As many of you are students, you already understand the importance of investing. You or your parents likely invested hard-earned dollars to pay for your tuition. At the SEC, our job is to see that when you begin investing on your own, you have a fair chance of having something to show for it at the end of the day.
Our job is to minimize risks that aren't a natural outgrowth of a smoothly functioning financial system. We try to keep dishonest financial advisers from stealing your money through fraud or illegal fees. We try to make sure the markets function properly, to minimize operational issues, so you don't lose money because a broker-dealer's trading platform broke down. And we try to ensure that you have timely and accurate information when you make your investment decisions. In other words, we work every day to have your back.
INVESTMENT IN TECHNOLOGY
And that requires an investment of our own. Although we by no means have all the resources we need, we're making an unprecedented investment in the cutting-edge technology we need to protect investors in today's blindingly fast and extraordinarily complex markets.
These new tools are giving us the ability to examine the way our markets function with greater precision and detail than we have ever seen. They allow us to plumb routine filings for irregularities that might signal illegal acts or suspect accounting. And they give investors — who must always be the first line of defense against fraud — a greater ability to chart a secure and informed course through a complex financial environment.
We are using this technology not just to make us stronger but to help make investors and other market participants smarter.
You are tomorrow's investors. So I would like to spend my time here talking about how we're using tomorrow's technology today.
We have a big job. About 4,000 SEC employees serve as advocates for 100 million individual investors. We oversee more than 35,000 registered entities, including publicly traded companies, national securities exchanges, investment advisers and broker-dealers, hedge funds, ratings agencies, mutual funds and clearing houses.
And an understaffed and outspent SEC simply can't be everywhere at once. But the technology we're bringing online helps by allowing us to leverage the talents and commitment of our staff to cover more ground and target our efforts far more precisely than would have been possible just a few years ago.
BETTER DATA ANALYSIS
I'm sure we all agree that a better understanding of the downpour of data generated by the markets every hour will lead to better regulation and better investor protection. Last year, the SEC put in motion two initiatives that will dramatically increase the quality and quantity of the data we receive, and improve our understanding of the way today's markets function.
The first of them is called MIDAS.
MIDAS stands for Market Information Data Analytics System. It captures all orders posted on the national exchanges, all modifications and cancellations of those orders, all trade execution of those orders and all off-exchange executions.
The result is an unprecedented aggregation of trading information data, one so comprehensive and arcane that I'll spare you a detailed discussion.
But let me say this: For the growing team of quant types now employed at the SEC, MIDAS is becoming the world's greatest data sandbox. And the staff is planning to use it to make the SEC a leader in its use of market data.
MIDAS has immediate applications. It can help us monitor and understand mini flash crashes, or pick up on possibly troublesome or illegal behavior, for example, by noting excessive cancellations of message traffic. But what's critical in the context of long-term investor protection is that it will give us dramatically better insight into the function of a market that moves many millions of dollars in millionths of a second. It will be like the first time scientists used high-speed photography and strobe lighting to see how a hummingbird's wings actually move.
This information has the capacity to give regulators — as well as academics and other stakeholders — unprecedented insight into the way markets work today. And it will make us better able to address concerns about market structure, monitor trends over time and provide information about the effects of any changes in rules, policies or market practices.
Analyses like these could be game changers.
MINING DATA FOR PUBLIC USE
But I see no reason for the SEC to monopolize the information. So I have asked the staff to explore ways to release the information that this data will reveal to the public.
Among other things, the staff likely will explore such subjects as:
The speed of quotes and subsequent cancellations. In today's market, where fractions of a second can be critical, this could give us a clearer picture of the potential effects of rules like those requiring quotes to have a minimum time in force.
The full depth of book for liquid as well as illiquid stocks. This could help us understand how overall market depth is affected, as top-of-book spreads narrow or widen, and the potential effects of a tick-size pilot and other rule changes.
Intraday volatility. Analysis of volatility within a trading day could allow us to better understand anomalous trading and extraordinary intraday spikes, as well as monitor the impact of rules that limit stock volatility, once fully implemented.
In short, MIDAS will enable us to examine the fundamental mechanics of today's high-speed markets in a way that has never been done at the agency.
One of the things that fascinates me about this is that people have been arguing for years — based almost entirely on instinct and intuition — about things like time-in-force rules or the ultimate impact of high-frequency trading on liquidity. But until now, no one has had clear data on these subjects. Not only did we not know if the solutions being proposed would work, we couldn't really determine the extent of the problems.
Now we will. And whatever the data reveals, the SEC will have substantially deeper understanding of where and how we should and can act to make the markets more stable for investors.
CONSOLIDATED AUDIT TRAIL
Important as it is, though, MIDAS is just the beginning.
Last year, the commission voted to require the national securities exchanges and other self-regulating financial organizations to establish a marketwide consolidated audit trail, a system to significantly enhance the SEC's ability to monitor and analyze trading activity.
Where MIDAS collects vast quantities of public data, CAT will capture nonpublic data, as well — not just trades and when they were executed but also, for example, the identities of the parties to the trades.
CAT has the potential to expand our ability to analyze market structure and behavior even further than MIDAS, while improving our ability to reconstruct and pinpoint the source of trading irregularities, and investigate suspicious activity.
I can't overestimate the importance of CAT. Comprehensive public and nonpublic data about the market — coming from a single system — could be the most important regulatory development in my lifetime. But we need to do it right. As I have said before, it is my personal hope and expectation that the plan submitted for the agency's consideration ultimately will produce a consolidated audit trail that is more timely, accurate, complete and accessible, and one that will provide for the eventual expansion to additional instruments such as OTC equities, fixed income and futures. I believe this is an extraordinary initiative and — if properly constructed — the consolidated audit trail will be the backbone for our future surveillance and policymaking efforts. And it will set the stage for expansion to additional instruments and markets, as well. My long-term vision is a consolidated audit trail that spans products, markets and the globe.
Regardless of our ability to analyze data and act on our findings, if the market structure is not stable, our other investor protection efforts are much less likely to succeed.
We saw how automated markets and computer-driven trading can go awry when technical issues in Knight Capital [Group Inc.]'s trading and routing software caused it to establish positions erroneously in nearly 150 stocks, ultimately costing the company $440 million. And we saw it when the flash crash moved the markets up and down hundreds of points in a matter of minutes.
This is of great concern to us, and it is why we have approved a series of measures to reduce the harm resulting from these failures, measures such as circuit breakers that halt trading when a stock price moves too far, too fast, and the more sophisticated “limit up, limit down” mechanism de-signed to prevent excessive volatility in individual stocks.
But rather than just trying to reduce the impact of these disruptions, we're trying to eliminate the causes by focusing on systems compliance and integrity.
I have asked the staff to accelerate work on a regulation aimed at improving systems compliance and integrity. We call it Reg SCI.
Reg SCI has its roots in “Black Monday,” when, 25 years ago, investors overwhelmed Wall Street's existing automated systems during a sell-off that drove the Dow's most significant one-day decline. Recognizing that automation, along with accelerating trading volume and speed, were hallmarks of modern exchanges, the SEC staff issued an automation review policy, or ARP, aimed at ensuring the stability and security of the interlocking technologies on which today's markets are constructed.
But the ARP is a policy statement, not a rule.
As chairman, I am working hard with my fellow commissioners to transform those voluntary guidelines into mandatory rules.
While some of the details of the proposal remain under discussion, I would expect this new rule to require, among other things:
• That the core technology of the exchanges, significant alternative trading systems and clearing agencies meet certain standards.
• That these entities conduct business continuity testing.
• That they provide certain notifications regarding systems disruptions and other events.
SEC CORE FUNCTIONS
As important and fundamental to our mission as market structure issues are, most of the SEC's investor protection efforts are devoted to a few core functions. So the SEC is developing analytical tools that will generate real-time risk analytics in support of our filings review, examination, enforcement and rule-making activities. These analytical tools will enable the SEC to proactively review registrant and corporate financial filing documents — along with other data — to determine whether an investigation, examination or enforcement action is warranted.
The SEC simply doesn't have the staff to go knocking on the door of every financial firm we oversee. So we created a risk-based targeting algorithm for our national examination program that analyzes information obtained from SEC filings and other sources to identify the firms most likely to be putting investors at risk. From there, the staff conducts further due diligence to focus each examination on the activities, products and relationships that present the greatest risks to investors and our markets. This approach has resulted in a significant increase both in what we call “major findings” and referrals to our enforcement unit.
In addition to this systematic-risk scoping for our examinations, we're also using tactical risk-scoping models, such as the aberrational-performance inquiry, or API — an analytical model that uses performance data to identify hedge fund advisers worthy of further review. For example, if an equity fund consistently outperforms similar funds over a period of years or has suspiciously consistent, positive results, it would be identified and examined.
With this model, Ponzi schemers, whose funds tend to claim suspiciously high and consistent returns, can be identified and singled out for scrutiny much earlier and much more comprehensively. In fact, the API has already resulted in seven enforcement actions.
REVIEW OF FILINGS
We're using this type of thinking to create other tools that aid oversight, as well. Within the SEC, we have teams of professionals who scrutinize corporate filings. But the sheer number of filings makes it challenging to discover when a disclosure contains deliberate omissions or lies. So our Division of Risk, Strategy and Financial Innovation, which we think of as our in-house think tank, is designing a new model to provide a set of quantitative analytics that examines routine regulatory filings by publicly traded companies and assesses the degree to which a company's financial statements appear out of step with its peers and therefore worthy of a closer look.
Through this model, we hope to be able to identify earnings management — which is a nice way of saying manipulative or even fraudulent accounting practices — more quickly. We are looking at both risk indicators (factors that are directly associated with earnings management) and risk inducers (factors that are associated with strong firm incentives to manage earnings).
For example, an accounting policy risk indicator might be a high proportion of off-balance-sheet transactions. Though much off-balance-sheet financing is legitimate, this type of reporting has been used in many of the largest accounting scandals, including Enron [Corp].
On the other hand, a risk inducer would be performance problems — particularly if they are considered short-term — that give management incentives to inflate current earnings in exchange for lower future earnings.
In addition, we are developing an even more sophisticated tool designed to give us enhanced detection capabilities with respect to irregularities contained in a company's written disclosure, as well as its numbers. As with the financial statement analytics, the technology we are bringing on line will compare a company's filings to its peers' — noting what topics are discussed thoroughly and what topics are not, and drawing our attention to outliers.
To me, this technology offers a 21st-century approach to disclosure review that is designed to pick up where I left off in my 2010 public remarks on the management discussion and analysis rules, which are at the core of public-company disclosure. To those in the business of preparing this critical disclosure today, I implore you to redouble your efforts to explain your company's successes and challenges.
The irregularities these tools uncover won't necessarily indicate malfeasance, nor will they detect all fraud. Their primary goal is to aid investors by improving the quality of financial disclosures. But they also help us see around corners — picking up clues that something is wrong before an investor calls in a tip, or someone's check doesn't clear.
Beyond all that, our enforcement program is being supported by several key innovations, as well.
One primary tool is our automated tips, complaints and referrals system. This system allows us to analyze and compare information received “over the transom.” TCR is the SEC's first automated, nationwide system for handling tips, ensuring, for example, that information received in a phone call to our Chicago office can be matched to an e-mail received in Miami, and that suspicious patterns can be tracked and investigated.
Another proprietary tool is what we call the automated blue-sheet analysis platform. This tool links records of significant company news — like an acquisition announcement — with trading data to help investigators identify suspicious trading patterns. Once we spot suspicious trading patterns, we can use other investigative techniques to fill in details and build a case.
The enforcement division also recently implemented a mechanism that can translate tens of thousands of phone records or lines of trading data into a visual representation that maps previously undiscovered relationships between participants in complex sets of transactions — perhaps conspiracies — something simply beyond the capacity of an individual investigator, or even a team of investigators, to detect. In a case filed last year, the SEC used this tool to track down an ad hoc insider network of nine people that stretched from Philadelphia to Hong Kong and “earned” over $1.8 million in illicit profits based on confidential information about an insurance holding company's merger negotiations with a Japanese firm.
We also have launched an audio-searching technology that allows phonetic searches of voice recordings. Such technology can help SEC attorneys find potentially relevant evidence from conversations between brokers and their customers, and conduct targeted searches without having to listen to hundreds or thousands of hours of talk.
And we have introduced an e-Discovery tool that can work its way through multimillion-page document haystacks to find the needle that an investigator needs to sew up a case. As impressive as e-Discovery's speed is its ability to use a kind of fuzzy logic similar to what Amazon uses to suggest additional books or movies based on your past searches. Many law firms have similar tools, but no firm handles nearly the quantity of data we do — six terabytes of data a month, or the equivalent of 614 million printed e-mail pages.
Technology can't replace the genius and experience of the people that work at the SEC. But like a wiretap that lets the cop stay on the beat while the bad guy's phone calls are monitored, it can allow them to be in many more places at once.
Lastly, for all our work protecting investors, we can't be physically present when a scam artist approaches an investor. So the first line of investor protection must be investors themselves. I like to think of my adorable, if fictional, Aunt Millie, who has a small portfolio but who is not able to analyze public information the way SEC staff members or an investment professional might.
Even though Aunt Millie's analytic expertise is limited, she should still have access to comprehensible information about matters such as liquidity, fees and the background of investment professionals who approach her. It will help her prevent unpleasant surprises and costly errors.
The SEC and the Financial Industry Regulatory Authority Inc. have long made information on the background of investment professionals available to the public. Last year, we combined two systems that covered two types of investment professionals, brokers and investment advisers so that investors no longer have to check two places to get that information. What's more, because of simple search engine optimization features, investors who Google their financial adviser are more likely to be sent to that combined system, even if they never knew of its existence.
Today, we're in the process of updating sec.gov, and already that investment is paying dividends. Before the recent upgrades, we were getting 200 million hits a month. Last month, we got 1.6 billion. But as they say, you ain't seen nothin' yet. It's not enough to make the information available. We need to make it more useful. For example, I believe that more expansive use of layered information, something as simple as embedding hyperlinks, would be a major step forward. We already do that with mutual funds' online summary prospectuses. There is no reason that Aunt Millie shouldn't be able to use hyperlinks in other types of online disclosure documents to move in an instant between summary information to more-detailed information whenever she wishes.
Similarly, I believe embedded pop-ups should appear to explain standard terms in plain English to Aunt Millie, who's not entirely sure what “amortization” means. This could allow investors to undertake a more independent evaluation of a potential investment outside of the presence of the financial professional who's making a pitch. And we need to provide more tools like Finra's mutual fund expense calculator to assist Aunt Millie in understanding the impact of various investment choices.
REGULATION IS IMPERFECT
All investments involve risk, and there is a chance that investors can lose their money with any investment made. I will never see a regulation, algorithm or civil penalty that successfully eliminates risk from our financial marketplace. Financial professionals, regulators, technologists and investors are human beings, and they will never be perfect, nor will the software on which technology relies.
We're still learning, but we're learning fast. And we believe that our innovative technologies, supported by in-house expertise and accumulating experience, and embedded into multiple layers of protective effort, are already lifting SEC performance and will continue to do so.
Our goal is to protect investors with technology that has the precision and sophistication to reflect the markets in which they invest today. And we're closer to that goal than we have ever been.