Customer Reviews

Plight of the Fortune Tellers: Why We Need to Manage Financial Risk Differently

5 star:		(5)
4 star:		(5)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

In my opinion this is the most valuable book on investment risk management of the past few years. Yet, no equations! However, with cogent arguments and literate prose, Rebonato lays out a case against the unfortunately prevalent misuse of statistical models in risk management.

Second edition should fix the minor annoyances, like "manger" for "manager" (appearing several times) and "form" for "from" (ditto), but the content should be read by everyone with interest in the area.

Especially welcomed are his arguments. Rather than setting up straw swans and knocking them down, or simply labeling alternative views as offensive or idiotic, he carefully sets out deep background for thinking about risk, and for thinking about probabilities, then shows how and why the well-meaning (and useful in the right context) VaR ideas are on a trajectory that is likely to go horribly wrong.

What to do? Unfortunately, the problem is hard and there are likely no easy solutions. But thinking correctly (my word) about the problem lets us roll up our sleeves and work on the right parts of the problem.

Investment management is all about risk management. We want to understand the risks in front of us, accept the risks we think we can get paid properly for, and avoid the ones where the bet is not in our favor. The Rumsfeldian "unknown unknowns" are the ones that are likely to cause the most damage. Those are what should keep us up at night trying to imagine. If they become "known unknowns", e.g. liquidity and linkage risks which showed up July/Aug 2007, we can get to work understanding and managing them.

Best (financial/investment) book of the year.

Rebonato challenges the "frequentist" approach to probabilities employed by stock analysts and rating agencies and finds lots to worry about. He says that although looking back at the past gives you masses of data that can be parsed and analyzed lots of different ways, it gives a dangerously miselading sense of security that future probabilities can be systematically determined with great prescision. The problem is that that whole thing is based on the idea that market moves are like coin flips, or monte carlo simulations, which say that while market prices change and fluctuate, that their underlying structure never actually changes. In fact, the probablities that really count are the those in the future, not those of the past. To predict those you need to understand what Rebonato dubs "subjective probability" - which while much more qualitative as opposed to mathematical, can actually be much more accurate and predictive. This is well worth thinking about, and is clearly explained for you to make your own judgment.

The main theme is that risk management is not about measuring risk, or assessing probabilities, but it is about making decisions under uncertainty. The author says that the existing framework of risk management, which is heavily based on "frequentist" approach to probabilities (i.e. repeatability under identical conditions, weak prior beliefs, etc.) does not necessarily serve for decision-usefulness associated with managing risks; "subjective" (Bayesian) probabilities tend to be better suited to the purposes. Focusing on the outcome of decisions relieves us from dogmatic probabilists and allows us eclectically to arrive at the best prediction we can, using whatever tool we have at our disposal. While the author's argument appears to make a lot of sense, the Bayesian probabilities brings in subjectivity such as prior information/knowledge, which in itself seems helpful, I wonder what if we are not confident of such prior information, as we cannot know what we cannot anticipate (i.e. an "unknown unknown": an uncertainty that is unanticipated)? Or put it differently, if we already have had good, reliable prior information about whatever the risk we attempt to assess, then, we would not have much to worry about to begin with, I presume..... Well, we probably should not try to rely on statistical approach to such an extremely high percentile to be considered effectively meaningless (I hasten to throw in my disclaimer here that I am not proficient enough in statistics to discuss the matter in detail!)
Those who have found Nassim Nicholas Taleb's "The Black Swan" and "Fooled by Randomness" fascinating would be intrigued by this timely, engaging , and highly accessible account, which provides not only professional risk managers but also amateur investors like me with numerous insights.

This book is one of the many that have come out in the last few years that has addressed the virtues and vices of financial modeling. Many of these books are devoted to the proposition that modeling has caused deep problems in the financial markets, but the evidence they present for this assertion is typically very weak. Considering the scale of modeling in financial institutions throughout the world, it would be naïve to assume that modeling has not influenced the markets, but it would also be unjustified from an empirical standpoint to say that modeling has been the predominant influence in market degradation. But if one believes that modeling has played the major role in this regard, then there will be a strong temptation to seek alternative methodologies for optimizing the risk/return trade-off.

The author is one of these, as can be ascertained early on in the book where he refers to data as giving "power to actions and decisions." However, the author is aware of the problems with the misguided imputation of power to concepts or ideas that are applied to contexts that are extremely rare in human experience. Thus he devotes several pages of the book to the "frequentist" interpretation of probability, and offers the Bayesian alternative. This is not to say that the frequentist approach should be completely discarded, for he discusses contexts where it is appropriate. One of these concerns the need for say a 99.9 percentile in some implementations of the Basel II accords. Such a level of confidence will be very problematic from the standpoint of validation given the paucity of real historical data. The author also offers suggestions for how risk managers are to clean up their act in the final chapter of the book.

He also discusses the possible use of belief theory in risk management, but apparently he is not aware that this approach has been tried in some contexts. In fact, this reviewer has applied some of the concepts from belief-theory to the problem of mortgage-broker scoring. Belief theory even has a "belief calculus" that has been applied to the modeling of financial portfolios, with the goal of learning how the returns change as new information is obtained on the factors that impact the portfolio. The belief calculus is similar to what is done using Bayesian networks, but with belief functions used to model the dependencies in the factors. Belief theory abandons the additive principle of probability theory, in that the 'belief mass', or "degree of belief" that is assigned to certain sets does not have to sum to one. However, belief functions is that they can be expressed as a probability using the so-called 'pignistic transformation', but one will obtain a loss in information if this is done. The author asserts that belief theory is a viable methodology, in that one's "confidence" that a certain event or number of events is about to occur is expressed by the willingness to "bet on" that event or events. But "credo" is Latin for "I believe" and "pignus" is Latin for a wage or bet, and certainly in everyday conversation one frequently hears "it is my belief that this will happen....I would bet a month's wages."

There is no arguing that decision making is the real essence of risk management, but does this have to involve subjective judgments, as the author seems to imply, or can it be done by a suitable collection of algorithms that are sophisticated enough to deal with most contingencies? If so, could this be taken one step further and allow the decision-making process to be automated, possibly using intelligent machines? Given the advances in artificial intelligence, this scenario is getting more plausible. But in all approaches to risk management, whether automated or not, one must still answer whether the human or machine estimation of probabilities of events is meaningful and how to assess if this is the case. Will this involve the use of traditional statistics or will some other approach be used?

Prospect theory, also discussed in the book, has certainly been a useful paradigm in risk management, to the degree that it has been utilized. But indeed how can one really know what concepts or methodologies are actually being employed by senior risk managers? In many cases, the analyst or modeler makes assumption that the management is using the results of the modeling efforts, but instead the management is relying on intuition or guesswork to make risk decisions, and completely ignoring the data from the models. In addition, distinguishing the impact of decisions based on modeling versus those based on intuition is more difficult than is realized at first glance.

Another important point to make here is that the Bayesian approach to the calculation of probabilities may not be part of the model itself, but frequently plays a major role in the validation and empirical support for the model. A similar situation occurs in other fields, such as physics, where Bayesian calculations permeate experimental confirmation of theories, but where the theories are stated in a frequentist framework.

Given the extreme events in the fixed-income sector at the present time, it is difficult to argue with the author's claim that financial risk management must be done in a different way. In fact, just this month a popular technology journal referred to a meeting of a couple of hundred of the more well-known financial modelers, who declared the summer of 2007 to be the worse ever for financial modeling. So the author is not alone in his opinions. But due to bureaucratic inertia and resistance from the status quo, finding the right time to implement these changes can be problematic, even when there is unanimous agreement that these changes are necessary.

This book was an early warning to the professional risk management teams of banks about what is predictable and to what confidence intervals events can be realistically viewed within. The hard scienced professionals within the risk management divisions of bank have developed, over the last few decades, more and more sophisticated tools to try to measure financial risks. The author believes that both data mining and distribution fitting are being over used to give a sense of false security on the stresses of a portfolio. He believes that backward looking frequentist methodologies for stress testing run into the problem of having too few examples of similar economic backdrops to make statistically significant conclusions and he believes that fitting known distributions to observable data and abstracting to what the mass under the tails is misguided for a multitute of reasons. All of these criticisms most people think go without saying, but glancing through risk literature in which authors claim that they believe with more than 99.9x% confidence that they will be able to manage some risk remind us that common sense doesnt always prevail in sophisticated institutions.

The author argues his perspective from various angles. The one which has most merit is probably the idea that paramater estimation in noisy environments makes it impossible to have high confidence in ones paramater assumptions. Sensitivities of distribution of outcomes to parameter estimates are often of large orders of magnitude in the tails of distributions. All of us who hear people describe losses as 6 sigma events that should have never happened, the very description of it as an event of exceedingly low probability which seems to happen with much higher probability than the people who claim their uniqueness, makes us believe it is almost certainly the assumed model of the world was wrong, not we were just exceedingly unlucky to have witnessed an event that happens once in the solar systems existence. The author also talks about the world and its distribution of outcomes is non-stationary, on all time scales. He describes the fact that rare events, by definition of being rare and our relevant history too short to have significant frequency estimations, cannot be fit into probability estimations with high precision.

The author also encourages risk management to delve much more deeply into subjective probability assesments. He does not believe the past will tell us the future and data mining might bring about a sense of false confidence without proper analysis of future risks. US housing is a fantastic example of this, models were all backward looking and thus probabilities of collapse were skewed based off historical stresses not being relevant to future stresses, a data set of 70 yrs would not have been sufficient to cause warning of the danger that we experienced. Sometimes over reliance on historical data as a predictor of future returns can cause bubbles in themselves by creating the illusion of riskless gains in risky markets in excess of risk free rates.

Most of the book gets to what almost all investment professionals realize. Risk isnt math, there are elements of math which help us aggregate risks, but their utility is bounded both by the data we have as well as the assumptions we make. Why do people talk about confidence intervals that are not only untestable but also nonsensical. To talk about a portfolio not being able to sustain a certain amount of loss given a time scale of human civilization really doesnt need to be argued as garbage, we all know it is, so why does it arise? Part of it is trying to bring a sense of confidence to the investors in banks at the creditor and equity level. Do the investors believe the confidence? Probably not, but it sounds like the bank is being prudent at least. This book was written and published before the financial crisis, its points, had they been internalized might have prevented the magnitude of misallocation of resources but i doubt the losses would have been avoided altogether. I think the author did a good job in articulating where risk management has been going wrong. Risk management he articulates requires better integration of those who develop the mathematical machinery and those who make investment decisions at the trading and executive level and most importantly the need for better incentive allignment of all of them with the long run of the firm. If this could be achieved, we would have come a long way.

This book should be required reading for every probstat course. It is intended for financial managers primarily; as such the formal math has intentionally been limited to basics. But what has been lost in rigor is balanced by broad-brush qualitative description and concrete examples that both bring home & motivate the applicability of Bayesian probability and help develop the reader's intuition.

Having completed grad courses in probability, comms theory, and information theory within the past five years, I do not believe that I am exaggerating in saying that any student of math/engineering/science will benefit from a reading of "Plight", either by itself or accompanying a textbook. Rigorous presentations of highly technical material (aka "textbooks") often fail to get across the vitality of the discipline - "rigor mortis", one textbook author called this. Rebonato's book is an excellent antidote.

Not least because it is so gracefully written.. I am re-reading the book after a lapse of a year or so and find it as delightful as on first reading.

"Risk management" sounds dry as dust, and yet ... it has just this morning reminded me of recent personal experience.

I recently spent a couple of hours with a financial management consultant. Among their other resources, his company leans heavily on market simulations. Indeed, their web site provides a Monte Carlo simulator that customers can use to evaluate market exposure associated with an investment strategy. In Chapter 7 under the section heading "Beautiful Monte Carlo" and more generally in that chapter Rebonato discusses some of the shortfalls of financial modeling. His reservations will directly inform my decisions.

In the second case, a very talented associate recently took one of his creations into the field for trials. Those of us with visibility into his project were surprised that performance did not meet expectations. Indeed, performance wasn't even close to what was needed.

As Rebonato points out:

1) The reliability of one's predictions depends critically on the quantity of *relevant* data one has collected. It is my impression that although my colleague collected quite a lot of data in simulations prior to the field trials, those simulations may not have been directly relevant to the field trials.

2) Unless you are aware of the limitations of your toolbox, your tools may not inform you. They may indeed blind you. If you wish to avoid unpleasant surprises, particularly with low-likelihood events ("in the tails"), you should consider whether your tools are relevant to your problem *before* you embark on the application of those tools to your problem.

Perhaps a bit of a stretch to apply Rebonato's arguments to the experience, but that is part of the value of this book: it causes you to think.

I cannot recommend this book highly enough.

Reading this book is like reading the prophecies of Cassandra, whose fate was to speak the truth to the unbelieving. In the wake of the fiscal crisis, no one can question the observation that the practice of financial risk management has serious flaws. Business journalist Riccardo Rebonato's discussion of why and how financial institutions misunderstand and mismanage risk provides valuable insights. He works to make his ideas accessible beyond the narrow circles of financial economists and quantitative risk managers. He uses no equations in the text, and his few graphs are clear and accessible. He states his case against excess reliance on statistical methods in plain language. getAbstract believes that his analysis should interest any manager or regulator whose responsibilities include oversight of finance.

Plight of the Fortune Tellers: Why We Need to Manage Financial Risk Differently (New in Paper)
I am a student of finance and banking, and am very interested in the matter of risk management. Although I am not an expert on the subject, I can read the lines of Rebonato's Plight of the Fortune Tellers and realize that something is not being handled diligently on quantitative methods in finance. Sometimes you have to circumvent the mathematical rigor to make coherent and meaningful decisions. He wrote about risk management appealing to common sense!

Riccardo Rebonato wrote Plight of the Fortune Tellers in a way that is accessible to a relatively wide readership. The book deals with complex issues in financial risk, but includes no equations. However, this is not a book that will be interesting to those without some background in markets and finance. If you have not studied finance at all you would probably miss some of the points that Dr. Rebonato makes.

Considering the fact that the entire world financial system would have failed in 2008 without massive government intervention it is pretty clear that financial risk needs to be looked at in a different way. This book provides some foundation for this, but the issue is much larger, with a significant political dimension. What Dr. Rebonato does is provide an analysis of risk that you will rarely find in finance courses. In fact, I had just finished a finance course when I read this book. For one of the class projects we used five years of monthly return data (e.g., sixty monthly returns). On several occasions we estimated the 5% Value-at-Risk. What I had not considered and what Dr. Rebonato points out at some length was that this estimate was questionable. There were only a few data points out at that end of the curve. As Rebonato points out, trusting any analysis with so few data points requires careful thought.

The book follows this pattern: looking more deeply into the financial techniques and assumptions that are frequently used without many questions. The book also makes an argument for Bayesian statistics. Like most people (and the readers the book is aimed at) I studied classical "frequentist" statistics, so its hard for me to know how valid Rebonato's argument is. I did, however, order a copy ofDoing Bayesian Data Analysis: A Tutorial with R and BUGS.

At one point in the book Rebonato discusses why there should be risk officers or risk analysis professionals at investment banks. He had previously discusses cases where rogue employees had caused huge losses, but oddly he did not list "risk police" as a reason for risk professionals. This seems odd to be me because investment banks have a problem that risk analysis can address: traders are trading with other people's money. A trader stands to gain if their bets pay off. At worst they get fired if they lose and, as those at Long Term Capital Management found out, this does not necessarily end their career. Without risk professionals to oversee a bank's risk, there is the danger that the traders will "blow up" the bank. Or the hedge fund.

I recommend this book for anyone interested in finance and it is certainly approachable for someone with a business focused MBA background. For those with a more quantitative focus the book will provide an important perspective on what they have learned in class.

I should, perhaps, have given the book five stars. Rebanato sets himself the difficult task in writing a book for a more general readership. As a result, there is a limit to the detail provided. Rebanato argues for a Bayesian approach, but I have no idea how Bayesian statistics might be used. Given the objective of the book, this is understandable and probably unavoidable. But its still a frustrating feature of the book.

I have long like the insight of Bernstein, doctor turned author. If you like understanding how world business evolves on a very long time scale, then you will like this book. The adage that history repeats in cycles is driven home. Globalization, trade imbalances, plagues, and power struggles all occurred many times in the past. Some groups benefited; some didn't. Read and heed!