Keeping Up with the Quants: Your Guide to Understanding and Using Analytics, by Thomas H. Davenport and Jinho Kim, is an accessible and timely introduction to the data mining work of quantitative analysts. Aimed primarily at a business audience, it was written as a guide for the 1.5 million data-savvy managers the McKinsey Global Institute (MGI) estimated the United States will need by 2018.
Given the intense discussion in recent days of data mining in a non-business context, it seems that the audience for the book is much larger than 1.5 million. MGI may have vastly underestimated the number of people who already today need the type of education Keeping Up with the Quants provides, aiming as it does to “develop the analytical skill levels of novices.” Data—and the urge to collect and mine it—has taken by storm all corners of the globe and all types of organizations, and understanding what the analysts do is fast becoming a basic requirement for managing all aspects of our lives.
Davenport and Kim describe the process of quantitative analysis, from defining the question and the decision to be made on the basis of the analysis all the way to the presentation of the results of the analysis and taking action. They also show that creativity is not incompatible with analytical work and provide advice on how to develop quantitative analysis capabilities (e.g., “never trust numbers”) and how to work effectively with quants. The book is full of examples of successfully solving specific problems with quantitative analysis and its companion website provides additional resources and a place for readers to pose questions to the authors.
Except for a short sidebar titled “The danger of not thinking analytically,” the authors don’t dwell too much on negative examples. But the one they do bring up, credit default swaps (CDSs) sold by AIG Financial Products (AIGFP), cost the US taxpayers “somewhere around $85 billion.” The authors quote the New York Times Gretchen Morgan: “credit derivatives… are fashioned privately and beyond the ken of regulators—sometime even beyond the understanding of the executives peddling them.”
This is why “keeping up with the quants”—by reading this book and investing in other ways to get educated about quantitative analysis—is so important. What if the new quants or “data scientists” (which Davenport has correctly desc1ribed in the Harvard Business Review as having “the sexiest job of the 21st century”) will wield the same type of influence in other industries?
Numerous books and articles have been written, especially in the wake of the financial crisis, about the quants and how Wall Street “peddled” the products based on their models and quantitative analysis. If you think this will not happen again because now we know better, a quick look at the very early days of the “quantification” of Wall Street may highlight the parallels to other sectors (including the US government) discovering “data science” today. Let’s have it right from the horse’s mouth and at the moment (more or less) of creation.
It’s October 13, 1994, and Charles S. Sanford, Jr., Chairman of the Board of Bankers Trust Company, is delivering a talk titled “The Risk Management Revolution” at a conference co-sponsored by MIT and the American Mathematical Society, titled “The Legacy of Norbert Wiener.” Here are the highlights of Sanford’s speech:
I am here today to help you honor Norbert Wiener, whose work provided a mathematical foundation for many of the financial models that have fueled the changes now taking place. I am delighted to be here with many of the mathematicians and economic theorists who contributed crucial ideas to the risk management revolution…
..revolutionaries from both my world and yours, from business and academia, have created powerful new tools to identify and measure risks; to shed unwanted tasks; to acquire attractive risks; and to enable companies, governments, fiduciaries and individuals to reach their preferred balance of risk and return more easily and more efficiently than ever before… By developing tools that are both universal and precise, we have drastically reduced the gap between what the clients really need and what they can actually get…
Much more skill and knowledge is required of the providers of risk management instruments than that of users… because one valuable role of the provider is to shield the user from much of the complexity and difficulty of the underlying technology… (One can, after all, use a television without understanding the electronics inside that are required to produce pictures and sound).
…there are a number of people who are content with their current methods of risk management…. Though accountants in ancient Egypt may have kept perfect track of Pharaoh’s dominions by using a tally stick, I think our stockholders rest easier knowing that we use computers…. The technology that has created derivatives has given risk managers more options, options they should examine seriously before deciding that they prefer the old ways….
Beyond the risks to particular investors, [opponents] argue, derivatives pose a risk to the entire global financial system by linking markets and counterparties in new and dangerous ways. We believe, however, that derivatives are more likely to reduce systemic risk than to increase it; further, the benefits of derivatives would more than justify any residual systemic risk….while we cannot rule out the possibility that a chain reaction would hit the global markets, we could argue with equal plausibility that derivatives decrease system risk by diffusing risks more widely throughout the financial system.
A new class of finical institution is being created today (Bankers Trust is among the first)… not a commercial bank, not an investment bank, not an insurance company, the “risk merchant bank” is an amalgamation of all three types of institutions…. Successful risk merchant bankers… will attract highly educated people, because of the level of talent required for each specialty is rising rapidly. Ten years ago, an undergraduate engineering degree was adequate for risk modeling; today we need math Ph.D.’s…. [in 2020] automated analytics will routinely perform many tasks that from today’s perspective seem ultrasophisticated… Everyone will have access to real-time, comprehensive, marked-to-market financial statements so that they may monitor and manage their total net worth… While imagining all these innovations might threaten to make our heads swim, let me remind you that just as you do not need to understand quantum mechanics to use a transistor radio… you will not have to understand all the intricacies of particle finance to use it to improve your risk profile. That will be the job of the risk merchant banker.
It’s all there: The blind faith in computers and models, the familiar threats of “if you don’t join the revolution, you will be left behind,” the arrogance of “trust us, you don’t need to – and can’t – understand the products of our superior brains.” Do you see it happening today in other industries and sectors, in all types of organizations where executives don’t want to be left behind the “data revolution” and blindly trust quantitative analysis without understanding what data scientists do?
Note I used the word “may” in the title for this post. Keeping up with the Quants certainly helps explain quantitative analysis to anyone who doesn’t want to leave “the intricacies of particle finance” (or “particle retail” or “particle marketing” or “particle snooping”) to the quants. Reading the warnings of Data Scientists such as Kate Crawford about the biases and myths of big data (here and here) is also a good medicine for anyone who wants to be cured of dataitis.
But what if executives, policy-makers, everybody, don’t want to be cured? What if we fervently believe or want to believe that Wall Street’s mistakes will not—cannot—be repeated in other industries and sectors? What if we have already bought into the big data myth that “The effectiveness of data-mining is proportional to the size of the sample” (as the Wall Street Journal editorial pronounced categorically in defense of the NSA)? What if we can’t wait to see a revolution similar to “the risk management revolution” sweeping all industries?
The financial crisis didn’t happen only because new mathematical and information technology tools and the creative people to handle them became available to Wall Street. It also did not happen only because of executives “peddling” new financial products like Bankers Trust’s Sanford and AIGFP’s Cassano. It happened because of demand, not just because of supply.
Last week, the Wall Street Journal reported: “…banks and investors battered by CDOs during the financial crisis are increasingly willing to ignore bad memories in order to reach for higher returns. In markets ranging from commercial mortgage-backed securities to junk bonds, investors are eager to buy even the very riskiest investments, some of which now deliver yields of more 20% per year.” For now, it seems, hedge funds are the main customers for these new-old financial instruments. But how long before pension funds also start to “ignore bad memories”?
The pressure for “performance” is not unique to Wall Street. The new “science” for improving performance is based on data and its analysis. There is no doubt in my mind that in many situations it is much better to base decisions on data rather than on intuition or the fashion of the moment or on politics (of any kind). But basing your decisions on blind trust in numbers, without understanding the work of the people producing them, is a sure recipe for more crises, financial or otherwise, ahead.
[Originally published on Forbes.com]
Reblogged this on kwalitisme.
Thanks, looks like a good read, will pick it up!