Future Research in Finance

Please read Mark Thoma’s piece on Capital Market Theory after the Efficient Market Hypothesis. It’s an interesting paper and Thoma’s own observations add to it well.

Dimitri Vayanos and Paul Woolley argue that we do not need to abandon the assumption of rational expectations in order to better understand how asset markets function, but we do need to incorporate principal agent problems into asset market models.

Setting aside whether agents are rational, I am fully sold on the idea that we need to do a better job of incorporating principal agent problems into these models…

The story below is one where asset price “momentum” is generated from principal agent issues driven by incomplete information on the quality of investment fund managers…

I’m very interested in the study of capital markets post-crash. I need to do a mind dump on this topic, so I’ll add some wide-ranging thoughts, that will increase in zaniness until I fall asleep on the publish button.


– We’ve had credible rational agent theories explaining momentum going back at least 20 years with DeLong, Shleifer, Summers, and Waldmann (1990), “Positive-Feedback Investment Strategies and Destabilizing Rational Speculation.”

– Principal-agent modeling is going to get weird quickly. Take hedge funds: They are highly secretive, use opaque strategies, have no legal disclosure requirements, and have giant windows in how you can withdraw your money from the fund. Yet people line up around the block to get their money into them. Is that a principle-agent problem? What is the deal with those who want to invest – are they suckers? Or brilliant? Would Lakonishok’s work with window-dressing and mutual funds hold up?

– The lowest hanging fruit for the information crowd is uncertainty over the risk-free asset, which certainly caused many problems in the previous crisis. See the Lucas quote here.

– I hope people can bridge the gap between macro and finance. The Business College-Liberal Arts College gap doesn’t help. I’m sure there are money differences too. I remember how shocked I was when I heard a macro person sneer that finance theory, Black-Scholes, etc. was “a partial equilibrium thing” (as if it was the market for pizza!) – I thought we’d get along like PB&J. So naive back then….

Also macro people are the prima donnas of every department, in case you weren’t getting that sense from the internet discussions (how could you not?). Which makes finance people want to deal with them even less.

Efficient Markets

– I think the liability for formal economics to tackle the capital markets problem is that they assume that profit-maximizing transactions leads toward an equilibrium. Particularly an equilibrium that has some fairly beneficial properties to them. This is almost certainly true for “the market for pizza”, but it’s an assumption that may have to be weakened at least during initial investigations. Notice some quotes from Summers getting it in the New Yorker profile:

Summers told me that, as a graduate student, he first studied claims, made famous by economists at the University of Chicago, that financial markets are always rational and self-correcting. He said, “I encountered a sentence that was much quoted: ‘The efficient-market hypothesis is the best established fact in social sciences.’ Any sentence like that is a red flag to an ambitious academic.”….

Since his days tearing apart the E.M.H., Summers had had a healthy academic understanding that markets under severe stress could be thrown so out of whack that their self-correcting mechanisms broke down and required outside intervention….

Summers returned to his attacks on the efficient-market hypothesis. While it is true that the market is often self-correcting…at other times…the shocks to the financial system are so pronounced that the market does not self-correct, and eventually tips the economy into recession. Summers explained that although the relationship between supply and demand is the cornerstone of a self-correcting market, in some financial crises that relationship breaks down. “If you think about a security that is bought on margin,” Summers explained during a speech early last year, “when its price goes down there are margin calls which force liquidations, and more of it is sold. So a falling price is not a stabilizing mechanism, but it is potentially a destabilizing mechanism.”

Getting a theoretical and empirical handle on that would be a great start.

– Methods wise modeling the way things fall apart would best be done, as far as I can see, by getting one’s hands dirty with stochastic calculus instead of models with dynamic programming (which, as smart people have pointed out, is really a one-period model disguising itself as an infinite period model.) Maybe not, but probably so.

– The best theoretical tool still standing is something like Andrew Lo’s Adaptive Market Hypthoesis. So much of what constitutes meaning and form in financial markets changes the moment it’s observed. It almost has quantum properties that way – observing it changes the nature of it. Beta worked really well, until everyone modeled their risk on it. There was a January effect, until everyone learned about it. Etc. Financial markets are always cannibalizing the very terms that are trying to describe it.

Other Fields

– Economists should be willing sit back and let the sociologists, ethnographers and anthropologists get a solid crack at this. It’s telling that the two new books I find the most informative about the crisis are Fool’s Gold, written by someone with a PhD in social anthropology, and The Myth of the Rational Market, a history of ideas which I read as very similar to the sociologist MacKenzie’s An Engine Not a Camera, but without as much performance theory (and focused on asset markets as opposed to derivative markets).

– I think this is true because there’s a big performativity component to financial markets that Lo’s Model hints at, but also because I think the mathematical formalism may be too strangling in the first passes. For instance, if you held Minsky’s (who is doing a lot of work for a lot of people during this crisis) fantastic Financial Instability Hypothesis, the actual paper, next to, say, Polyani and/or a systems theorist, and then held it against Duffie’s Dynamic Asset Pricing Theory financial economics textbook, it would obviously go into the first pile. Could Minsky get tenure nowadays? Math-ing up Minsky’s ideas is a very good goal, but first we need the ideas fleshed out. And that requires getting the ideas.

– Speaking of ideas: let’s imagine an economist asks a feminist why women have so much more control over their sexual and economic lives nowadays. You can imagine the feminist discussing glass ceilings, “The Second Sex”, Judith Butler, and the economist cutting her off saying: “aww, you’re sweet. But really the technology process gave you The Pill, and there was a move from manufacturing to service industries, and that’s it. All your consciousness raising has done nothing.” I think that’s a fair stereotype. Given that, why are we sitting around listening to economists navel-gaze about their methods so much? I feel like so much of the debate around macro has been completely cut-off from the evolving nature of capital markets. You’d think economists would be the first to see a debate over whether or not Euler equations belong in a paper as being the kind of academic fashion that distracts us from the technology processes that are really determining things.

Now feminist theory obviously doesn’t set the interest rate (insert stimulus joke here). And I do believe ideas matter. Especially economic ones in a neoliberal age. But they matter because they help us construe a world that is simultaneously construing us. Why aren’t we forcing economists to look at the actual institutions and capital market structures and see whether or not their ideas are matching up to them anymore? Krugman’s big article on Dark Age macro made no references to anything outside academia.

Here I really like Perry Mehrling’s short 2005 The Development of Macroeconomics and the Revolution in Finance (“In fact, as I shall argue, neoKeynesian macroeconomics circa 1965 was destabilized not by the various internal theoretical problems that standard pedagogy emphasizes, but rather by fundamental changes in the institutional structure of the world that neoKeynesian macroeconomics had been developed to explain. We see this fundamental institutional change most clearly when we focus our attention on the financial sector”), and it comes with the highest recommendation. It’s all about placing the debates in macroeconomics against the actual capital markets, and in particular the question of who is setting the risk-free rate: the market or the government? I would love to see more work done in this vein.


– A good starting place, but I am not sure if economics has the tools to tackle it right now, is what does “the financialization of the economy” mean? This is what I’m dedicating my study to, and I’m willing to go to just about any source, and so far the answers have been weak. What does it mean that the share of business profits going to the finance sector has more than doubled from 15% to 35%?

The knee-jerk economics answer is “that’s no different than if the pizza sector was 35% of the economy and what are you going to do go around telling people how much pizza to eat?” The sophisticated economics answer is that it must be the result of consumption smoothing and business cycle smoothing and that this sector is growing because consumers are enjoying their services. Maybe that’s unfair, but I don’t see much else.

Is it the sign that an economy has matured, and the finance sector is just trying to eek out some extra growth? Does it have implications for IO or labor contacts? Is this a new form of global hegemony? Is it a byproduct of The Great Moderation? Is it just rent-seeking on the productive sectors? And most importantly, does this sector being this big increase the chances of financial disasters? I do not know, and I’m hoping to learn more in the future.

This entry was posted in Uncategorized. Bookmark the permalink.

11 Responses to Future Research in Finance

  1. Jean says:

    On explaining size and change in size of the financial sector, Philippon (2008) has some interesting insights. Instead of adopting a household/consumer behaviour perspective (to start with, the consumption of or the production of financial services is not a straightforward concept), he relates:
    – the evolution of the size of the sector to corporate finance developments (how best to finance development opportunities? internal/external and direct/indirect finance); and
    – the demand for intermediation to the pattern of industrial/technological developments (what is the typical firm with development opportunities?).

    The existence of/ability to capture rents is an additional layer worth exploring in relation with industrial organisation/corporate governance/wage developments in the sector. On the latter, I’m sure Philippon Reshef (2008) has no secret for you.

    Philippon (2008) The Evolution of the US Financial Industry from 1860 to 2007, mimeo NYU Stern

    Philippon Reshef (2008) Wages and Human Capital in the U.S. Financial Industry: 1909-2006, revise and resubmit to Journal of Finance

  2. Mike says:

    Awesome – I knew there was a literature I was missing, and I shouldn’t have been so dismissive. I’ll be checking out Philippon’s work. Thanks!

  3. Nicholas Mycroft says:

    In re “mathing up” Minsky, Steve Keen has done some interesting work.


  4. anon says:

    re: Hedge funds

    Seems likely to me that, a bit like the Madoff case, many investors may have assumed that the fact that regulators had no information about them meant that they could break laws with impunity — and they wanted in on these extra-normal profits. In other words the fact that hedge funds are opaque and unregulated is a marketable characteristic.

    Not unrelated to the fact that the DTCC seems to think that it’s just fine for the SEC to have to ask on a one at a time basis for CDS trade data — instead of just being given access to the whole database in real time: http://ftalphaville.ft.com/blog/2009/10/08/76531/sec-wishlist-derivatives-data-with-which-to-pursue-derivatives-fraud/

  5. Jason Ontko says:

    I have done some research on this and my answer is that the relative size of the financial sector in an economy grows as the division of labor and technology increases. For instance since education is more specialized with increased division of labor, you must go to specialized schools that are farther away from home and more expensive then liberal-arts schools, this needs to be financed. You then will move to a job center that has your specialty that is located only in a few places in the world spaced far apart from one another, that needs to be financed. Since your specialized job is more easily eliminated from changes in technology, you will have to retrain in another specialty and move to different job centers, this will need to be financed. Also businesses with increasing scale and specialization will also require more borrowing to produce their complex products in a global economy. Even small businesses now must utilize the global economy. I will even give up my sister as an anecdotal example. She runs a 14 person company that makes signs for supermarkets. She gets most of her sign making materials from Denmark. Since her market is very price sensitive and she enters into long term contracts, she needs to hedge exchange rate risk otherwise she could end up selling signs for less than what it costs to produce them. That hedging would not have been necessary in a simpler time when signs were made from a piece of paper and a marker.

    This is not to say the financial sector can not over grow itself as it recently did, it is saying that there is no optimal fixed relative size of the financial sector to the economy and that we should see stochastic growth in the relative size of the financial sector as economies become more specialized. What is the optimal size; is the main question and all the research papers I have read on the subject either don’t talk of optimal size or fail utterly to solve the problem. It is a hard thing to do quantitatively on the country level because you have global financial centers that serve more than their own economy. That will be the first hurdle you will have to overcome.

    Best of luck


  6. Not the Mike You're Looking For says:

    I really like Jason’s account of financialization and find it quite compelling.

    But we need to keep separate two distinct issues: (1) finance’s share of the economy; and (2) finance’s share of corporate profits. The former fits well into Jason’s story, but the latter is an entirely different animal. It suggests one of three possibilities: (a) monopoly power; (b) rent seeking; (c) genuine innovation (in its early stages). Otherwise, why aren’t these profits being competed away?

    What’s interesting is that, if I remember the Philippon and Reshef paper correctly, the authors connect increased profits to deregulation. Normally monopoly power and rent seeking are associated with the opposite. Yet I find it hard to believe that genuine innovation is driving financial profits.

  7. Mike says:

    Jason/Mike – awesome. Thank you for starting the (long) process of breaking this down into answerable questions.

    I’m also interested in how businesses became more “financialized.” GM was basically an auto loan company towards the end. One of the larger subprime lenders were subsidiaries of GE Capital, who bought up some of the bigger name mid-90s subprime lenders. GE Capital was made so mom-and-pop stores can finance having refrigerators, why (and how) did it change?

  8. Louis says:

    I don’t really know anything about finance, but from what I understand of Minsky’s ideas, there’s a kind of information theoretic quality to it. Minsky is essentially saying that after a while market transactions tell you nothing about the demand for the asset being bought and sold and only that there’s a high demand for loans to speculate on the asset.

    Waving my hands, it seems like the “Minsky Moment” would drop out of a finite information model, but that such a model would be pretty exotic in finance.

  9. jult52 says:

    Re: Not the Mike’s “(2) finance’s share of corporate profits.” please consider a technical alternative:

    US & int’l GAAP/IFRS rules are very flawed and do not properly reflect a financial firm’s economic performance as well as nonfinancial firms. That’s the key, I believe.

    Now cash flow (#1) depends on large financial firms floating large issues of debt constantly. How do lenders of money to financial firms determine credit risk? A crosscheck with cash production like it exists in nonfinancial firms doesn’t exist, so they rely on GAAP/IFRS. Which is flawed. It comes down to accounting rules masking the fact that the financial services industry ISN’T particularly profitable.

  10. Peter says:

    Mike – on financialization, a good piece is the sociologist Greta Krippner’s “The Financialization of the American economy,” in 2005 Socio-Economic Review. Good stuff.

  11. gappy says:

    Mike, nice post with a great breadth of ideas. I wanted to point out something however. When you say:

    Methods wise modeling the way things fall apart would best be done, as far as I can see, by getting one’s hands dirty with stochastic calculus instead of models with dynamic programming (which, as smart people have pointed out, is really a one-period model disguising itself as an infinite period model.) Maybe not, but probably so.

    I should point out that Dynamic Programming is not at all a one-period model. That’s what undergraduates exposed to infinite-horizon DP think. In practice, there are stopping-time problems (e.g. pricing American options), finite-horizon problems etc. Case in point: the Noise Trading literature starting with DeLong et al. (1988) and culminating with Shleifer-Vishny (1997) is using finite-stage DP (2-3 periods). And of course you can have DP in continuous time and a Stochastic Calculus framework as well. I don’t see a dichotomy between Stochastic Calculus and DP, but rather between discrete- and continuous-time formalism.

    I would like to add a couple of other observations. In order to discuss directions of research, we should define a clear testing protocol for EMH violation. Only then we can discuss about stylized facts regarding the violation, at least beside the uncontroversial momentum effect. Since the Adaptive EMH is really a tautology, we should furthermore situate violations in their historical context, since they tend to disappear after they are discovered (this point has been made by many, and among those I recommend Doron Avramov). In terms of approaches, surely the principal-agent framework is a favorite of economists, but I am skeptical it will have much explanatory power. In fact, I don’t know of any actual numerical successful prediction conducted using principal-agent. It’s all storytelling. Correct me if I am wrong.

    Another way to look at the problem is relax rationality and use something like cellular automata. Such parsimonious models have been useful in modeling complex physical phenomena (e.g., self-organized criticality). So far, research has been scattered and not really cross-disciplinary, because many physicists think they can colonize other fields ab initio. But I think it would be a promising direction.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s