain - doesn't provide any substantial social benefit.

Am I missing something? Is there something other than the private benefit associated with the exchange going on?

Both Samuelson and Krugman pay lips service to the subjective theory of value. If I want to pay a million dollars to get the information one nanosecond before everyone else, and if value is indeed subjective, then it doesn't matter if someone else's judgment says I paid too much or that the provider of information invested too much in gathering the information. Even if that someone else is a Nobel laureate, it is still just his personal, subjective judgment that has no more objective weight that anyone else's judgment.

Emily writes:

The point they're making isn't that you paid too much (that you are acting inefficiently for you), it's that you're expending resources in a way that benefits you, but the overall transaction is socially negative - in other words, that we're in David's case #3 scenario. This would be the case for any zero sum game transaction that eats up resources to play. These scenarios certainly exist and they don't require anyone to challenge the idea that you are correctly assessing the private value of something to you.

Hazel Meade writes:

I think the perceived problem here isn't really that it is "unproductive" but that it seems to be unfair to other people in the stock market. If you have access to the information 2 seconds early, you're obviously using it to make trades just a split second before the information hits the market. There doesn't seem to be any social value to THAT activity, although maybe someone can explain it to me.

Emily, the claim that a transaction is "socially negative" implies a quantitative judgment about my "private" benefit relative to other people's benefits. My private benefit must be less than the total loss experienced by others. This is where objectivity of value gets smuggled in.

Daniel Kuehn writes:

Emily I agree that Predrag is misunderstanding the claim (if you lead with assuming Samuelson only pays lip service to subjectivism that oughta send up serious red flags), but I'm not sure this even requires a #3 scenario. I think the point is simply that a lot (who knows how much) of finance may be resources dedicated to getting the better of what David calls a "transfer" of services. This is like Keynes's point that finance is like a casino. You don't have to assume that casinos have a negative impact on welfare. They don't. Insofar as people enjoy gambling they have a positive impact actually. But you wouldn't claim that they do much to efficiently reveal privately held information.

That is the usual claim about finance by economists. And it's certainly true to a large extent. The question posed here (and the interesting U of M example) is how much of finance is more akin to transfers of private surplus or casinos.

Daniel Kuehn writes:

* - transfer of *surpluses*

Jardinero1 writes:

I quibble with Samuelson's premise in the first place for three reasons.

1. There is always assymetry in the information delivery. There is no statistical proof that being first in line leads to being first in profit.

2. Assuming, for the sake of argument, the first in line is better off, earning a profit in finance entails two parts: knowing when to buy in and knowing when to sell out. To profit more than the rest, you would have to be first in line with perfect information at both ends.

3. Again, assuming, for the sake of argument, the first in line is better off, then those behind him. Those behind him are also better off knowing he's in front of them and so profit by following the leader. After each iteration of this follow the leader, the leader's advantage wears off and the followers advantage improves until both plateau with no advantage accruing to either.

MikeP writes:

Price discrimination. How does it work?

...especially when one of the price choices is "free".

Daniel, I don't think I misunderstood the question; I just focused on the meaning of "socially useless" in the first place.

In this particular case, even if we are in case 2., the fact that some market participants chose to pay a million dollars for early information, and others didn't, indicates something about their time- and place-specific personal knowledge of what they could do with that information.

Aaron Zierman writes:

I'm not sure I understand what Krugman means when he talks about "real resources being devoted to the socially useless task". Thomson-Reuters and their clients benefit or they would not spend the million dollars. The remainder of society benefits in that they (after a brief period of time) get the information for free.

Addtionally, since when does something have to be considered "socially useful"? I don't buy a steak in order to help the seller or the producer. I buy it because I'm hungry and it tastes good.

The simple fact is that there is a market for this information and that whoever values it most can purchase it.

Daniel Kuehn writes:

Predrag -
There are two very reasonable interpretations of "surplus" here, the question is what is the most reasonable in the context of the post.

1. Social meaning the sum of all private benefits, or
2. Social meaning the benefits to society besides the private benefits

Both uses are pretty common.

Since Krugman clearly said there are private benefits #1 would be wacky, don't you think? I know you don't like the guy but surely you don't think he's just a babbling idiot.

We ultimately think things like speculation are good because they elicit privately held information. Assuming David's #2 is right that's not what this annual million is doing. It's not bad obviously. It's kind of nice that U of M can appropriate some of the surplus this way. But it's not the normal sort of functionalist narrative we ascribe to speculation.

Roger McKinney writes:

I think everyone misunderstands Krugman. He doesn't write about economics. He only uses his expertise in economics to lend credibility to his political/social views. This is his typical bait and switch. He teases with a lead about economics to build credibility and then switches to a socialist argument about the value of different types of work.

I have no doubt that Krugman would agree completely with David on his economic analysis. But that's not what Krugman is arguing. Essentially, he is saying the "right" kind of people, people like him and Keynes and others with the "right", meaning left politics, all understand that finance is a waste of time and resources.

RPLong writes:

I think I agree with Predrag.

Suppose we are in a 2-person economy consisting of Ryan and Predrag Economicus. Predrag grows apples that he values at $5/bushel. Ryan catches fish that he values at $5/dozen.

In one sense, equilibrium would occur when Ryan trades a dozen fish for a bushel of apples. But what if Predrag values fish at $15/dozen and Ryan decides to raise his price in accordance with the actual demand for fish? Ryan gains a surplus, and Predrag does not suffer a loss of utility.

Now suppose Ryan decides to throw his surplus of apples into the mouth of a volcano. In one sense, this is a "socially useless" expenditure of apples, since neither Ryan nor Predrag gain apples or fish from Ryan's having tossed his surplus into a volcano.

...But in another sense, why on Earth would Ryan toss surplus apples into a volcano, except for his own personal utility?

Maybe Ryan ought to be convinced that he could gain more by putting his apples to an alternate use, and maybe Predrag would gain something from making that argument to Ryan. But, as it stands, Predrag suffers no loss, and Ryan spends his surplus according to his subjective utility.

That's my argument. Where have I gone wrong?

RPLong writes:

DK - Can you clarify or give an example of what might be a "benefit to society beside the private benefits?"

Daniel Kuehn writes:

I think that's right Ryan.

Now let's imagine a world where tossing apples into a volcano was generally thought to be a very important industry for allocating funds and assessing the value of business enterprises.

Wouldn't it be nice for someone to say "actually the whole apple-into-the-volcano thing is mostly just about Ryan satisfying his private utility, which is great but not what we often think it is".

jason braswell writes:

Daniel (or anyone),

Where would potential "ancillary" benefits of this behavior fall into the calculation? For example, getting information faster requires faster networks, and getting a lot of information fast requires big, fast networks and networking equipment. I can think of several examples of technological improvements that were funded by HFT players but have application outside of HFT as well. These improvements may well not have come (at least not as early) were it not for HFT races.

Shouldn't this count as social value?

Daniel Kuehn writes:

I am not standing by that analogy as the best one for explaining Krugman :-P I'm just trying to push it closer to the point that I think Krugman is actually making.

I think David's initial list was actually the best exposition so far - and Krugman is at #2 and noting that a lot of finance may just be about chasing rents (in the original Ricardian sense - not in the negative sense) around rather than eliciting new, valuable information.

Daniel Kuehn writes:

Jason - definitely

Daniel, I am familiar with those (and other) interpretations, and I would be happy to discuss them with you at some other occasion.

I was commenting on the literary exposition of those two economists in relation to the subjective theory of value, not my preferences about them as people.

I don't think it is necessary to put a value judgment on the conclusion that market prices reveal otherwise unobservable information. I personally think it is a remarkable insight, but some people may not see much value in it.

jason braswell writes:

I don't fully understand why everyone is SO confident about the zero-sum theory of HFT, especially in comparison to that theory as it applies to a lot of other activities.

Full disclosure: I work in HFT. Despite that, I'm pretty sympathetic to the zero-sum theory, and I think probably a good portion of HFT efforts are zero-sum.

Nonetheless, I think the Krugman-esque worries that a lot of people talk about are overblown for a lot of reasons:

1) There ARE ancillary benefits like the ones I talked about. Besides tech benefits, US exchange profits from HFT seem to encourage the creation and improvement of exchanges in other countries where traditional "slow" trading may not have been sufficient.

2) There's a very, very natural rate limiter to this behavior, and we're clearly already seeing now with the huge drop in HFT profits. You can't go below zero time! Given that HFT firms already think on microsecond scales, there's not a lot of slack left. As firms achieve similar speeds, competition will favor the ones who tease out truly new information in the markets.

3) There are a lot of behaviors we tolerate and enjoy where the social benefits are far from obvious. As much as I enjoy it, I'd consider economics blogging to be a great example of such. What is the real probability that regularly reading Krugman's blog (or this one) is going to, for example, make a person a more rational voter or agent? So much more so that that person's more "improved" behavior is going to spill over and create economic benefits for non-readers?

What's more likely is that blogs like Krugman's make a small number of people a little more narrow-minded and less rational as it stokes so much rah-rah groupthink.

Furthermore, the private rewards that Krugman experiences for his public activity seem likely to draw other economists to spend their time blogging instead of, say, writing awesome papers and digging up new, cool data?

Tracy W writes:

Daniel Kuhn: David outlines some possible explanations:
Possibility 1: Getting to see the information 2 seconds early helps pay for the information to be collected in the first place, ie the university wouldn't collect this information if it wasn't paid by Thompson-Reuters. In this case, there's a social benefit created.

Possibility 2: The university would collect this information anyway. But the university does something socially useful with this extra money. Social gain. Universities are generally regarded as being socially useful, which is why rich people occasionally donate money to them.

Possibility 3: The university would collect this information anyway, and it'll just waste the extra money.

Daniel Kuehn writes:

Tracy - right... I've referenced those three several times now.

I don't agree with your spin on #2 (which I think is what Krugman is communicating). It's simply a "transfer" (as David says) of surplus in that case. I don't think he's invoking any kind of universities-create-social-benefit assumption. This is the private gain that Krugman was talking about.

Tracy W writes:

[Comment removed pending confirmation of email address. Email the webmaster@econlib.org to request restoring this comment. A valid email address is required to post comments on EconLog and EconTalk.--Econlib Ed.]

David C writes:

There is a fourth possibility.

4) People gaining by being first imposes secondary costs on the market, and these costs exceed the benefits to the University of Michigan.


1) If the most successful people spend all their time trying to be first, innovation rates will decrease. (I suspect this is Krugman's view)

2) Thomson-Reuters is simply transferring money from one source to another, and extracting costs of their own to pay for the cost of transaction. It is possible that if Thomson-Reuters did not exist, another transaction of nearly equal or greater value would be made instead with lower transaction costs.

3) By transferring money in a way that does not improve the market, Thomson-Reuters is reducing the market share of firms which would be able to improve the market.

mike davis writes:

We should also consider one other positive benefit from things like HFT and Thomson-Reuters: Wider understanding of such practices may help convince noise traders to become passive investors. I'm sure almost of all of you get asked to give investment advice (I know! I wish people understood the differences between economics, finance and business.) And I'm sure almost all of you tell people to buy low cost index funds and then take the kids swimming instead of trying to time the market. People seldom take this advice. If they think information traders have a huge advantage, maybe they will.

On the other hand, now that I think about it, we should also consider the possibility that HFT, etc. makes people so cynical about markets that (1) the noise traders get even crazier as they they look for alternatives to conventional investments (like Vegas real estate), (2) noise traders just don't save as much, (3) people in general distrust the efficiency of all markets and so become more accepting of inefficient regulations.

Hazel Meade writes:

What if companies sold their quarterly reports to some investors 5 minutes before they were published, or 2 seconds before that to other investors?

How is this different?

If we're operating under the assumption that markets are more efficient with more information available to more participants, then why wouldn't it be most efficient to release the information 5 minutes and 2 seconds early to everyone?

mike davis writes:

One more thing, John Cochrane has an interesting paper with the provactive title "Is Finance Too Big". I gather some version will appear in JEP and there may be later versions out there but here is the link to the version on his blog. If you missed it, H


John B writes:

Saying that a lot of what finance does is socially unproductive because it's only shifting money around is like saying that the circulatory system is unproductive because it just shifts blood around.

Since not everyone acts on new data at the same time, the ultra-fast traders are helping to smooth out price adjustments and provide liquidity.

The reason HFT causes problems, like a lack of liquidity during an event like the flash crash or 1987, is due to poor algorithms and decision making. Portfolio insurance, the cause of the 1987 crash, is a program that specifically sells when prices go down and buys back in once prices have risen substantially. In other words, it creates problems because it violates the basic rule of buying low and selling high.

The more profitable the financial system is, the smoother markets are. But, don't argue that the financial system during the 2000s was profitable. Most of it wouldn't exist if it weren't for the Fed and TARP.

Roger McKinney writes:

The only objective way to tell if something is valuable is to offer it for sale. If someone will pay for a service, it's valuable to that person even if it's not valuable to Krugman. And who cares whether Krugman finds something valuable or not?

Roger McKinney writes:

Jason, I have read very few economists who had any respect for the stock market and that's probably because they are so ignorant of how it works and the role it plays in the economy. Ask any economist where he has invested his savings; most will tell you that they are in zero coupon bonds or something similar. Most just repeat Keynes' unbelievably asinine comments. BTW, the only time Keynes made money investing was when he traded on inside information for which he would go to jail today.

The only economist I have read who understood the stock market is Ludwig Lachmann. Check out his "Capital, Expectations and the Market Process" and "Capital and its Structure." They're not just about the stock market, but have brilliant insights into the vital role it plays. Here's an except:

The above throws some light on the function of the stock market in the market economy. Here, shares of different capital combinations, consisting essentially of fixed capital goods, are continually being evaluated. The stock exchange not only registers success and failure, but also expresses expectations about the prospects of plans already set in motion. The stock exchange may be viewed as the central forward market for future capital yields of indefinite horizon. Buyers and sellers on the exchange express their expectations about the chances of various plans, and thereby also evaluate the underlying capital combinations.
david writes:

You can bid wastefully for a transfer, if the competing bids are wasted rather than paid to the auctioneer. The auctioneer might still auction anyway: they still earn your bid less their private valuation, which may be positive.

Arthur_500 writes:

One mans waste is another mans treasure.

Does everything in life need to have a social benefit? One cannot conduct commerce unless it benefits 'socially?'

What about a pet rock? Someone made money on this useless idea because someone thought it was silly enough to buy it. There is no redeeming social value in the pet rock so would Mr. Krugman feel it appropriate to eliminate this type of commerce?

Actually I am sure Mr. Krugman would like to have another government czar decide what is valuable commerce and what is not. May I be that czar? Discussion by Mr. Krugman in a syndicated column for his private enrichment when he should be focusing on teaching or research that has 'real' social value would be outlawed.

Please Mr. Krugman, go away. Go far far away. We are tired of listening to your ideas that have no real merit and then getting told over and over that you won a prize for research that you deny through your actions.

Since when does mankind need to limit its actions and economy to only social merit?

Wojtek Grabski writes:

Hazel, that's my thinking too. The implication of the article is that the practice is unfair. This is strictly true, but not because of any economic argument.

It is common practice for theatre companies to offer tickets in advance to select clients. These clients pay a fee for the privilege, and anyone who wants to pay is entitled to it.

The part of the referenced arrangement that irks people is that the same menu of choices isn't offered to all potential clients. If the university posted a schedule of fees for early access then there would be no quibble with their actions. Socialist angst over the current practice is that it exposes that seemingly free information actually isn't, and that an institution that purports to supply it as a social benefit is acting within Locke's definition of unjust pricing -- they're giving one client a preferential price.

The economic argument for this action is of course that it probably fetches them a higher price; which is the combined value of the money, which pays for the polling, and the appearance of providing a free service to the community.

prometheefeu writes:

Of course, this isn't the best example because at worst, we just have a transfer of surplus to the University. However, there are the cases of firms buying very expensive real estate near stock market exchanges. I very much doubt that data centers are the highest value use of real estate in the middle of New York. If those firms colluded so they could all put their datacenters outside the city on substantially cheaper land, both the firms and their customers could save a lot. Or an easier way to do this would be to change order execution algorithms by batching orders for 1 second and randomizing their order before executing. That wouldn't impact the price discovery mechanism, but there would be little benefit from being a few millisecond before the other guys.

ezny writes:

Aren't there two mistakes with the argument (one is Krugman's the other Henderson's)?

a) Krugman says the expenditure is socially useless (Samuelson, instead, talks of 'perverse rewards'). We are talking about the distribution of profits from the exploitation of information in financial markets - social uselessness or otherwise doesn't come into it. It might be perverse to get a huge reward from knowing things first but if information were distributed more evenly the same reward would simply be spread more widely. By selling the info, U Michigan takes a cut from the extra profits of the early movers.

b) U Michigan needs to pay for information gathering. But information is a public good and therefore it should be available freely and financed through taxation rather than limiting access to the highest bidder.

Steve Wise writes:


I have to disagree with your second assertion. Information is NOT a public good. Information, methods, and work product (especially in today's world) are in many cases proprietary and hence protected under multiple statutes.

Companies, universities, and private foundations spend a fortune attempting to create informational or analytical work products which give them an advantage.

Should such an organization choose to offer access to developed work products they are well within their right to collect economic rents from parties willing to pay for access (or in this case, priority access).

Comments for this entry have been closed
Return to top