PsyFi Search

Wednesday 17 November 2010

Arbitraging Embeddedness

Society, Sublimely

Although the British Prime Minister Margaret Thatcher once opined that “there is no such thing as society”, society generally begs to disagree. Indeed British society, enraged by her insistence on an unrepresentative and unfair tax, combined to help force her from power. In reality we’re all sublimely swimming along in a swirling pool of social relationships, which drag us unwittingly in various directions, while we blithely assume we’re actually in charge of our lives.

This is what sociologists study and economic sociologists, as you might expect, are particularly interested in how these social forces drive financial markets. Self-interest being what it is, it’s perhaps not surprising that they suggest that the behaviour of markets is less driven by individual behavioural biases and more by social factors. What’s more, they may even be right.

Economic Embeddedness

Economic sociology has a long history, but its recent development can be dated to a paper by Mark Granovetter entitled Economic Action and Social Structure: The Problem of Embeddedness, which outlined the idea that economic relationships between individuals and corporations are embedded in social networks. In itself this may not seem too exciting, but he went on to speculate that these types of relationships simply don’t appear in the standard economic models, which could have been specifically designed to eliminate them.

Economic models go about eradicating any trace of social relationships in a number of ways but the assumption of laissez-faire competition leading to efficient markets is one of them: because if markets are driven in this way then the issues of social relationships between the various actors simply goes away. As Granovetter states:
“In a much quoted line, Adam Smith complained that “people of the same trade seldom meet together, even for merriment and diversion, but the conversation ends in a conspiracy against the public, or in some contrivance to raise prices. His laissez-faire politics allowed few solutions to this problem …”.
Regulating Social Behaviour

So, the idea is that the introduction of insider dealing rules, anti-cartel legislation and the like is designed to remove the possibility of humans actually interacting. Thus, potentially, much of the superstructure of modern economics is an attempt to argue away the existence of social relations – since, if these really have an impact on economic behaviour, then the existing models don’t have a way of dealing with them. Best to ignore the issue or legislate it away, then.

This is as true of behavioural economics as of the more traditional kind – as we saw in Behavioral Finance’s Smoking Gun – because this is founded on the same principles as standard economic models. Simply adding on the notion that behavioural biases are caused by the limitations of the brain's ability to process information doesn't change anything very much.


In contrast embeddedness sees economic behaviour – such as trust or malfeasance – as being based around social relationships: we tend to trust people we’ve happily dealt with previously and avoid those who chiselled us. Only this may not happen in the real-world – maybe we have no choice but to use an untrustworthy supplier, or maybe we’re just the sort of person that continually ends up in abusive relationships.

Basically embeddedness argues that it isn’t guaranteed that the market will self-regulate and exhibit efficiency – it could equally well end up establishing social relationships that encourage perverse and deceitful behaviour: there’s nothing deterministic about it. This, if you’re interested in developing vaguely scientific models, is a nightmare, because without determinism it’s darn difficult to make predictions.

Models Endeth Relationships

However, there’s another strand of argument that can be used against the idea that social relationships drive markets. Simply put, most modern institutions are driven by computer models, not personal relationships. People interact with terminals, not other people. In this automated, inhuman world there is, surely, no place for embeddedness?

Well, Daniel Beunza and David Stark reckon differently. Firstly, they argue that not only are social relationships behind much market behaviour but that it’s also a much more likely explanation for extreme market events – Black Swans – than the ideas of behavioural economics. In Models, Reflexivity and Systemic Risk they suggest that Nassim Taleb is wrong:
“The Black Swan argument presents financial actors as hopelessly unreflexive about the limitations of their models. According to it, traders either ignore what every finance academic already knows – namely, that extreme events happen – or lack the reflexive capacity to act on it. Confronted by this argument, we ask, why should we deny to financial actors the capacity for reflexivity that we prize and praise in our own profession?”
Which rather opens up the obvious question of, if they’re not caused by behavioral biases, what is causing them? To answer this the researchers draw on an ethnographic study of a merger arbitrage trading desk. A what of what, you might ask?

Ethnographic Arbitrage

Ethnographic studies are most commonly associated with anthropology, for the good reason that it's quite difficult to transplant whole communities into a laboratory and start running scientifically controlled experiments on them while maintaining any pretence at ecological validity. They're carefully documented, real world studies of individual, but hopefully typical, situations. You might more commonly have come across them in the form of a case study.

The study here centred on a group of traders who attempt to make money by exploiting pricing differentials in corporate takeover situations. By analysing the probability of success or failure of any given merger they take positions with the aim of exploiting their insights. This, of course, is a game played by lots of different teams from lots of different companies and there are winners and losers. Interestingly, though, sometimes everyone loses.

Isolated Yet Social

Even though these traders are essentially operating in isolation from their peers they are, in fact, embedded in a complex set of social relationships because they’re continually analysing the market data to determine what their competitors are doing. It's this reflexive approach to the task at hand – continually questioning their beliefs – that embeds them in their social network.

The trouble being that all this reflexivity, as the traders test each other’s trading strategies to check their own, sometimes brings disaster. A case in point was the GE-Honeywell deal in 2001 which collapsed when the European Commission regulator blocked the deal. This cost assorted traders over $2.8 billion and was completely unforeseen by the industry, despite its attempts to check and cross-check every possibility.


So, obviously, the actions of the European regulator were completely unforeseen and unforeseeable – a veritable Black Swan? Well, as it happens, not really: the media was all over the possibility yet the carefully reflexive arbitrage traders chose to ignore the risk. In fact, worse, they increased their positions when they crosschecked what everyone else was doing and ended up with unwarranted confidence in their positions:
“Each fund erroneously takes the others’ lack of visible concern … as reassurance that the merger will be completed. The added confidence leads each fund to increase its position, compounding the losses when the merger is canceled.”
The researchers here argue that traders specifically set out to look for cognitive dissonance – indications that their ideas are wrong. Occasionally, though, if enough people miss an important point this leads not to dissonance but to resonance, triggered through the use of common modelling tools and leading to “cognitive interdependence”.

This is the same sort of problem we see at play in groups of investors in tight or extended social networks. Think share clubs or bulletin boards or Twitter or any forum that allows people to interact: merger arbitrage traders work damn hard to eliminate these problems, and still they get it wrong on occasion. Most other people don't even recognise that these issues exist

Rationally Wrong

That social networks play a part in financial behaviour shouldn’t really be a surprise. The difference here is that the individual actors aren’t transmitting irrational behavioral biases and causing a cumulative cascade of aggregated idiocy, but they’re attempting to act pretty much rationally and making a mistake. In certain situations this can get amplified into something that looks like gross stupidity, but is more a facet of the way the particular social relationships that connect individual actors operate.

This is an explanation that neither relies on the inadequacy of quantitative modelling nor the dominance of behavioral biases in market participants to explain the apparently mad behaviour of markets. It doesn’t deny the reality of either of these two causes either, but connects them through social interaction to offer an approach to systemic risk that transcends both. What's more, it affects us all, because none of us exists independent of the rest of society.

In our modern world for the most part nuclear power stations don’t spontaneously explode and aircraft don't fly themselves into the ground. Yet financial markets and groups of investors do both with alarming regularity. If social interaction is behind some of this behaviour we're not going to solve that problem until we start to build it into our risk models – and to do that, we've got to start by acknowledging the issue. Otherwise we're likely to get the same results Margaret Thatcher did.

Related articles: On Incentives, Agency and Aqueducts, Behavioral Finance's Smoking Gun, Black Swans, Tsunamis and Cardiac Arrests


  1. For a perfect example, Bernanke’s biggest flaw is that he thinks this is only a credit contraction recession, so he can fix it with credit expansion tools. He doesn’t realize that he is up against (1) a generational shift to the Gen Xers who got buried in student loans, and view credit very differently from the Baby Boomers, (2) the retirement of said Boomers, the pig in the python with the largest cohort turning 48 ( the year of peak consumption per Harry Dent) in 2005 – did something change after that? And (3) technology, which slashes prices every year (OMG – deflation!).
    No one around him will tell him he's using the wrong tools - his social set, such as it is - are basically yes-men. the book launched yesterday – Bernanke isn’t going to change – given his history, he can’t – and Obama doesn’t have any idea how awful his appointments were. It's inflation coming, not deflation. Don't ever bet against the Fed.

  2. If social interaction is behind some of this behaviour we're not going to solve that problem until we start to build it into our risk models – and to do that, we've got to start by acknowledging the issue.

    Here's my optimistic take in two points:

    1) If we have no choice but to soon begin acknowledging the issue (I believe that, if we do not, we will soon see our free market economy collapse), then we WILL soon acknowledge it; and

    2) If we acknowledge it, we will take the biggest step forward in our understanding of how investing works that has yet been seen in the history of markets.