Wednesday, March 14, 2012

Can people be fooled again and again? (Goldman Sachs edition)

A very interesting article in the NYT today, in which Greg Smith gives his reasons for resigning as an executive director at Goldman Sachs. Apparently, he's tired of his firm's practice of ripping off its clients:
What are...quick ways to become a leader [at Goldman today]? a) Execute on the firm’s “axes,” which is Goldman-speak for persuading your clients to invest in the stocks or other products that we are trying to get rid of because they are not seen as having a lot of potential profit. b) “Hunt Elephants.” In English: get your clients — some of whom are sophisticated, and some of whom aren’t — to trade whatever will bring the biggest profit to Goldman. Call me old-fashioned, but I don’t like selling my clients a product that is wrong for them... 
It makes me ill how callously people talk about ripping their clients off. Over the last 12 months I have seen five different managing directors refer to their own clients as “muppets,” sometimes over internal e-mail. Even after the S.E.C., Fabulous Fab, Abacus, God’s work, Carl Levin, Vampire Squids? No humility? I mean, come on. Integrity? It is eroding. I don’t know of any illegal behavior, but will people push the envelope and pitch lucrative and complicated products to clients even if they are not the simplest investments or the ones most directly aligned with the client’s goals? Absolutely. Every day, in fact. 
It astounds me how little senior management gets a basic truth: If clients don’t trust you they will eventually stop doing business with you. It doesn’t matter how smart you are.
If Smith is to be believed, and Goldman makes a large portion of its money by tricking clients into making financial mistakes, then this raises an extremely interesting question: How has Goldman gotten away with this for so long? 

Economists generally believe that you can't fool people for very long. In fact, this idea is the cornerstone of many very important models. George Akerlof's legendary "Lemons" Model, for example, assumes that if people know that they might be getting ripped off, they'll just exit the market (and hence avoid getting ripped off). In macro, Robert Lucas' Rational Expectations Hypothesis rests on the idea that policymakers can't repeatedly trick agents in the economy.

So before we conclude that something is very wrong with basic economic axioms, let's think of some ways that Greg Smith's report might be consistent with our prior beliefs:

Possibility 1: Goldman's clients aren't really getting ripped off that much. In real markets, every transaction involves some "surplus" for both the buyer and the seller. It may be that Goldman's "ripping off" of clients is just the way Goldman takes its share of the surplus - i.e., that getting slightly "ripped off" is an inevitable but bearable cost of doing business with Goldman. It might be that all of Goldman employees' braggodochio ("We took those clients for a ride, hahaha") might just be hot air.

Possibility 2: The government somehow forces people to do business with Goldman. In some markets, the government guarantees a firm a monopoly, through regulation, preferential deal-making, etc. An example is utility companies; another is the notorious ratings agencies. It may be that some combination of government regulation, preferential deal-making, or implicit bailout guarantee may be forcing businesses to do business with Goldman instead of some smaller or less well-known or well-connected (but more honest) investment bank.

Possibility 3: Learning takes a long time. Smith says that "if clients don’t trust you they will eventually stop doing business with you." But how long is "eventually"? Maybe if your firm hires a bunch of the smartest (and least scrupulous) finance people, they can keep tricking people for decades on end. Maybe Goldman will die now that people have caught on.

But suppose these don't fully explain the situation? As scientists - or even just as rational thinkers - shouldn't we entertain the possibility that people can be fooled again and again, and simply never learn that they're being fooled? Last fall, I saw George Akerlof (he of the Lemons Model) give a talk called "Phishing for Phools," in which he raised exactly this possibility. How could that happen? Well, let's think of some ways:

Possibility 4: Behavioral effects are very strong. It may be that some sort of deep-seated psychological processes in human beings are too strong to be overcome by rational learning within any reasonable time scale. In other words, maybe we are built to just keep believing the same lies over and over and over. That may sound improbable, but sometimes when I look at U.S. politicians' campaign promises (How long have Republicans been promising to cut spending?). I wonder. Alternatively, people may un-learn what they learned in the past.

Possibility 5: The influx of trick-able people may be faster than the learning process. If enough suckers are born every minute, it may be that the total number of suckers grows over time, even if some fraction of the suckers are always wising up. Thus, there may always be more clients for Goldman to rip the faces off of, even if no one keeps getting their face ripped off forever.

Possibility 6: Rapid structural change may make learning impossible. As Andrew Lo demonstrates in this paper, if the underlying structure of the economy changes at about the same rate that we learn about it, there may never be a model that allows us to understand the economy. This goes for individuals' rationality as well. If the financial world undergoes continuous structural transformation, it may be possible for a Goldman Sachs to keep ripping people off as new structural changes emerge.

I'm sure there are some possibilities I've left off of this list. But I think that figuring out the relevant importance of these effects is crucial if we want to understand what we should do - if we should do anything at all - about the shenanigans of firms like Goldman Sachs. I don't pretend to have the answer. But simply assuming that people can't be fooled is a luxury that we economists may no longer have.

Update: A bunch of people are putting forth another hypothesis: that Goldman has only been ripping people off for a relatively short time. But in Liar's Poker, Michael Lewis discusses a culture at Salomon Brothers in the 1980s that is almost identical to what Greg Smith describes at Goldman. Furthermore, Lewis says that all the big U.S. investment banks (of which Goldman was one) were doing the same thing, and that Europeans often marveled at the willingness of American investors to keep their money within a small circle of oligopolistic investment banks that were obviously ripping them off. Yes, Michael Lewis' book is just an anecdote, but then again, so is Greg Smith's article. 25 years of face-ripping is a long time.

Update 2: Justin Fox also pooh-poohs the idea that Goldman's face-ripping is a recent phenomenon.


  1. having worked with GS many many times, you go into every deal questioning every detail. Anyone who gets into bed with them first asks how are they making money off the deal, and how much. Where's the catch? How far does this go? They want to build and maintain the relationship and get to the next level, get repeat business.

    Of course they try to upsell you!!! I didn't need that extra cheesecake for dessert either, but the cute waitress said it would be so yummy!

    Nothing in the NYT article is really news to people who have done business with Goldman. So why do business with Goldman when they are so clearly using you?

    Because you get to use them back. No illusions, just business.

  2. "I know it's crooked, but it's the only game in town." - Canada Bill Jones

    If you need to invest a large sum of money and your choice is between someone like Goldman Sachs who you know are going to skim the till and MF Global who will simply make the whole till disappear then a rational person will choose the Goldman Sachs of the world even if they know they are being played for fools.

    Of course retail investors are washing their hands of Wall Street and Wall Street bonuses are falling because some people refuse on principle to be taken advantage of.

  3. Even Warren Buffett used Goldman Sachs to get deals done. The fees are high but they've got a good network.

  4. Avik Roy presents another possibility at Forbes: Goldman hasn't actually been ripping people off for that long. He argues that the change in corporate culture didn't really happen until GS went public in the 90's, after which the financial incentives shifted from encouraging long-term growth to demanding high quarterly profits.

  5. We're not living in 1810 anymore. The world is extremely complicated, and for many things you're going to keep being fooled unless you spend days, weeks, months, years, full time educating yourself in the area in question. Often – and this is very important -- you have no idea you've been fooled; you have no idea that there was a vastly better deal out there, or that years later you'll pay for what you didn't know (see: And, of course, anyone living in this world, not a freshwater model, knows people are superbusy, and have little desire to constantly study to not get ripped off, even if the time did magically exist.

    I know this very well for a number of reasons: (i) I've taught personal finance for close to 10 years to thousands of students at the University of Arizona, and hundreds of thousands through my business National Personal Finance Education. It takes a ton of learning, not to get ripped off, or make pretty, or very, sub-optimal decisions. (ii) At one time I was an agent at an auto insurance brokerage. Many agents (salespeople) made a great living constantly ripping off customers, and there was an endless supply who didn't have a clue, or the time or self-discipline to shop and find out. For more on this see:

    And as you said, people only live so long, "There's a sucker born every minute". Again, we don't live in a freshwater model; people aren't infinitely long lived. There's always a new crop of young people who haven't learned yet (again see:, and by the time the older people have learned, they may not have much longer to live.

  6. Smith's claim is that Goldman hasn't t been this way for long. A broader claim is that investment banks in general weren't t always like this. Notably. 3 of 5 major investment banks recently failed. A possible explanations is that evryone agrees that you can't fool people forever, so back when investment banks were partnerships they were rationally honest and now their shareholders are among those fools being separated from their cash.

    I think I can think of a 7th explanation based on uh old DeLong et al noise trader models people chose between two suboptimal strategies -- trusting investment banks and buying safe assets ( what Atriis means when he reconnens the money under your mattress strategy. The loss from playing it safe (largely the huge equity premium) implies that getting cheated by investment banks is a relatively ok strategy. So people don't t stick to the same mistake but move back and forth between 2 ( really more). GS is fine till people try the strategy of buying and holding the market -- in fact one doesn't need expert advice to beat money under the mattress. When people Lear that, GS is doomed.

    Making the same mistake again and again and again is implusible, making one different mistake after another is very plausible. This doesn't appear in models of learning, because the allowed mistakes are few by assumption. The result that people eventually act Nashly is the direct result of this assumption made for convenience and not from the interesting hypothesis that we can learn.

  7. Anonymous11:26 PM

    I think the issue at hand is that Goldman shifted from being a financial advisor to actively trading, and many of its large institutional clients didn't realize what was happening and got financially raped.

    Everyone is slowly learning now that you can't really expect Goldman to give you "advice" so much as you can expect Goldman to take your money, but its way too late for the pension funds, municipalities, etc that expected they were getting sound financial advice.

    This doesn't mean Goldman will fail, it means Goldman will make most of its money trading, and less money advising.

  8. Most recent Anon:

    But doesn't that mean the Volcker Rule will slaughter Goldman?

  9. 7. The tricksters learn faster than the marks. Yes, rational people will make an effort not to be fooled. But other rational people will make an effort to fool them. How can we say a priori that the first effect will completely dominate? (Especially if the tricksters, as they usually are, are a small group of specialists and the marks are widely dispersed and only engage in the transaction occasionally.)

    I've never understood why, even in principle the Lucas argument should apply that people end up holding a true model of the world, as opposed to holding some weighted average of the true model of the world, and the models it would be most beneficial to the people they interact with for them to hold.

    (Where people they interact with includes themselves: the most useful beliefs aren't always the most enjoyable ones. Like, say, the beliefs that you are not a financial genius, and that the Goldman guy is not your friend.)

  10. JW:

    So why don't the marks exit the market?

  11. AngryKrugman1:21 AM

    Does the fact investors themselves are usually simply agents play a role? They don't necessarily have the same interests as the people giving them money to invest--principle-agent problems exist in all these situations. They may, for instance, use Goldman for plausible denialbility--using Goldman to complete a deal sends a signal to one's bosses. Or, they may simply have misaligned incentives--they get the benefits of a deal with Goldman without having to bear the costs.

  12. Noah-

    My point is that we could just as well ask, Why don't the con artists convince the marks to stay in the market? When it is rational for you to decide to X, and it is rational for me to persuade you to do Y, then "people behave rationally" no longer tells us anything about the outcome.

  13. No rational person would put their money in an investment firm.

    A thousand studies have shown this, during a very long time.

    Still they exist.

    And today you seem surprised that people do not have the
    information/knowledge they should?

    REH – seriously?

    There are millions of articles and books written on economics, billions of hours spent by rather intelligent people, and what do we get for it? That the current dominant view on knowledge acquisition, expectation formation and human data computation is “know everything, present and future, and gets everything more or less correct” is unspeakably sad (but hey, you might argue that it has progressed from the “know everything present and gets it right” we had 150 years ago). How much more recourses has to be spent in order to reach at least a somewhat nuanced view? Is it even worth it?

    Someone should make a cost benefit analysis on the subject of economics – maybe the recourses could find a better use outside this completely stagnated line of inquiry.

    PS: love your posts – woke up on the wrong side of the bed this morning.

  14. I think this was realized a long time ago by PT Barnum. You can fool all of the people some of the time, and some of the people all of the time. Economists need to learn what hucksters, con artists and investment bankers already know.


  15. Richard H. Serlin hit the nail on its flat, metallic head.

    It comes down to information costs. There is a real cost to acquiring and processing information about every single thing you do. Even sophisticated pension funds don't have enough resources on hand to invest thousands of man-hours figuring out every detail of every trade with GS.

    More importantly, it's oftentimes impossible to determine whether you got the best possible outcome, especially in finance where literally everything is relative.

    Ultimately, market discipline only works up to the constraints imposed by information costs and some ripping off is inevitable.

  16. PS: Noah then asks "why don't the people getting ripped off exit the market".

    I'd say (1) there are exit costs and (2) the people that get ripped off are ultimately the investors in the mutual funds and pension funds that GS rips off.

    So the managers of those funds can pass on the "rip off" costs to those investors without suffering substantial declines in their salaries because the investors are even LESS aware of what goes on when you trade with the big i-banks.

  17. PSS: My last Finance PhD class was a while back but I remember an excellent paper that discussed the continued participation of uninformed investors in the markets with the argument that those investors know they are at an information disadvantage and therefore build in buffers by requiring greater returns from certain investments (so that the Sharpe ratio is ultimately optimal) and by betting on the long right tail of highly illiquid assets where most of the ripping off takes place.

    Sorry for the many, many posts.

  18. Phil Koop12:22 PM

    "Goldman's clients aren't really getting ripped off that much."

    That one won't fly, unless you are claiming that only Goldman clients are rational agents capable of learning. Otherwise, why don't the GS iBankers ever learn that they aren't really ripping anyone off?

  19. I think part of the answer is that the counterparties of GS are agents, not principals. A sophisticated hedge fund may not get tricked much by GS, but a pension fund, insurance company, or corporate treasury office might. The agents at those counterparties making the decision to deal w/ GS (and other Wall Street firms) may not even realize when they've been fooled, or for that matter care enough to follow up. There's little oversight of their performance, just procedure, and dealing w/ "reputable" counterparties like Goldman is seen as a safe career move. Goldman may actually take care of these agents personally (taking them to drinks, buying them dinner, getting them tickets to events), which functions as a nice kickback.

  20. Patrick said...

    " Even Warren Buffett used Goldman Sachs to get deals done. The fees are high but they've got a good network."

    Noah, I believe that you didn't mention this - network effects. There are a limited number of major players, and if you are playing, you'll have to deal with them.

  21. A "natural oligopoly" argument? Yeah, that does seem like another possibility.

  22. JW:

    My point is that we could just as well ask, Why don't the con artists convince the marks to stay in the market?

    So people never find out that Goldman tricked them?

    But then wouldn't Goldman's competitors come along and say "Hey clients, Goldman's ripping you off, switch to us instead!"?

    I think there still needs to be some explanation as to why this doesn't happen.

  23. Noah:

    There are two answers:

    Munger, The Psychology of Human Misjudgment---read his Harvard Speech

    and Daniel Kahneman’s Thinking, Fast and Slow (Munger on steroids)

    add lazy and friction and you will get the answer

  24. The clients are getting bragging rights and status. Just like for food and wine they are paying a higher markup because Goldman is a brand they can show off to their friends.

  25. What's missing from Mr. Smith's Op Ed is that he's really only talking about the sales and trading division of Goldman, even though he pretends he has insights into the entire firm. What's missing from the conventional analysis of Mr. Smith's Op Ed is the recognition that trading is a zero sum game. Goldman can only make money at the expense of its "clients" who are often on each side of a trade simultaneously. Meanwhile, Goldman's trading clients can only make money at the expense of Goldman or other market participants. Everyone, in fact, is trying to "rip everyone elses face off."

  26. Goldman Sachs did exactly the same stuff in the run-up to the Great Crash of 29, wherein they sold over quarter of a billion dollars in securities in less than a month, in '29 dollars! In his classic book, The Great Crash 1929, Galbraith, in the chapter "In Goldman Sachs we Trust" reports on this interesting interchange at a 1932 Congressional hearing.

    "Senator Couzens: Did Goldman Sachs and Company organize the Goldman Sachs Trading Company?

    Mr. Sachs: Yes sir.

    Senator Couzens: And it sold its stock to the public?

    Mr. Sachs: A portion of it. The firm invested originally in 10 percent of the entire issue for the sum of $10,000,000.

    Senator Couzens: And the other 90 percent was sold to the public?

    Mr. Sachs: Yes, sir.

    Senator Couzens: And at what price?

    Mr. Sachs: At $104. That is the old stock...the stock was split two for one.

    Senator Couzens: And what is the price of the stock now?

    Mr. Sachs: Approximately one and three quarters"

    The more things change....

    John H.

  27. Anonymous2:29 PM

    There's a certain amount of prestige in being ripped off by Goldman Sachs. Just like people like to brag about sending their kids to expensive private (but sometimes very un-elite) college, or investing in the latest Madoff fund, people/companies know they've made it when they're working with Goldman Sachs.

    A mafia analogy could also be appropriate. You're paying a little extra, but it could be better than the alternative.