The Information Dilemma

Today I spent a bunch of time at the Strata Conference in NYC catching some great speakers, speaking on a few panels and catching up with old friends. All in all, it was a productive and enjoyable day. 

However, one of the panels on which I participated, the Son of Money:Tech, was a revival of a terrific conference that was kicked off in 2008 but ultimately mothballed in the wake of the financial crisis. When the conference was held in 2008 it felt cool, integrating the discussion of what we now call the “Big Data value stack” into the realities of Wall Street, hedge funds and even consumer finance. It was the ultimate NYC conference. Fast forward to today. Our 45-minute panel discussion was just getting going when it sadly had to end, and we had just begun digging into some of the pressing issues of the day, e.g., historically high market correlations, the democratization of journalism coupled with the paradoxical importance of brand in an increasingly fragmented landscape, the massive barriers to entry in the low latency trading game, etc. It left both the audience and the discussion leaders wanting for more. In fact, I’d argue that a Money:Tech conference, circa 2012, would be far more thought-provoking and high-impact today than it was only three years ago. Cloud computing and storage is exponentially greater today than it was in 2008. Semantic technologies have continued to make large strides. Both the NoSQL movement as well as advances in relational databases have materially altered the face of web-scale and real-time analytics. And innovations in crowd sourcing and and the leveraging of contributory databases, together with machine learning, has further advanced the field of predictive analytics. 

Just musing about these topics in an unstructured manner during the panel made me think about how complicated the world has become. Making money is harder - and more costly - than ever before. Yes, more powerful open-source tools are available. Yes, the rise in open data has created analytical sandboxes the likes of which couldn’t be imagined even five years ago. Yes, the cloud has made massively scalable pay-as-you-go storage and processing accessible to even the earliest start-ups. However, as compute and data have become cheaper and more readily available, the data deluge has made it ever harder to extract tradable signal from the sea of content. Content that is structured and unstructured; in heterogeneous formats; with and without (affordable) historical archives; and assembled along different time scales. These are complex problems that often require teams of highly skilled data analysts and scientists to solve. The kinds of teams that are very expensive and available to only the very rich (think Bridgewater, Renaissance Technologies, Two Sigma, AQR, Citadel, Teza and Goldman Sachs). So does this mean that the mega-quants are so firmly entrenched that no one can hope to successfully compete against them? Well…

Hedge funds and Wall Street firms, like technology companies, can often become victims of their own success. It is very hard for the mega-quant fund to live on the razor’s edge printing hundreds of millions or billions of dollars of gains, year in, year out. There are always younger, scrappier, hungrier and more nimble competitors just looking for the smallest opportunity to earn their way into the club. It is hard to remain #1 in all aspects of what makes trading firms successful, and the most successful proactively disrupt themselves just as the most successful companies do (be they technology companies or widget makers). But it is a steep challenge. Then there is the issue of time scale. While there are very clear benefits of having vast resources to compete in the low latency end of the continuum (co-location, data teams, huge data budgets, etc.), as one moves towards the longer-term thematic long/short strategies the benefits of scale begin to melt away. Good research and good trade implementation, coupled with prudent risk management, can go a long way towards generating differentiated alpha. Co-location? Massive real-time data and historical data costs? Investment in real-time execution platforms? Not necessary. That said, these strategies are far more volatile and place more capital at risk than the low latency quant strategies. So as with anything, there is no free lunch.

These are just a few of the topics that surfaced during today’s panel. It was a pleasure kicking it around with my pals Rob Passarella (of Dow Jones) and Paul Kedrosky (of our Solar System), and I look forward to a redux in 2012. 

Tweet

The real-time web, a/k/a Wall Street 2.0

I have spent lots of time at the intersection of real-time data and technology. Like 20 years, in fact. Things that absolutely amazed me on Wall Street - real-time stream processing, massive databases of well-structured and indexed data, co-location - amaze me no more. The funny thing is, however, that much of the stuff that Wall Street figured out two decades ago is making its way into the real-time web. It’s kind of like “back to the future” - we had glimmers of the future back in the 1980s as Wall Street pioneered techniques and technologies that are now being applied to the ad markets as well as other real-time environments. TIBCO? Instinet? These are the companies that should be given much of the credit for showing us the possible. And the merely possible is now a reality.

Data is accumulating at such a rapid rate that batch processing for many functions is grossly inefficient if not prohibitively costly. The amount of big iron - or cloud capacity - needed to analyze and index a day’s worth of a decent sized ad exchange’s billions of impressions is mind-boggling. Nobody in their right mind would take this approach to data analysis. Instead, data is being analyzed on the fly with massive amounts kept in memory, eliminating the need for the scale of compute required to run terabytes of accumulated data though a batch process. Where did we learn this from? Wall Street. Why? Out of necessity. The main difference between then and now is that quants who run models against historical data don’t need to run them overnight, walk away, and come back 10 hours later to get the results. They can often get real-time or near-real time number crunching done on a massive scale by using the power of distributed computing, an approach that simply didn’t exist on a commercial basis back in the late 1980s and much of the 1990s.  

I see this trend being played out among the most sophisticated advertising technology companies as well as new entrants in the search space. Data is being treated as bits flowing across an exchange, both actionable in real-time while being indexed and stored for historical analysis and algorithmic development. At infinity, the data architectures and analytic frameworks of Wall Street and the real-time web will become one and the same. The requirements of each will be identical. But is important to remember our roots. Wall Street showed us the way. Make no mistake about that.

Tweet

How to rob a bank

This was one of those questions on Quora that caused me so sit back and ponder. But it didn’t take long before the best and easiest method of robbery popped into my mind (assuming staying out of jail is a priority): derivatives trading. Now I’m not saying that this is de-facto how derivatives traders operate; I’m simply saying that what is described below is a tried-and-true approach for extracting value for oneself without truly creating value for the firm and its shareholders. It has happened countless times throughout Wall Street and hedge fund history, and I’m sure it will happen again (it’s probably happening now).

Here is a straight-forward seven-step process for getting in the position to use derivatives trading to “rob the bank:”

1. Go to a fancy college, do well in math, CS and stats, and study those brain teasers that trading desks love to give applicants. If you didn’t go to a fancy college, work yours and your parents connections to try and secure a position. Absent connections, leverage your alumni network. Do whatever you need to do to get an interview.

2. Trade a little on your own, so you can say “I trade my own book” when asked, and have a trade idea in mind when you go for the interview. You will be asked this question.

3. Once you get the interview impress, show passion, intensity and how you are laser focused on making tons of money for your firm (and yourself). Take chances, be brash, stand out from the crowd.

4. Congratulations, you did it! Join the trading desk as a go-fer, indentured servant, whatever. Ask lots of questions without being annoying, and make sure you are studying the mechanics of how trades get done and how they get reflected in the firm’s risk systems.

5. Gain the confidence of others, build reputation, and position yourself for securing your own book. This can take a little time but be patient: It’s worth it.

6. When you get your book, really understand how P&L recognition works. If P&L is calculated on concepts like “inception gain” or “theoretical profit,” you’re in good shape. Also figure out the implied cost of capital. If your firm is charging you LIBOR flat, yippie! Another good fact. If it’s LIBOR plus a credit spread for the duration of the trade, less good but you can live with it if your P&L has the right revenue recognition characteristics.

7. What you’d now really like to do is enter into a large notional amount of longer-dated, short volatility trades. Book profit as theta decays, assuming you don’t blow up. if you do, well, your downside is zero and you can get a job at another firm for even having been in the position to take such massive risk. If you don’t, well hey now, all that time decay flows right to your bottom line. And with a large enough notional amount, your 8-12% of P&L can add up to some serious coin.

Did you create value? No. Did you trade well? If being lucky counts, then yes. And if you get paid and shortly thereafter the chickens come home to roost and your book blows up? Well, you’ve just robbed the bank!

Tweet

Insider Trading and Expert Networks - Efficiency vs. Fairness

Not a day goes by that we don’t read about the SEC’s deepening investigation into insider trading and its heightened scrutiny of expert networks. To some this is little more than a witch-hunt fueled by a chastened-yet-highly-motivated regulator; to others this is a boon to rooting out information asymmetry that has created an increasingly uneven playing field among investors. As always, the truth lies somewhere in between, but I believe to truly understand the issues one needs to look at the underlying principles: efficiency and fairness.

Many market theorists have argued that “insider trading” (e.g., acting on information that is not publicly available) enhances market efficiency, smooths price volatility and reduces the likelihood of price shocks arising from unexpected events. Where insider trading was once thought to be the province of people skulking around, listening to whispers from company executives, placing trades and delivering bags of money representing a share of the winnings, there is now an entire industry that has emerged to institutionalize the collection of hard-to-obtain information: expert networks. Note that I distinguish between hard-to-obtain information and material non-public information, thought the SEC is currently doing a deep-dive into whether such hard-to-obtain information has, in fact, crossed into the realm of material non-public information. So the “If it looks like a duck…” test has gotten much more complicated. Ivan Boesky and Dennis Levine - now that was insider trading and looked like it without much analysis required. However, are phone calls to professionals with domain expertise insider trading or simply gathering additional research towards validating or invalidating an investment thesis? Well…

From a theoretical perspective, I think it is hard to argue that insider trading doesn’t enhance the smooth functioning of the markets. More high-quality information is out in the marketplace being incorporated into stock prices. News of material events would leak out and be factored into trading activity prior to its release, limiting the potential shock that would arise were the information to be released into the market all at once. Therefore, stock prices would better reflect all the relevant information that is available, more closely representing the true intrinsic value of public companies. What is a boon for market efficiency, however, does not necessarily promote fairness. But that then begs the question - what exactly is fair?

As it relates to the research process, I think fairness should be judged by whether a motivated investor could learn the information through resourcefulness and hard work. If an investor was focused on better understanding the prospects for a particular product, could they read published research, explore primary sources and perform surveys? Sure. Would it be difficult? Absolutely. But it could be done if the individual was sufficiently motivated. Alternatively, they could obtain this information by paying money for it via an expert network. Is it “fair” that someone has the resources to pay for access to experts where another might not? I think so. Those who have built careers around investing and are willing and able to spend money to streamline their investment process are ok with me - provided that the information to which they have access could be obtained by a highly motivated person. What isn’t fair, however, is when experts provide information that couldn’t be obtained through any amount of hard work, e.g., if an official at the FDA was on an expert network panel and had non-public information about the likelihood of a drug making it through clinical trials, sharing this information with a subscriber would unquestionably be unfair. Another example would be a company executive handicapping the prospects of an M&A deal. In these circumstances the recipient of the information would have the ability to trade on it, capturing the benefit between the current stock price and the target stock price reflecting the information they have received. Does this promote efficient markets? Yes. Does it support fair markets? Clearly not.

I believe the current set of cases being brought by the SEC will create a further stratification of the expert network marketplace. The behemoth - Gerson Lehrman Group - will consolidate and enhance their position due to their strong compliance culture and the time they’ve invested with the SEC in figuring out the appropriate business model. Smaller players, however, will be driven out by the costs associated with building and supporting the necessary compliance infrastructure, almost like a Sarbanes-Oxley standard for the expert industry. This will create a brighter line between what constitutes hard-to-obtain information and material non-public information.

Market efficiency is an important goal to which we should always be aspiring, but not at the expense of fairness. But let’s be clear: fairness does not mean equal. It means equal opportunity. And the sharing of material non-public information does not permit equal opportunity.

Tweet

Bond Analytics: Taking an Open Souce Approach

Much has been written about what’s wrong with the rating agencies: structural conflicts-of-interest; intellectually over-matched; lacking in creativity and orthogonal thinking. Some believe the rating agency industry as we know it is on death row. To be honest, I tend to agree. Attempts at resuscitation are unlikely to yield the material changes required. The biggest problem: lack of transparency and insight into the multi-trillion dollar debt market. Much as the OTC derivatives market needs an overhaul, the opaque corners of the debt market need to move out of the shadows as well.

One of the unique aspects of the debt market is its mind-numbing diversity and dimensionality: maturity, amortization, optionality, collateral, seniority, etc. A “one size fits all” approach simply does not work for the bond market, and it is questionable as to whether a single entity has the intellectual horsepower and access to the resources necessary to effectively and efficiently analyze its range of securities. Large, seemingly intractable software problems have benefited from the massive collaboration available through the open source movement. This has been an effective method for not only addressing a core problem, but for keeping up-to-date and relevant as technology evolves. It has also been a vehicle for value-added service providers to build on top of these solutions (e.g. Red Hat/Linux, Lucid Imagination/Lucene, etc.) for specific use cases, providing needed service levels and documentation, etc. While not a panacea, the open source movement has effectively harnessed the world’s intellectual capital and applied it to big problems relevant to a broad array of constituencies.

If an open source approach has worked so well in software, why not apply it to the ratings problem? Whether or not ratings should be required for institutional investors to buy certain securities is not the issue; the essential point is getting better transparency into and analysis of instruments constituting the investable universe. Imagine a university or a large institutional investor seeding the open source initiative by putting their own debt ratings models into the public domain and allowing others to contribute to its development. I can see a suite of open source libraries by type of instrument, with a new industry emerging to deliver additional analytics, data and recommendations on top of these libraries. There would need to be a Wikipedia-type board of curators, ensuring that additions to the libraries are sensible and increase the stock of intellectual capital. But I can’t see why such an approach wouldn’t address the biggest problems facing the ratings industry today.

Combining bond analysis and the open source movement could deliver:

• Transparency;
• Unbiased input;
• Access to a global talent pool;
• Opportunities for specialized applications to be delivered in tandem; and
• Institutional-grade analytics and research available to all.

I haven’t seen or heard of a better solution to the problem, and the problem certainly isn’t going away. If we as a financial community are committed to such an approach, it is bound to be successful. Let’s give it a shot.

Tweet