Models, Morals, and Management in a Trading Room

Most investment banks, certainly the larger ones, have functions outside of investment banking itself, among them trading, securities research, asset management, and wealth management.  Each subgroup has its own structure, culture, and rules by which it interacts with the other parts of the organization.

This posting looks at a trading operation, by way of Daniel Beunza’s book, Taking the Floor.  Like Karen Ho’s analysis of the norms of corporate finance and M&A within investment banks (which was the topic of the last posting), Beunza’s evaluation developed over many years; the reader experiences his discoveries and changing conclusions as each layer of the onion is peeled away.

Trading is now substantially different than it was when Beunza was first exposed to the business in 1998.  It is much more automated (although that varies by asset class and security type), and — because of regulatory changes following the financial crisis — investment banks don’t dominate institutional trading flows to the extent that they once did.  Those changes are part of the story.

“New people, new practices, and new tools” altered social relations and communications within firms, and with clients and counterparties.  Types of trades that had worked in the past no longer did, while other opportunities opened up because of the evolution of derivatives and multi-asset strategies.  Deregulation and technological improvements fostered outsized risks before the crisis, then the Volcker Rule changed the game again after it.

“Models, Morals, and Management in a Wall Street Trading Room” is an apt subtitle for the book.  Those three factors are explored in depth.

A remarkable gap

Beunza was intrigued by what he saw as “a remarkable gap in the academic understanding of finance” regarding “the managerial and organizational aspect of Wall Street.”  He observed that his fellow academics fell into one of two camps:  “orthodox economists, who studied capital markets from the standpoint of rational choice” and “behavioral economists, who explored the ways in which decision-making biases impacted financial markets.”  Missing was the social aspect of markets, within and between the organizations that are the major actors within the ecosystem.

The first part of Beunza’s work involved interviewing traders and observing their work.  He longed for “the habituation and familiarity that arises from closer involvement,” so Bob (the head of the trading operation through whom he had arranged the interviews) gave him a pass, desk, computer, etc. for several months.  In subsequent years, Beunza would revisit the firm, which he called “International Securities” to hide its identity, and continue to interview the people involved, even after they had moved on.

Bob

Bob is the main protagonist in the book.  He was in many ways an anomaly, “a curious dude” whose approach to running a trading room differed from others on the Street.  Agreeing to have an academic embed within his staff was one example of Bob’s willingness to stand apart.

Beunza realized that the trading room at International Securities “was not representative of the average investment bank on Wall Street, but illustrative of what a possible solution to the challenges that plagued the financial industry might be.”  Bob sought ideas from outside the industry rather than just replicating what others were doing, and rethought established ways of structuring a trading room and managing traders.

At first, Beunza thought that Bob’s ideas gained traction through his moral authority as a role model.  Later on, Beunza learned that prior to the time of his research, Bob was more forceful in creating a number of rules that would he thought would lead to the culture that he sought.  By the time of Beunza’s period of involvement those rules were well entrenched, so he hadn’t realized the management decisions that led to the environment that he witnessed.  (That’s a good lesson for those trying to evaluate an organization.  The norms in place can be the product of evolution, but often there are direct actions at some point in years past that were instrumental in framing them.)

In many ways, Bob was trying to hold onto the ethos of the Wall Street partnerships that had disappeared in the 1990s.  One of the other interviewees reflected on those structures:  “The partners did two things that people don’t do these days:  they did their own risk analysis, and they criticized each other’s risk positions at the partners’ meetings.”  A former banker in the City of London characterized the impact:

In the old broking firms, career structure was straightforward.  The goal was partnership.  Once this was removed [in the move to a corporate form] status was dethroned and cash became king.

The structure had changed, the culture had changed, and risk positions were increasingly quantified and “managed” via models.

Models

Beunza’s research took place during a period when models (especially Value at Risk, commonly known as VaR) were increasingly used to summarize positions at the trader level and across the entire firm.  The problems with models and the increasing dependence upon them are a major theme of the book.

The “performativity” of models in the financial world make them different than models in the realm of the hard sciences.  Two examples, the portfolio work of Harry Markowitz and the Black-Scholes option pricing formula, were descriptive at first and then something more than that, as their adoption shaped how market participants (and markets) behaved.  The theories behind the models became self-fulfilling:

New quantitative representations had led to new practices and trading strategies.  In turn, these strategies often created a new financial product, and thus an opportunity for yet another representation.

As models become widely adopted, the returns that were available to pioneers in their initial use are “competed away” and new models are built on top of the old ones, “leading to an overlap in financial properties and interdependencies across models,” in what Beunza calls a “performative spiral.”  Those layers of belief in models are rarely challenged once they are entrenched.

But problems lurk.  According to Beunza, Bob was concerned that the use of models “creates an illusory simplification that is attractive but misleading.”  That can lead to “model-based moral disengagement,” where “the self-regulatory function of individuals ceases to apply.”  Beunza walks through how the historical relationships changed within firms and with outside parties once models overcame judgment as the governing factor of behavior across an organization.

In a column in the Journal of Portfolio Management many years ago, Peter Bernstein wrote:

Models are not immutable.  Indeed, the more immutable we believe them to be, the more useless they are likely to become.

Risk management

The risk management department of a firm is culturally different and usually separate from trading, with little political power relative to those who are viewed as being on the front lines making money.  One interviewee remarked that the function “should really be called risk measurement,” since it “was not effective in instilling restraint.”  Bob said:

One guy gets the rewards and the other gets the responsibility.  One guy is the hero, the other guy is the scold, right?  How the hell can that work?

Over time, VaR became the measure of risk that was the essence of communication between traders and risk analysts and firm management.  What was meant to be a management tool turned into a management substitute.  It allowed for “control at a distance,” so that a firm could:

grow and diversify into multiple markets and asset classes without appearing to lose its ability to quantify and aggregate the various risks it entered into.

A single metric was used to define risk, even though it did so in a narrow way that overlooked the widespread adoption of the model across the industry — and its shortcomings.  The tool that was supposed to manage risk completely missed the true risks that built up in individual firms and across the industry during the years leading up to the financial crisis.

Everything is seating charts

Over the years, Bloomberg columnist Matt Levine has titled a number of items “Everything is seating charts,” because issues of status and interaction in an organization are driven by who sits where.

When he came to International Securities, Bob created a trading room structure unlike others on the Street, and periodically he would shake up the layout of the different specialty desks in the trading room if he thought things were stale or that market relationships were evolving in ways that required new combinations of people.  For example, Bob began to see three properties of securities — fundamental, quantitative, and volatility — as driving strategies in disparate directions, and he thought that should feed into the organizational design.

Trading rooms are complex social entities, where visual cues and background noise provide clues, and proximity is a critical factor in whether and how people work together.  Without attempts at integration, traders tended to remain in their own silos.  According to Bob, at most firms, if you ventured into another area:

There were these cold looks.  Somebody would ask, “What do you want?” in a defensive tone.  People do not like you to watch them trade.

Those tendencies lead to a lack of information sharing, which Bob wanted to avoid.  So he organized the traders “to encourage collaboration and reduce status differences,” drawing from network theory to arrive at the configuration of the room.

Managing traders

A trading room is a community.  Bob described how he saw himself within it:

My role is as a cooling rod.  I walk the floor.  I talk to people about non-substantial issues.  I try to find out who is stressed.

But he needed to overrule traders’ decisions and force them to liquidate losing positions, a delicate balancing act that “created a complex relationship” with them.  He was also mindful that a trader’s intuition might be hindered by a forced analysis, that the initial implementation of a good idea “could be hampered by managerial supervision.”  Good managers need to read the room and protect their people from undue interference — but they also need to objectively judge their work and act accordingly.

Most traders are happy being traders, so they are not angling to move out of the trading room to another role.  That lack of interest in hierarchical advancement feeds into their focus on maximizing income, and that can be a source of much friction.  At other firms, traders often “experienced their outsized bonus as insufficient and unfair,” so Bob used a simple arithmetic approach for calculation, avoiding subjective issues and not cutting special deals with those who thought they deserved them.

Examples of trades and their effect on traders are included in the book, among them some merger arbitrage situations.  One trader offered this memorable comparison:  “The best predictor of your skill at merger arbitrage is your potential as a taxi dispatcher — the skill to do lots of things in parallel.

In general, complex trades are sought after, because that’s where bigger returns can be found.  But the theoretical limits to arbitrage become very real when you have a trade on that doesn’t move in the way that you think it should.

Periods of opportunity appear and disappear, so traders need to have the ability to take advantage of the openings that exist while restraining themselves when the odds are poor.  The standard models may not provide that warning, which is where the morals and management come in.

Universal concerns

Many of the topics addressed in the book apply beyond the confines of a trading floor:

~ While formal “speed bumps” can be used within an organization (or by regulators), morals are more effective as a means of social control.  Norms are powerful.

~ Bob:  “Money is not a good driver of cooperation,” especially when people are getting paid massive bonuses for their bottom-line contributions.  (This is an important issue across organizations — with beliefs on the topic that differ significantly based upon the type of entity and its culture.)

~ Beunza wrote of “the reactive and social nature of the trading room:  people resisted being acted upon, whether it was being exposed to the screams of others, seen by others to make a flawed argument, or being told to whom they should talk.”

~ The range of investor types and their strategies, resources, and concerns is staggering.  An investment bank sees flows from investors for a variety of reasons; understanding the motivations of the different players in the ecosystem is crucial to getting a sense of what is driving changes in the market.

~ No matter what kind of entity you seek to create — asset manager, advisory firm, etc. — you will mostly imitate what others have done before you, for better or worse, unless you purposely set out to build something that addresses the weaknesses of the standard approach.

~ Invariably, the best performers in a cycle are the ones who get destroyed when circumstances change in a meaningful way.

~ After a time trying to study International Securities, Beunza realized that his approach had been limited, because he was only dealing with traders and their managers.  That is similar to a key failing of many due diligence analyses, namely, a constriction of inputs.  Those deemed to be most important get too much attention when a representative sampling of people from throughout the environment under study will provide a more complete picture.

~ Bob:  “Communication is the central driver to adding value in business.”

Understanding an organization

As with the other sources in this series, Taking the Floor gives a detailed and valuable look at the workings of one part of the ecosystem.  An anthropological view can offer deep insight into why an organization does what it does, providing perspectives that can be missed by others.