Bookstaber: past performance is no guide to future risks

Veteran risk chief says trading gains in the wake of LTCM’s demise forged love of agent-based modelling

Richard Bookstaber

Rick Bookstaber’s CV reads like a Who’s Who of Wall Street, stuffed with spells at its most storied institutions and stints working with some of its most colourful personalities.

Bookstaber, now in his 70s and still looking to stretch the industry’s horizons on approaches to risk modelling, was the first formal head of risk at Morgan Stanley, where he had a ringside seat for Black Monday. Next, he became chief risk officer at Salomon Brothers, just after its more scandalous years, where, working alongside Jamie Dimon, he spotted an opportunity that yielded big gains in the wake of Long Term Capital Management’s (LTCM) demise.

He then moved to Bridgewater Associates, where he rubbed shoulders with fellow economic history enthusiast Ray Dalio, and experienced its famous ‘radical transparency‘ doctrine up close. Finally, he had a hand in curbing some of the finance industry’s more freewheeling instincts during a post-crisis stint with US regulators.

But despite – or perhaps because of – a lot of time spent managing risks at some of the industry’s largest firms, as well as a hand in crafting the Volcker Rule while at the US Treasury, Bookstaber has a deep suspicion of traditional approaches to risk management. He believes the financial industry needs to shake its reliance on quantitative methods, and experiment with more heterodox techniques, such as agent-based modelling and factor-based risk analysis.

His new firm, Fabric, aims to apply both techniques to the wealth management industry, a sector he describes as underserved, with the aim of advising firms on more effective transfer of idiosyncratic risks they face in their given market.

“Every investment adviser knows the mantra that past performance is not indicative of future results, right? They say that for returns – but nobody thinks about that being true for risk,” Bookstaber says.

Every investment adviser knows the mantra that past performance is not indicative of future results, right? They say that for returns – but nobody thinks about that being true for risk

Rick Bookstaber

He says that accusations of short-termism can be levelled at the risk management field as readily as any other area of finance, pointing to trading desks and market risk managers accustomed to deriving forward-looking risk measures – such as value-at-risk and expected shortfall – from the last few years’ worth of market volatility and returns data.

While Bookstaber isn’t arguing for classical risk metrics to be ignored, he argues that many firms focus on them too closely, and over too short a timescale, to the detriment of factors such as depth of market liquidity, leverage exposure and concentration risk. It has taken a paradigm shift in thinking from the industry post-crisis for risk to cease being a second-order concern, Bookstaber laments, something people “had to look at it for due diligence, and so on”.

He first became interested in agent-based modelling – an approach that Fabric has sought to commoditise – in the early 1990s, attending conferences at New Mexico’s Sante Fe Institute, a think-tank dedicated to the study of complexity. Agent-based modelling differs from traditional financial modelling in its scale. While banks and other financial companies will use arrays of various models for simulating the behaviour and performance of particular instruments and portfolios, agent-based modelling seeks to capture the interactions driving complex, rules-based systems, such as regional markets or entire sectors.

Biography – Rick Bookstaber

2020–present: Co-founder and head of risk, Fabric

2015–2020: Chief risk officer, University of California Office of the President

2009–2015: Research principal, US Department of the Treasury and senior policy adviser, US Securities and Exchange Commission

2007–2009: Risk manager, Bridgewater Associates

1998–2002: Head of risk management, Moore Capital Management

1994–1998: Chief risk officer, Salomon Brothers

1984–1994: Head of risk, Morgan Stanley

Working at Morgan Stanley at the time, Bookstaber – who had earned a PhD in economics from the Massachusetts Institute of Technology, and spent time teaching at Brigham Young University – recalls being struck by a conference at Sante Fe in 1993, where practitioners from fields including finance and physics discussed the use of agent-based models for understanding market complexity.

“They were developing what were called artificial stock markets. Right away, I realised this was the way to go,” he says.

A year later, Bookstaber moved to Salomon Brothers to work as head of risk management, where Dimon later became his direct boss. He began developing proprietary models employing the agent-based techniques during his time there – something that proved instructive in the wake of the collapse and bailout of hedge fund Long Term Capital Management in 1998.

“After LTCM, the markets were going crazy. We had really pulled back from big positions, thanks to Jamie Dimon and me, because we could see the risks that were growing,” Bookstaber says.

But roiling markets meant the bank was still left nursing a mark-to-market loss of $100 million on its Treasury desk, he says: “Nowadays, that’s not a lot of money for a trading operation, but at the time, it was enough to really matter to the desk and the fixed income division.”

It was here that Bookstaber spied an opportunity. With LTCM unwinding positions and selling off assets at any price it could, the spread between 19.5-year and 20-year bonds had become unreasonable – and unsustainable, Bookstaber argued. “We should be buying that position,” he told colleagues.

Down $100 million and with their head of risk management proposing an eminently risky strategy, his co-workers reacted with incredulity. But the gambit paid off handsomely for Salomon, he says.

“I knew the risks and the opportunities, and it was more than prudent to do,” Bookstaber says of the episode.

The fall of LTCM was spurred by defaults in the Russian debt market, a segment to which LTCM had no exposures of its own. The hedge fund was, however, highly leveraged in markets full of firms with Russian attachments – firms forced to deploy capital when the ruble collapsed. This contagion eventually spread to the Danish mortgage market, where LTCM had significant exposures.

“After that debacle, I realised which dynamics were essential and which could be modelled within an agent-based framework,” he says. “There’s a shock to the market: people who are highly leveraged have to sell. The selling, in an already-weak market, drives prices down even more, and you get this cascade dynamic. Then, the market gets to a point where it’s not liquid any more, and [participants] can’t sell what they want to sell. So, they sell whatever they can, and you get contagion – other markets getting embroiled – and a risk that spreads.”

“You can’t model that with standard approaches, but you can with agent-based modelling,” Bookstaber concludes.

Long-term vision

After stints as a risk management head at Moore Capital Management and investment firms Ziff Brothers and FrontPoint, Bookstaber landed at Bridgewater Associates in 2007. Working in an office near Dalio, Bookstaber found something of a fellow traveller in the firm’s billionaire founder. Both were history buffs with a shared conviction that understanding periods and events far outside of the usual market data intervals could provide valuable insights into modern phenomena.

“When I was [at Bridgewater], Ray was looking at the Weimar Republic. I thought, ‘Jeez!’” Bookstaber laughs. “But Bridgewater has a qualitative, long-term approach, a respect for long-term history, and what you can learn from knowing you’re in a market that has existed for decades and decades.”

This point brings Bookstaber to one of his key critiques of modern finance practice, which focuses on the centrality of quantitative modelling, a lack of attention paid to qualitative phenomena, and what he sees as overinvesting in data from the immediate past.

“People who are quantitatively oriented don’t like to think about the fact that we’re in a world of institutions, a world of central banks,” he says.

“One thing I do – which is kind of the anathema to someone with a quant background – is go through the markets from 1936 on, and ask the question: ‘What is it that leads to crises? That leads to major breaks in the market? How often do they occur, and how often does recovery take?’”

Richard Bookstaber
Rick Bookstaber practising Brazilian jiu-jitsu

Bookstaber left Bridgewater after two years, taking a government job in 2009. Working for the US Treasury and the Securities and Exchange Commission in the Obama administration, he helped draft the Volcker Rule, a key section in the post-crisis Dodd-Frank Act designed to prevent further bank collapses. Within the broader act, the rule – named after former Federal Reserve chairman Paul Volcker, and reviled by free marketeers – effectively functioned as a ban on short-term prop trading in certain asset classes, although the requirements have since been relaxed.

There was also a chance for Bookstaber to engage his passion for agent-based modelling by leading a project for the watchdog to model systemic instabilities and assess vulnerabilities in the financial system.

“If you understand market structure – if you know the credit conditions, levels of institutional leverage, where they are in terms of inventory, the broad stance of the Fed – you have the essential ingredients to understand how vulnerable the system is to a shock,” Bookstaber says.

“Material risks that can really matter to the market – not: ‘Oh, what’s the VaR right now?’ Things that can lead to prices going down 25 or 35 per cent, and possibly even propagate into a recession.”

With Fabric, he hopes to continue pushing such ideas and influence future generations of risk managers. The ideal risk manager is someone with eclectic tastes, he adds, and the capacity to understand a wide range of subjects.

“It’s hard to build a model that will forecast returns and say where prices are, but it’s another thing to say where risk might be given some scenario occurs,” he says.

“You have quant people that love computers and numbers, and you have people that look at markets and think long-term and out of the box. You don’t tend to find people doing all that together – and to really deal with risk management well, you need to have both.”

Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.

To access these options, along with all other subscription benefits, please contact info@risk.net or view our subscription options here: http://subscriptions.risk.net/subscribe

You are currently unable to copy this content. Please contact info@risk.net to find out more.

Most read articles loading...

You need to sign in to use this feature. If you don’t have a Risk.net account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here