What do you think?
Rate this book
Radical Uncertainty: Decision-Making Beyond the Numbers
Some uncertainties are resolvable. The insurance industry’s actuarial tables and the gambler’s roulette wheel both yield to the tools of probability theory. Most situations in life, however, involve a deeper kind of uncertainty, a radical uncertainty for which historical data provide no useful guidance to future outcomes. Radical uncertainty concerns events whose determinants are insufficiently understood for probabilities to be known or forecasting possible. Before President Barack Obama made the fateful decision to send in the Navy Seals, his advisers offered him wildly divergent estimates of the odds that Osama bin Laden would be in the Abbottabad compound. In 2000, no one—not least Steve Jobs—knew what a smartphone was; how could anyone have predicted how many would be sold in 2020? And financial advisers who confidently provide the information required in the standard retirement planning package—what will interest rates, the cost of living, and your state of health be in 2050?—demonstrate only that their advice is worthless.
The limits of certainty demonstrate the power of human judgment over artificial intelligence. In most critical decisions there can be no forecasts or probability distributions on which we might sensibly rely. Instead of inventing numbers to fill the gaps in our knowledge, we should adopt business, political, and personal strategies that will be robust to alternative futures and resilient to unpredictable events. Within the security of such a robust and resilient reference narrative, uncertainty can be embraced, because it is the source of creativity, excitement, and profit.
The limits of certainty demonstrate the power of human judgment over artificial intelligence. In most critical decisions there can be no forecasts or probability distributions on which we might sensibly rely. Instead of inventing numbers to fill the gaps in our knowledge, we should adopt business, political, and personal strategies that will be robust to alternative futures and resilient to unpredictable events. Within the security of such a robust and resilient reference narrative, uncertainty can be embraced, because it is the source of creativity, excitement, and profit.
544 pages, ebook
First published March 5, 2020
About the author
John Kay
53 books126 followersI was born in Edinburgh, Scotland in 1948, and completed my schooling and undergraduate education in that city: I am fortunate to have lived most of my life in beautiful places. I went to the University of Edinburgh to study mathematics. But, after taking a subsidiary course in economics, I decided that I wanted to be an economist. The notion that one might understand society better through the application of rigorous and logical analysis excited me – and it still does. After graduating from Edinburgh, I went to Nuffield College, Oxford. I worked there under James Mirrlees, who was in due course to win the Nobel Prize for his contributions to economic theory.
On Mirrlees’s advice, I applied for and to my astonishment got a permanent teaching post in the University of Oxford at the embarrassingly early age of 21. Oxford is a collegiate university – members of the faculty generally have both University and College appointments. This post carried with it a fellowship at St John’s College, an association which I have maintained and enjoyed ever since. Through the 1970s I developed a conventional academic career, publishing in academic journals, and writing my first book Concentration in Modern Industry (with Leslie Hannah, an economic historian colleague). My particular interests were in public finance and industrial organisation.
But my rationale for studying economics had, from the beginning, been concerned for application. My career began to change direction when I was asked to join a group to review the structure of the British tax system. This group was established under the auspices of a newly established think tank, the Institute for Fiscal Studies, and was headed by James Meade, another economist who had achieved the ultimate distinction of a Nobel prize. Meade’s rigour was as demanding as that of Mirrlees, (both delivered it with extraordinary personal charm). But the most important effect of my experience with the Meade Committee was that I began to develop a taste for the popular exposition of economic concepts.
In this vein, I wrote (with Mervyn King, now Governor of the Bank of England) a more personal account of issues in taxation, The British Tax System which ran through five editions.
Pursuing these interests, I moved from Oxford and joined the Institute for Fiscal Studies as its first research director. Soon after I became Director of the Institute. IFS developed into (and remains) one of Britain’s leading think tanks, respected and feared by policymakers and journalists for its fiercely independent analysis of fiscal issues.
After seven years, I decided it was time to move on. The success of IFS had been built on serious economics accompanied by a commitment to popularisation and application. If this could be done for public policy issues, could the same be done in the area of business policy? This was the thinking that led me to accept a chair at the London Business School in 1986 and, at the same time, to establish a consulting company, London Economics.
Over the next few years, the application of economics to business issues became my intellectual focus. During this period, London Economics grew rapidly, largely on the back of the wave of privatisation and regulatory change in Britain in the 1980s. By 1991, managing the company had become a major responsibility. I revised my arrangements with London Business School. My new contract was as Visiting Professor but my job as executive chairman of London Economics took the larger part of my time. London Economics grew until, by its tenth anniversary, its annual turnover exceeded £10m with offices in three continents and assignments in over sixty countries.
London Economics gave me insights into the business world. These came both through consultancy work with major corporations and first hand observation of the growth and development of a small business. Other activities have enriched my more scholarly work by broadening the experi
On Mirrlees’s advice, I applied for and to my astonishment got a permanent teaching post in the University of Oxford at the embarrassingly early age of 21. Oxford is a collegiate university – members of the faculty generally have both University and College appointments. This post carried with it a fellowship at St John’s College, an association which I have maintained and enjoyed ever since. Through the 1970s I developed a conventional academic career, publishing in academic journals, and writing my first book Concentration in Modern Industry (with Leslie Hannah, an economic historian colleague). My particular interests were in public finance and industrial organisation.
But my rationale for studying economics had, from the beginning, been concerned for application. My career began to change direction when I was asked to join a group to review the structure of the British tax system. This group was established under the auspices of a newly established think tank, the Institute for Fiscal Studies, and was headed by James Meade, another economist who had achieved the ultimate distinction of a Nobel prize. Meade’s rigour was as demanding as that of Mirrlees, (both delivered it with extraordinary personal charm). But the most important effect of my experience with the Meade Committee was that I began to develop a taste for the popular exposition of economic concepts.
In this vein, I wrote (with Mervyn King, now Governor of the Bank of England) a more personal account of issues in taxation, The British Tax System which ran through five editions.
Pursuing these interests, I moved from Oxford and joined the Institute for Fiscal Studies as its first research director. Soon after I became Director of the Institute. IFS developed into (and remains) one of Britain’s leading think tanks, respected and feared by policymakers and journalists for its fiercely independent analysis of fiscal issues.
After seven years, I decided it was time to move on. The success of IFS had been built on serious economics accompanied by a commitment to popularisation and application. If this could be done for public policy issues, could the same be done in the area of business policy? This was the thinking that led me to accept a chair at the London Business School in 1986 and, at the same time, to establish a consulting company, London Economics.
Over the next few years, the application of economics to business issues became my intellectual focus. During this period, London Economics grew rapidly, largely on the back of the wave of privatisation and regulatory change in Britain in the 1980s. By 1991, managing the company had become a major responsibility. I revised my arrangements with London Business School. My new contract was as Visiting Professor but my job as executive chairman of London Economics took the larger part of my time. London Economics grew until, by its tenth anniversary, its annual turnover exceeded £10m with offices in three continents and assignments in over sixty countries.
London Economics gave me insights into the business world. These came both through consultancy work with major corporations and first hand observation of the growth and development of a small business. Other activities have enriched my more scholarly work by broadening the experi
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 30 of 100 reviews
March 17, 2020
For a hyper-discursive one-idea book, this was actually rather good!
Here’s the idea:
Economics would do well to turn its mathematical prowess away from the idea that economic agents are busy maximizing something and must focus on “what’s going on here.”
And “what’s going on” is quite simple: we all have some sort of dream! (the word the authors use is “narrative.”) As we gather information, we assess how this new information influences our dream and adapt accordingly.
The End.
This is neither Mervyn King’s nor John Kay’s best book, but that’s a mighty high bar. John Kay is one of my Gods (for all his contributions to the FT) and author of three excellent books, including “The Truth About Markets” (though I’d give a miss to the rather grumpy “Other People’s Money.”) Mervyn King is the author of “The End of Alchemy,” a truly AMAZING, one-lesson-per-paragraph treatise about the 2008 crisis and what central banks ought to do to avoid the next one.
Speaking of which, they’re totally on-the-ball: top of page 40 they as-good-as predict the current coronavirus crisis. And in pages 408-409 they introduce the concept that “the large corporation is the most important actor in the modern economy, and it is surprising how little attention has been given to the economics of organisation.”
So there you have it: to get out of our bind, to understand the economy and to be able to make decent predictions about the future, we need to study the narrative that’s in the head of the captains of industry and predict how they will react to new information, be that incremental or some type of shock. That’s the main takeaway here.
But that’s to sell the book short. It’s full of amazing stories, references millions of recent books without ever being overbearing and makes fun of David Viniar on pages 6, 9, 58, 68, 109, 150, 176, 202, 235 and 366.
It’s kept me good company for a week and I recommend it to anybody looking for what strongly feels like the last book from two very good authors.
Here’s the idea:
Economics would do well to turn its mathematical prowess away from the idea that economic agents are busy maximizing something and must focus on “what’s going on here.”
And “what’s going on” is quite simple: we all have some sort of dream! (the word the authors use is “narrative.”) As we gather information, we assess how this new information influences our dream and adapt accordingly.
The End.
This is neither Mervyn King’s nor John Kay’s best book, but that’s a mighty high bar. John Kay is one of my Gods (for all his contributions to the FT) and author of three excellent books, including “The Truth About Markets” (though I’d give a miss to the rather grumpy “Other People’s Money.”) Mervyn King is the author of “The End of Alchemy,” a truly AMAZING, one-lesson-per-paragraph treatise about the 2008 crisis and what central banks ought to do to avoid the next one.
Speaking of which, they’re totally on-the-ball: top of page 40 they as-good-as predict the current coronavirus crisis. And in pages 408-409 they introduce the concept that “the large corporation is the most important actor in the modern economy, and it is surprising how little attention has been given to the economics of organisation.”
So there you have it: to get out of our bind, to understand the economy and to be able to make decent predictions about the future, we need to study the narrative that’s in the head of the captains of industry and predict how they will react to new information, be that incremental or some type of shock. That’s the main takeaway here.
But that’s to sell the book short. It’s full of amazing stories, references millions of recent books without ever being overbearing and makes fun of David Viniar on pages 6, 9, 58, 68, 109, 150, 176, 202, 235 and 366.
It’s kept me good company for a week and I recommend it to anybody looking for what strongly feels like the last book from two very good authors.
April 18, 2020
I've read both Kay and King's individual books and am a fan of both. This book needed some editing help, but the insights are worthwhile, incredibly relevant, and essential for everyone to understand--especially for financial regulators.
March 26, 2020
This is a book by two well known British economists, both of whom are accomplished scholars and analysts of economics, corporate strategy, and finance. The premise is fairly straightforward and starts from the work of Frank Knight (1929) who distinguished between risk (a lack of knowledge that can be parameterized) and uncertainty (a lack of knowledge that we cannot parameterize and about which we know little). The study of risk gives us a range of productive fields of study linked to statistical decision making and problem solving. The study of uncertainty is as or more important to human decisions but also much less amenable to structured programs of statistical analysis. Sometimes we just do not know what to say or do about uncertainties and available measurements and numbers will not prove to be useful.
Fair enough, but then the plot thickens. The economists at Chicago that followed Knight figured out ways to parameterize uncertainty (or so they thought) - leading to subjective probability analysis and a number of other variants — in effect collapsing Knight’s distinction and launching decision analysis and financial economics to new heights - as fields that could apply rigorous analysis to all areas of human interaction. Kay and King see this as an original sin of sorts that launched finance and banking on the road that led to overstretched models detached from reality and ultimately to the market collapse of 2007-2008.
Yes, it appears that ideas can make a difference. The book has a critical edge to its view of how economics and finance developed.
So what to do about it? Kay and King argue that uncertainty should be rediscovered and re-embraced as Knight first intended. They focus on this as “Radical Uncertainty” and develop the idea of how our understanding of the world and of decision making changes with this change of focus. This is an argument for abductive reasoning rather than deduction or induction. It focuses on stories and narratives and argues that while deduction and axioms retain importance, they do so in the communication of policies and decisions rather than in the actual decision making and choice. What is also crucially important in a situation is to find out “what is going on” in a situation before rushing to measure everything and apply general statistical models whether or not they are really applicable.
What follows is an amazing book that takes the reader on a tour of contemporary areas of decision analysis, behavioral economics, finance, and policy studies to show what is wrong with the conventional reasoning and what is possible to achieve by embracing radical uncertainty. I cannot begin to do justice to the range of topics under one roof in this book. This is also a compelling discussion of academic versus applied policy approaches to decision making, along with the ways to evaluate the success and usefulness of different analytic approaches. This makes the book a crucial addition to the reading list of anyone seeking to learn about the advisory professions.
It is always possible to take issue with this or that point in a book, especially one on the state of economics and its variants. However, I read this at a time when an unknown virus was racing across the world, when I was confined to home in little more than house arrest, and when the capital markets had declined more (and more rapidly) in two weeks than they had in the prior 80 years. While this was happening, the number of identified virus cases spiked all over the world, the healthcare system was threatened with crashing, and the economy was grinding to a complete halt.
Under the circumstances, a book about radical uncertainty seems very much on target and I think that the authors have a lot to say that resonates today.
Fair enough, but then the plot thickens. The economists at Chicago that followed Knight figured out ways to parameterize uncertainty (or so they thought) - leading to subjective probability analysis and a number of other variants — in effect collapsing Knight’s distinction and launching decision analysis and financial economics to new heights - as fields that could apply rigorous analysis to all areas of human interaction. Kay and King see this as an original sin of sorts that launched finance and banking on the road that led to overstretched models detached from reality and ultimately to the market collapse of 2007-2008.
Yes, it appears that ideas can make a difference. The book has a critical edge to its view of how economics and finance developed.
So what to do about it? Kay and King argue that uncertainty should be rediscovered and re-embraced as Knight first intended. They focus on this as “Radical Uncertainty” and develop the idea of how our understanding of the world and of decision making changes with this change of focus. This is an argument for abductive reasoning rather than deduction or induction. It focuses on stories and narratives and argues that while deduction and axioms retain importance, they do so in the communication of policies and decisions rather than in the actual decision making and choice. What is also crucially important in a situation is to find out “what is going on” in a situation before rushing to measure everything and apply general statistical models whether or not they are really applicable.
What follows is an amazing book that takes the reader on a tour of contemporary areas of decision analysis, behavioral economics, finance, and policy studies to show what is wrong with the conventional reasoning and what is possible to achieve by embracing radical uncertainty. I cannot begin to do justice to the range of topics under one roof in this book. This is also a compelling discussion of academic versus applied policy approaches to decision making, along with the ways to evaluate the success and usefulness of different analytic approaches. This makes the book a crucial addition to the reading list of anyone seeking to learn about the advisory professions.
It is always possible to take issue with this or that point in a book, especially one on the state of economics and its variants. However, I read this at a time when an unknown virus was racing across the world, when I was confined to home in little more than house arrest, and when the capital markets had declined more (and more rapidly) in two weeks than they had in the prior 80 years. While this was happening, the number of identified virus cases spiked all over the world, the healthcare system was threatened with crashing, and the economy was grinding to a complete halt.
Under the circumstances, a book about radical uncertainty seems very much on target and I think that the authors have a lot to say that resonates today.
August 23, 2020
Sorry nothing here if you have already read Nassim Taleb’s Black Swan and Antifragility.
The book is 140,000 words long and the main idea could perhaps be summarized in one-tenth of the book. Many chapters can actually be culled as they bore only cursory relevance to the topic of “radical uncertainty”.
Understandably, this book contains is a lot of attack on mainstream economics. However again nothing original here and much better read would be Richard Thaler’s Misbehaving or Danny Kahneman’s Thinking Fast and Slow or Michael Lewis’s Undoing Project.
Interestingly, this book also threw punches at Kahneman’s style behavioral science research saying that those were also “small-world” theories that would lead to “theory-induced” blindness.
The alternative being proposed is that the so called “biases” are honed over many years of evolution (very Taleb once again) and therefore serve some superior purposes. (What they call “evolutionary rationality” vs “axiomatic rationality”)
They also propose that we should embrace narratives and crowd wisdom to tackle radical uncertainty. And that basically no model can replace human ingenuity (a la Blink by Malcolm Gladwell) and that we should trust our gut feeling / expert hunch.
Overall a decent read with some interesting anecdotal stories but lacking original ideas.
The book is 140,000 words long and the main idea could perhaps be summarized in one-tenth of the book. Many chapters can actually be culled as they bore only cursory relevance to the topic of “radical uncertainty”.
Understandably, this book contains is a lot of attack on mainstream economics. However again nothing original here and much better read would be Richard Thaler’s Misbehaving or Danny Kahneman’s Thinking Fast and Slow or Michael Lewis’s Undoing Project.
Interestingly, this book also threw punches at Kahneman’s style behavioral science research saying that those were also “small-world” theories that would lead to “theory-induced” blindness.
The alternative being proposed is that the so called “biases” are honed over many years of evolution (very Taleb once again) and therefore serve some superior purposes. (What they call “evolutionary rationality” vs “axiomatic rationality”)
They also propose that we should embrace narratives and crowd wisdom to tackle radical uncertainty. And that basically no model can replace human ingenuity (a la Blink by Malcolm Gladwell) and that we should trust our gut feeling / expert hunch.
Overall a decent read with some interesting anecdotal stories but lacking original ideas.
August 7, 2020
I admit to starting this book and being a bit bored because I read a lot about uncertainty and risk and how that affects decision making. I have read so many books about mathematics, statistics, modelling, economics, human decision making, etc. So for the first few chapters, I will admit that my attitude was all wrong, and I was thinking "what is this book contributing that's new?" If I am really honest, it wasn't until Chapter 8 that it completely clicked and I realised that this book is truly different, and I finally understood exactly what it was contributing that was so important. The authors really do have a unique perspective that is borne out of real world experience and a healthy skepticism towards faux precision and (my new favourite term) "mathiness". We have all experienced it - decision makers who find the allure of numbers irresistible, even if many of the assumptions in their theories and models are, frankly, b.s. Garbage in, garbage out, as they say....and yet so much of our lives are ruled by really bad models and really sub-par theories. It happens even more in economics than in policy areas, but what has always worried me is how economists have managed to frame their work as robust when it is on such shaky ground. This book is logical, strongly grounded in an understanding of economics as well as real world decision contexts, and really well researched. Having read a lot of Tversky and Kahneman, who I felt had made some pretty great advancements on the very poor understanding of human behaviour in economics, I especially learned a lot from their critiques of the aforementioned authors' study designs, and it has made me think differently about their work and the conclusions that have been drawn from it. From a personal point of view, as someone who draws a lot on communicative rationality, framing, and narratives in my work, I was so pleased to find these threads so prominently appearing in this work, and it really will expose many more people to these important aspects of argumentation and reasoning. So while I do think this book starts a bit slow and could make more of its unique contribution from the outset, I will say that if you stick with it, you will have a sort of "aha" moment. Depending on how much of this genre you have read, it may take you more or less than 8 chapters, but if you are interested in the topics covered in this book, I do think you will get a lot out of it that you won't get from any of the other standard books that people recommend on the topic. It is one of the few popular non-fiction books I have read in recent years that I think hangs together logically incredibly well, and manages to marry both academic ideas and practical policy and societal concerns.
June 5, 2020
Radical Uncertainty makes an important contribution to the current debate. It challenges the role of technocratic experts, and the mathematical models that they hide behind, in decision making. It puts into question the Nostradamus level pretensions of social scientists and policy makers attempting to predict and control an uncertain future. The book feels especially timely and relevant during the coronavirus epidemic. And yet, it would be a mistake to think radical uncertainty has finally been "captured in a bottle"; by its very nature, it cannot be captured nor its risks quantified.
Now, I may be biased: radical uncertainty and decision making under it are part of my own research agenda as well. But the authors really nail the story for the most part. They provide an exciting and broad ranging narrative that touches on economics, public policy, the history of mathematics and probability theory, war strategy, city planning, banking, the financial crisis of 2008, etc... The real life examples are both exciting and relevant. From real life politics to real life human drama, the book uses narrative explanations to drive home complex topics. But at the same time the book does not shy away from technical subjects either. The authors scrupulously trace the history of, and criticize the facile use of, statistical models, rational choice calculi, Homo economicus reasoning, expected utility theories, etc.. They argue that in the realm of complex phenomena, we have to stop relying on probabilistic reasoning and embrace "I don't know" as the new mantra. This leads to a new emphasis on evolutionary explanations, community based explanations, and models of bounded rationality. And it leads to a reappraisal of evolved traits, quirks and all, as useful guides to action. Our evolution has given us the ability to make seemingly irrational but adaptive choices so we should rely on them rather than academic models of "rationality" or "optimality" - and we definitely shouldn't try to blame people for not trusting experts or scientific planners.
While I do not have many substantive problems with the book, I do have two. One problem is a small one. The small problem is that the book ignores the insights on radical uncertainty by the Austrian school, especially F.A. Hayek, while simultaneously (and correctly) focusing on the insights of Knight and Keynes. This leads to a distorted view of the history of economics. And this also leads to a failure to see the market order as a coping strategy for radical uncertainty. Although the authors recognize the role of entrepreneurship as a mediator of radical uncertainty (mostly thanks to Knight's influence), they fail to connect the dots to Schumpeter and Hayek. For an Austrian perspective that makes a case for markets as a coping mechanism that deals with uncertainty beyond mere rational choice theory, see e.g. Mark Pennington's excellent Robust Political Economy: Classical Liberalism and the Future of Public Policy. However, this is only a small problem as the authors do not lean heavily into any particular sole author.
The other problem with the book is more serious. For in their desire to undermine the hegemonic dominance of probabilistic reasoning, the authors go too far in the other direction. In effect, they throw out the mathematical baby with the probabilistic bath water. This leads to some implausible and unwarranted conclusions. In particular, they make evolutionary arguments against the adaptive value of mathematical computation. Their argument, to paraphrase, is that if probabilistic computation had adaptive value, evolution would have selected for it. This argument is central to the author's anti-computation world view. In the first instance, this neglect of the power of mathematics to cope with radical uncertainty might appear as only a minor blip in a book that contains so many different aspects. But I fear that it constitutes serious error. For this reason, I will dedicate the remainder of my review to explaining what is wrong with it.
There are three major problems with the anti-computational view. I will explain them in order:
1) Firstly, adaptation is backward looking, not forward looking. Just because something had adaptive (or survival) value in the past does not mean that it has adaptive (or survival) value in the future. A lot of our ancestral human behaviour might become maladaptive as social circumstances change. Arguably this is already true e.g. in our penchant for violence or craving for sugary foods, etc. Our biological makeup does not have enough time, or strong enough selection pressures, to change. Our technological and cultural world evolves faster than our biological makeup can and this leads to our psychological responses, including our decision making skills, lagging behind. Now, we might have good reasons to trust our evolved instincts more than most innovations since the former have millions of years of evolution, and proven survival value, to back them up. But we cannot assume that their adaptive value has remained constant given how rapidly the social world has been changing in the past few hundreds of years, accelerating to incredible heights. This means that our strategies and skills must change to adapt to the circumstances. If this is so, and we have every reason to assume it is, then probabilistic reasoning and mathematical computation, although counter-intuitive and alien to our natural instincts, may have adaptive value after all?
2) Secondly, it is simply not true that probabilistic computation has not been selected for by natural selection. Across all clades of the evolutionary tree, the phenotypical and neurochemical routines of animals and plants, i.e. their behaviour and mentation, can be modeled using mathematical functions that are very similar to computational algorithms. Computation is ubiquitous in the natural world; and a lot of it is mathematically precise. The same goes for human beings too: our motivational structure, neural network, psychological profile, behavioural responses, etc., can be computationally explained - not perfectly, of course, and certainly not yet. But everything points to the fact that sound mathematical models of human decision making under radical uncertainty can be developed, and these can be used to give insights to better decision making strategies. Relying on those aspects of our evolutionary ancestry that have given us flexibility and learning mechanisms to cope with a changing world is not incompatible with actively developing new and better ones. Maybe the best thing we can do is to try to correct for what is wrong with our inherited skill set while maintaining the basic evolutionary capacity for radically uncertain decision making?
3) Thirdly, mathematics has been constantly developing, and the authors downplay the generative capacity of mathematics. The new sciences of chaos theory, complexity theory, agent based modelling, etc., rely on advanced mathematical models that are perfectly capable of modelling radical uncertainty without resorting to narrative models. The same also applies for A.I. and the applied mathematics of computer programming. The authors fail to understand the evolutionary capacity of complex computational systems for generating complex adaptive strategies of survival. New models of computer evolution, just like biological evolution, can be powerful tools of coping with radical uncertainty - precisely because they rely on mathematics as the language of life. Again, the insight is that computation is an indispensable tool of coping with radical uncertainty.
The book is not perfect. I think that the anti-computational approach is less than satisfactory. Computation has a place in developing better decision making and policy approaches to radical uncertainty. Computer A.I. can be a valuable companion in the human quest to improve our adaptive capacities. However, the rest of the book is so wonderful and interesting, so I highly recommend it. There aren't enough books out there on this very important and timely topic. Now is the time to recognize radical uncertainty as a key factor of our economic and political life. It should lead us to recognize the failures of simplistic models. It should make us stop demanding the impossible of leaders who know very little and can control very little of the uncertain world they live in.
Now, I may be biased: radical uncertainty and decision making under it are part of my own research agenda as well. But the authors really nail the story for the most part. They provide an exciting and broad ranging narrative that touches on economics, public policy, the history of mathematics and probability theory, war strategy, city planning, banking, the financial crisis of 2008, etc... The real life examples are both exciting and relevant. From real life politics to real life human drama, the book uses narrative explanations to drive home complex topics. But at the same time the book does not shy away from technical subjects either. The authors scrupulously trace the history of, and criticize the facile use of, statistical models, rational choice calculi, Homo economicus reasoning, expected utility theories, etc.. They argue that in the realm of complex phenomena, we have to stop relying on probabilistic reasoning and embrace "I don't know" as the new mantra. This leads to a new emphasis on evolutionary explanations, community based explanations, and models of bounded rationality. And it leads to a reappraisal of evolved traits, quirks and all, as useful guides to action. Our evolution has given us the ability to make seemingly irrational but adaptive choices so we should rely on them rather than academic models of "rationality" or "optimality" - and we definitely shouldn't try to blame people for not trusting experts or scientific planners.
While I do not have many substantive problems with the book, I do have two. One problem is a small one. The small problem is that the book ignores the insights on radical uncertainty by the Austrian school, especially F.A. Hayek, while simultaneously (and correctly) focusing on the insights of Knight and Keynes. This leads to a distorted view of the history of economics. And this also leads to a failure to see the market order as a coping strategy for radical uncertainty. Although the authors recognize the role of entrepreneurship as a mediator of radical uncertainty (mostly thanks to Knight's influence), they fail to connect the dots to Schumpeter and Hayek. For an Austrian perspective that makes a case for markets as a coping mechanism that deals with uncertainty beyond mere rational choice theory, see e.g. Mark Pennington's excellent Robust Political Economy: Classical Liberalism and the Future of Public Policy. However, this is only a small problem as the authors do not lean heavily into any particular sole author.
The other problem with the book is more serious. For in their desire to undermine the hegemonic dominance of probabilistic reasoning, the authors go too far in the other direction. In effect, they throw out the mathematical baby with the probabilistic bath water. This leads to some implausible and unwarranted conclusions. In particular, they make evolutionary arguments against the adaptive value of mathematical computation. Their argument, to paraphrase, is that if probabilistic computation had adaptive value, evolution would have selected for it. This argument is central to the author's anti-computation world view. In the first instance, this neglect of the power of mathematics to cope with radical uncertainty might appear as only a minor blip in a book that contains so many different aspects. But I fear that it constitutes serious error. For this reason, I will dedicate the remainder of my review to explaining what is wrong with it.
There are three major problems with the anti-computational view. I will explain them in order:
1) Firstly, adaptation is backward looking, not forward looking. Just because something had adaptive (or survival) value in the past does not mean that it has adaptive (or survival) value in the future. A lot of our ancestral human behaviour might become maladaptive as social circumstances change. Arguably this is already true e.g. in our penchant for violence or craving for sugary foods, etc. Our biological makeup does not have enough time, or strong enough selection pressures, to change. Our technological and cultural world evolves faster than our biological makeup can and this leads to our psychological responses, including our decision making skills, lagging behind. Now, we might have good reasons to trust our evolved instincts more than most innovations since the former have millions of years of evolution, and proven survival value, to back them up. But we cannot assume that their adaptive value has remained constant given how rapidly the social world has been changing in the past few hundreds of years, accelerating to incredible heights. This means that our strategies and skills must change to adapt to the circumstances. If this is so, and we have every reason to assume it is, then probabilistic reasoning and mathematical computation, although counter-intuitive and alien to our natural instincts, may have adaptive value after all?
2) Secondly, it is simply not true that probabilistic computation has not been selected for by natural selection. Across all clades of the evolutionary tree, the phenotypical and neurochemical routines of animals and plants, i.e. their behaviour and mentation, can be modeled using mathematical functions that are very similar to computational algorithms. Computation is ubiquitous in the natural world; and a lot of it is mathematically precise. The same goes for human beings too: our motivational structure, neural network, psychological profile, behavioural responses, etc., can be computationally explained - not perfectly, of course, and certainly not yet. But everything points to the fact that sound mathematical models of human decision making under radical uncertainty can be developed, and these can be used to give insights to better decision making strategies. Relying on those aspects of our evolutionary ancestry that have given us flexibility and learning mechanisms to cope with a changing world is not incompatible with actively developing new and better ones. Maybe the best thing we can do is to try to correct for what is wrong with our inherited skill set while maintaining the basic evolutionary capacity for radically uncertain decision making?
3) Thirdly, mathematics has been constantly developing, and the authors downplay the generative capacity of mathematics. The new sciences of chaos theory, complexity theory, agent based modelling, etc., rely on advanced mathematical models that are perfectly capable of modelling radical uncertainty without resorting to narrative models. The same also applies for A.I. and the applied mathematics of computer programming. The authors fail to understand the evolutionary capacity of complex computational systems for generating complex adaptive strategies of survival. New models of computer evolution, just like biological evolution, can be powerful tools of coping with radical uncertainty - precisely because they rely on mathematics as the language of life. Again, the insight is that computation is an indispensable tool of coping with radical uncertainty.
The book is not perfect. I think that the anti-computational approach is less than satisfactory. Computation has a place in developing better decision making and policy approaches to radical uncertainty. Computer A.I. can be a valuable companion in the human quest to improve our adaptive capacities. However, the rest of the book is so wonderful and interesting, so I highly recommend it. There aren't enough books out there on this very important and timely topic. Now is the time to recognize radical uncertainty as a key factor of our economic and political life. It should lead us to recognize the failures of simplistic models. It should make us stop demanding the impossible of leaders who know very little and can control very little of the uncertain world they live in.
May 24, 2020
I've spent some time this year thinking about contingencies in life - COVID has forced all of us to do that. I am not sure who recommended the book to me but I am glad whoever it was did.
The book's fundamental premise is that there is a profound difference between situations with definable probabilities (a roulette wheel is one such situation) and ones where about the best that we can do is to try to figure out what is going on. The authors are skeptical of those who assert that we can learn from modeling things where understanding the constraints of the problem are either poorly defined or unknown.
The book goes through a series of examples where the folly of such "predictive" models is well understood - starting with Thomas Malthus and extending through the Club of Rome and Paul Ehrlich and Ray Kurzweil's Singularity theory they chronicle the idiocy of those types of massive predictions. What is ludicrous about these efforts by promoters of these speculations suggest that a "distinguished" panel of "experts" cam gloss over what is unknown with a fancy model with 1000 variables.
Even in areas like financial projections there are numerous examples of experts missing the basic trends. The authors quote the then head of Barrings bank soon before it collapsed suggesting that the bank was in dandy shape.
I got some cautions from this book. First, all of us should be skeptical of consensus of experts on any complex issue, without first understanding whether the experts have any expert knowledge which can be applied to the question at hand. Second, understanding that a set of solutions to a complex problem poorly understood can be worse than doing nothing at all.
There is a joke in Economics (the authors are economists) which asks "Why do economists use decimal points? " The answer is "it shows greater precision." This book asks us to question the absolute assumption that we can safely project all things if we just think carefully about them. It is a good place to start but this book helps us consider how to approach complex problems from ones where the limits of our knowledge is a constraint.
The book's fundamental premise is that there is a profound difference between situations with definable probabilities (a roulette wheel is one such situation) and ones where about the best that we can do is to try to figure out what is going on. The authors are skeptical of those who assert that we can learn from modeling things where understanding the constraints of the problem are either poorly defined or unknown.
The book goes through a series of examples where the folly of such "predictive" models is well understood - starting with Thomas Malthus and extending through the Club of Rome and Paul Ehrlich and Ray Kurzweil's Singularity theory they chronicle the idiocy of those types of massive predictions. What is ludicrous about these efforts by promoters of these speculations suggest that a "distinguished" panel of "experts" cam gloss over what is unknown with a fancy model with 1000 variables.
Even in areas like financial projections there are numerous examples of experts missing the basic trends. The authors quote the then head of Barrings bank soon before it collapsed suggesting that the bank was in dandy shape.
I got some cautions from this book. First, all of us should be skeptical of consensus of experts on any complex issue, without first understanding whether the experts have any expert knowledge which can be applied to the question at hand. Second, understanding that a set of solutions to a complex problem poorly understood can be worse than doing nothing at all.
There is a joke in Economics (the authors are economists) which asks "Why do economists use decimal points? " The answer is "it shows greater precision." This book asks us to question the absolute assumption that we can safely project all things if we just think carefully about them. It is a good place to start but this book helps us consider how to approach complex problems from ones where the limits of our knowledge is a constraint.
June 7, 2020
I didn't rate this book so highly because I learned so much from it (most of it confirmed my existing views) on the topic, but because I think it should be read by anyone who uses (policymakers, military strategists, company executives) or produces (economists, actuaries, transport planners) forecasts involving events that are fundamentally uncertain. Really. Don't read reviews, just read the book.
April 13, 2021
A great book on radical uncertainty or deep uncertainty! It fits very well with my study in which model-baded policy analysis is key. I wonder why nobody at my faculty has recommended it yet! Well, I will definitely tell people to read it! And now that you have read this, feel addressed, get the book and get started!
April 6, 2021
I have to admit to struggling through this book. It's not that it's uninteresting, it's just that it needs a good edit for length. The authors take 400 pages to say what they could equally say in 150 to 200 pages. I would have preferred a more c0mpact statement of the argument rather than a discursive approach to the topic.
Why did I rad it all the way through? Because it has an important argument. One that needs to be embraced and absorbed into our daily lives. The argument is simple. Our knowledge of the future is incomplete. This lack of completeness reflects a future that is knowable and one that is unknowable. For the undiscerned aspects of the knowable future, we can assign risk probabilities. For the undiscerned aspects of the unknowable future, we can ascribe no objective uncertain probabilities, even is we assign subjective ones.
It is the pretence of objectivity in which our subjectivities is couched that gets us into trouble. When we ascribe a probability to an uncertain future, we are really saying that we have no idea what the outcome will be, but couching that statement with a degree of objectivity and precision. The book is about how to manage an uncertain world, and abandoning the pretence of objectivity and precision is the first start.
That path got us into all sorts of trouble in the global financial crisis. The certainty with which balance sheets were valued proved to be nowhere near the true value of bank balance sheets. Partly because the value of the underlying assets was not fully understood. Partly because of a collective delusion about the degree of certainty over the reliability of a sequence of valuations. And mainly because the certainty of our precise looking calculations appeared to be valid. Until they weren't. We were given a modern take on Andersen's Emperor's New Clothes.
Given that we know this, what can we do about it? The authors are less certain in this area. Apart from more general platitudes, such as 'be more resilient' or 'adopt more robustness', the reader is left wondering what to do. Obviously, we have to accept uncertainty. We have to be more aware of it as we encounter it in our lives. We have to develop a curiosity that allows us to ask 'what is going on here?' Beyond that, the reader is left without too much assistance.
Perhaps that was the core of my disappointment. After wading through a long exposition of the problem, when I reached the end of the book there were very few solutions on offer. I felt that part of the book could be reasonably strengthened. Because I felt that it was only dealt with lightly, that left me disappointed with the book.
It is, however, an important book for that reason. It does articulate the problem. The problem is not being addressed and it is one that we are likely to bump into again in the not too distant future. It can help us to see with a bit more clarity, even if it doesn't really help us in adapting to that change. I'm not sure how far I would recommend the book. Reading it represents quite an investment of time. It's not an easy read and it takes a bit of determination to make progress. After making that effort, many readers may be wondering, like me, if the effort was at all worth it?
Why did I rad it all the way through? Because it has an important argument. One that needs to be embraced and absorbed into our daily lives. The argument is simple. Our knowledge of the future is incomplete. This lack of completeness reflects a future that is knowable and one that is unknowable. For the undiscerned aspects of the knowable future, we can assign risk probabilities. For the undiscerned aspects of the unknowable future, we can ascribe no objective uncertain probabilities, even is we assign subjective ones.
It is the pretence of objectivity in which our subjectivities is couched that gets us into trouble. When we ascribe a probability to an uncertain future, we are really saying that we have no idea what the outcome will be, but couching that statement with a degree of objectivity and precision. The book is about how to manage an uncertain world, and abandoning the pretence of objectivity and precision is the first start.
That path got us into all sorts of trouble in the global financial crisis. The certainty with which balance sheets were valued proved to be nowhere near the true value of bank balance sheets. Partly because the value of the underlying assets was not fully understood. Partly because of a collective delusion about the degree of certainty over the reliability of a sequence of valuations. And mainly because the certainty of our precise looking calculations appeared to be valid. Until they weren't. We were given a modern take on Andersen's Emperor's New Clothes.
Given that we know this, what can we do about it? The authors are less certain in this area. Apart from more general platitudes, such as 'be more resilient' or 'adopt more robustness', the reader is left wondering what to do. Obviously, we have to accept uncertainty. We have to be more aware of it as we encounter it in our lives. We have to develop a curiosity that allows us to ask 'what is going on here?' Beyond that, the reader is left without too much assistance.
Perhaps that was the core of my disappointment. After wading through a long exposition of the problem, when I reached the end of the book there were very few solutions on offer. I felt that part of the book could be reasonably strengthened. Because I felt that it was only dealt with lightly, that left me disappointed with the book.
It is, however, an important book for that reason. It does articulate the problem. The problem is not being addressed and it is one that we are likely to bump into again in the not too distant future. It can help us to see with a bit more clarity, even if it doesn't really help us in adapting to that change. I'm not sure how far I would recommend the book. Reading it represents quite an investment of time. It's not an easy read and it takes a bit of determination to make progress. After making that effort, many readers may be wondering, like me, if the effort was at all worth it?
April 4, 2020
I am not finished, but I will say that the central thesis of this book is compelling. I am not convinced that the same couldn't have been accomplished in a shorter book, but I will have more to say when I am finished.
Having now finished, I will suggest that this is a good book and an important addition to the literature of probabilities, decision making and uncertainty. But there remains the problem of repetition. I understood the central thesis fifty pages in, and would have appreciated a shorter version.
It is now several days later and I am reading Frank Ramsey, A Sheer Excess of Powers, by Cheryl Misak. I came to this biography via references to Frank Ramsey in Radical Uncertainty. The time spent with Frank Ramsey is making me appreciate, even more, the work in Radical Uncertainty. I have a vague impression that the book is important and should be read by all sorts of people. I say vague because, after 67 years, I am still impressed at how many people spend their lives in an "unexamined" way. Can we really expect administrators, doctors, lawyers, politicians, etc., to take the time to understand "decision making beyond the numbers"? They should!
Having now finished, I will suggest that this is a good book and an important addition to the literature of probabilities, decision making and uncertainty. But there remains the problem of repetition. I understood the central thesis fifty pages in, and would have appreciated a shorter version.
It is now several days later and I am reading Frank Ramsey, A Sheer Excess of Powers, by Cheryl Misak. I came to this biography via references to Frank Ramsey in Radical Uncertainty. The time spent with Frank Ramsey is making me appreciate, even more, the work in Radical Uncertainty. I have a vague impression that the book is important and should be read by all sorts of people. I say vague because, after 67 years, I am still impressed at how many people spend their lives in an "unexamined" way. Can we really expect administrators, doctors, lawyers, politicians, etc., to take the time to understand "decision making beyond the numbers"? They should!
May 8, 2020
Kay and King do something quite remarkable here - they attempt to draw our (economists', policymakers' etc.) attention to the notion of what they call "radical uncertainty", which, with a bit of oversimplification, may stand for the inherent unpredictability of the future. While properly acknowledged, authors claim, the existence of radical uncertainty undermines the rationale for various quantitative predicions and models and ridicules the "pretence of knowledge" and yearning for certainty present not least in financial sector risk models. Furthermore, they masterfully show how, in light of the radical uncertainty, the modern microeconomic axioms of "rational" human behaviour are misguided and ultimately dangerous. However, they do not buy into the "biases" rhetoric of behavioural economics as well, showing that the former are useful heuristics that help us cope with the radical uncertainty. Masterful achievement, compulsory reading for economics students, practictioners, policymakers and everyone who is interested in forecasting and risk management.
June 10, 2021
The principle of ergodicity has led the field of economics astray. Contrary to how the book presents itself, Kay and King didn’t develop this criticism...Post-Keynesians have been articulating and debating the theory of uncertainty for decades, this is no secret. There is no mention of leading scholars and contributors, and for this reason I could not enjoy the text. The authors use interesting and engaging examples to illustrate their point, and this is why I brought myself to finish the book. The book doesn’t provide any meaningful solutions and feels like it is constantly repeating the same narrative (ironic) with slightly different examples.
Economics needs to look inwards and assess its errors. However, stealing the basics ideas of others is not how we do that.
Economics needs to look inwards and assess its errors. However, stealing the basics ideas of others is not how we do that.
April 18, 2021
no thank you
November 29, 2020
I luckily stumbled upon this book title's during the time when I was thinking of buying an medical insurance and I needed help in taking decision. I could not make my mind properly since I already have one from my employer for a fixed amount, should I go ahead buy an additional one personally to cover any expenditures exceeding the amount given by my employer?
If I were to go ahead and buy the medical insurance, is that a good decision or am I just falling into the fear created by advertisement scams or am I surrendering to my natural human fallibility. What percentage would I be right and how much can I go wrong?
I also needed to buy one for my parents. My father is a alcoholic and heavy smoker and he is retired and he is at a stage if something were to happen to him, he would be readily give up and would not wish for me to run helter-skelter and spend my savings on him. Then is medical insurance worth for him? I need to buy one for my mother, she takes good care of her health and a missing insurance is something I would not want to regret about if something were to happen.
Which insurance should I buy, there are many companies with options branching out like a maze and with the tiny but lengthy terms and conditions. Will there be any customer care hassles on the one I would be choosing? Will I ever fully understand on what grounds can claims be rejected before selecting one? Will I select the one that I just require or will I give in to the sales persons selling extra items that I would never require. How do I keep track of all my investments, insurance, bank balance, save myself internet frauds, oh and the numerous sms and emails that I am bound to receive once I sign up for an medical insurance. Should I buy an insurance for my house, water logging is common every monsoon. Is there an option that covers my house and medical insurance both under same policy?
I was getting overwhelmed.
Reading this book helped me gain a perspective on how we tend to choose the option that we think has maximum utility (our own calculations could be flawed in those calculations!) compared to the options that are good enough for survival and good Enough is sufficient. Many good enough decisions have brought the human race so far.
The book is not only about this but there are few examples and explanations about human behavior, probability, mathematics, origins of insurance industry, economic theory, what are the choices did Obama have before deciding to go-ahead with the operation on the suspected Osama hideout. What percentage was he or the military sure about Osama's presence and was that percentage good enough to go ahead?
There is an interesting argument to a human behavior that Daniel Kahneman described in Thinking Fast and Slow. I had consumed Thinking Fast and Slow agreeing with all along. What a mistake! I gained worthy insights. I have highlighted many interesting portions that I would want to re-read later but during my first read I was only looking for what I wanted to read. I look forward to the near future when I re-visit the highlighted portions to get additional insights on other topics.
There are many references of certain economic theories, which would require more effort to grasp this book completely but a good enough grasp works equally well :p
If I were to go ahead and buy the medical insurance, is that a good decision or am I just falling into the fear created by advertisement scams or am I surrendering to my natural human fallibility. What percentage would I be right and how much can I go wrong?
I also needed to buy one for my parents. My father is a alcoholic and heavy smoker and he is retired and he is at a stage if something were to happen to him, he would be readily give up and would not wish for me to run helter-skelter and spend my savings on him. Then is medical insurance worth for him? I need to buy one for my mother, she takes good care of her health and a missing insurance is something I would not want to regret about if something were to happen.
Which insurance should I buy, there are many companies with options branching out like a maze and with the tiny but lengthy terms and conditions. Will there be any customer care hassles on the one I would be choosing? Will I ever fully understand on what grounds can claims be rejected before selecting one? Will I select the one that I just require or will I give in to the sales persons selling extra items that I would never require. How do I keep track of all my investments, insurance, bank balance, save myself internet frauds, oh and the numerous sms and emails that I am bound to receive once I sign up for an medical insurance. Should I buy an insurance for my house, water logging is common every monsoon. Is there an option that covers my house and medical insurance both under same policy?
I was getting overwhelmed.
Reading this book helped me gain a perspective on how we tend to choose the option that we think has maximum utility (our own calculations could be flawed in those calculations!) compared to the options that are good enough for survival and good Enough is sufficient. Many good enough decisions have brought the human race so far.
The book is not only about this but there are few examples and explanations about human behavior, probability, mathematics, origins of insurance industry, economic theory, what are the choices did Obama have before deciding to go-ahead with the operation on the suspected Osama hideout. What percentage was he or the military sure about Osama's presence and was that percentage good enough to go ahead?
There is an interesting argument to a human behavior that Daniel Kahneman described in Thinking Fast and Slow. I had consumed Thinking Fast and Slow agreeing with all along. What a mistake! I gained worthy insights. I have highlighted many interesting portions that I would want to re-read later but during my first read I was only looking for what I wanted to read. I look forward to the near future when I re-visit the highlighted portions to get additional insights on other topics.
There are many references of certain economic theories, which would require more effort to grasp this book completely but a good enough grasp works equally well :p
November 30, 2020
Radical Uncertainty covers a lot of what is now commonly trodden ground in the behavioral economics genre, so I would not consider the conclusions it draws to be especially novel. However, Kay's exposition of the ideas is one of the best I have seen, and it essentially comes down to this: you cannot model or assign probabilities to poorly defined problems.
The more reports and articles you read with any form of analysis, the more you will encounter claims of; x% chance of the next GFC, y% of the next 9/11 or z% of the next covid-19. In many regards, the rise in statistical analysis across all factions of society has been a positive thing. However, unprecedented increases in data available has lead to a demand for putting a number on everything. The result has been a fetishization of mathematical models. Radical Uncertainty aptly presents the limits inherent in modelling events that are vaguely defined, dynamic and complex - such as those listed above. You can endlessly argue what is defined as 'the next economic crisis', the number of variables are innumerable and prophylactic moves can alter outcomes. To use one of Kay's phrases, uncertain events tend to be defined by non-stationary statistical distributions. That is, probabilities of events are in a constant flux, as opposed to a coin flip which is constant at 50/50. Further complicating things, data and assumptions used in models are selected by people. A key example of this is in the decision of prior probabilities in Bayesian models. Bayesian reasoning is sensible in a well-defined problem, and humans choosing inputs is not in itself problematic. But for uncertain events, different reasoning leads to radically different priors, turning it into a political exercise.
Another probabilistic illiteracy Kay has contention with is the use of confidence percentages in conversation. For example, someone declaring, "I am 80% sure that Philadelphia is the capital of Pennsylvania". As Kay notes, this statement of confidence does not have the clear and objective meaning in the frequentist sense of drawing outcomes from a probability distribution. Either it is the capital or it is not (it isn't). He concludes that that this type of statement is only useful for ranking - you believe one event is more likely than another, but, probabilistically, assigning a percentage makes no sense.
Radical Uncertainty has a range of further interesting topics covered to those I have mentioned. It could certainly have been edited down without losing its substance. However, I believe the main essence of the content is important and accessible to anyone, with no statistical background being required. 5 stars.
The more reports and articles you read with any form of analysis, the more you will encounter claims of; x% chance of the next GFC, y% of the next 9/11 or z% of the next covid-19. In many regards, the rise in statistical analysis across all factions of society has been a positive thing. However, unprecedented increases in data available has lead to a demand for putting a number on everything. The result has been a fetishization of mathematical models. Radical Uncertainty aptly presents the limits inherent in modelling events that are vaguely defined, dynamic and complex - such as those listed above. You can endlessly argue what is defined as 'the next economic crisis', the number of variables are innumerable and prophylactic moves can alter outcomes. To use one of Kay's phrases, uncertain events tend to be defined by non-stationary statistical distributions. That is, probabilities of events are in a constant flux, as opposed to a coin flip which is constant at 50/50. Further complicating things, data and assumptions used in models are selected by people. A key example of this is in the decision of prior probabilities in Bayesian models. Bayesian reasoning is sensible in a well-defined problem, and humans choosing inputs is not in itself problematic. But for uncertain events, different reasoning leads to radically different priors, turning it into a political exercise.
Another probabilistic illiteracy Kay has contention with is the use of confidence percentages in conversation. For example, someone declaring, "I am 80% sure that Philadelphia is the capital of Pennsylvania". As Kay notes, this statement of confidence does not have the clear and objective meaning in the frequentist sense of drawing outcomes from a probability distribution. Either it is the capital or it is not (it isn't). He concludes that that this type of statement is only useful for ranking - you believe one event is more likely than another, but, probabilistically, assigning a percentage makes no sense.
Radical Uncertainty has a range of further interesting topics covered to those I have mentioned. It could certainly have been edited down without losing its substance. However, I believe the main essence of the content is important and accessible to anyone, with no statistical background being required. 5 stars.
August 3, 2021
Although at times the text can be circuitous, its premise is logical and legitimate, and the authors' expertise, intellect and engaging writing style equips them to deliver a clear and powerful message regarding the role of uncertainty in Economics.
The essence of the book is that the Economics profession’s fixation with quantitative methods and modelling (which they refer as "mathiness”), even when the underlying assumptions are plainly unrealistic, has led to overstretched theories detached from reality that end up having real effects, such as to the market collapse of 2007-2008. Instead, Economics would do well to focus on the question of “what is going on here?”. The idea is that every individual, community, or organisation has a personal narrative: a story they are trying to pursue. Assessing “what is going?” on simply refers to the process of collecting information, analysing how this new information influences our dream, and adapting accordingly.
The authors begin by distinguishing risk (a lack of knowledge that can be parameterised) and uncertainty (a lack of knowledge that we cannot parameterise and about which we know little). The study of risk provides us with a variety of disciplines linked to statistical decision making and problem solving. The study of uncertainty is just as important (or more) to decision-making but also much less tractable to statistical analysis. Sometimes admitting that we simply do not know can be the best thing we can do.
The authors then take the reader on a voyage of finance, behavioural economics, politics, policy studies, and decision-making fields to demonstrate the fallaciousness of the majority of the contemporary economic and financial modelling, the potential gravity of its consequences, and what is possible to achieve if we were to embrace radical uncertainty. Their critiques challenge the pretentiousness of social scientists and policymakers in their certainty to be able to predict and control an uncertain future, which is rather fitting in the current coronavirus epidemic.
Given the rapid rise of computing power, the direction Economics is going in, and the current socioeconomic climate, this is certainly a book worth reading. My only major criticism is that the book could have been 100 pages shorter without losing depth, but I suppose academics will always be academics.
The essence of the book is that the Economics profession’s fixation with quantitative methods and modelling (which they refer as "mathiness”), even when the underlying assumptions are plainly unrealistic, has led to overstretched theories detached from reality that end up having real effects, such as to the market collapse of 2007-2008. Instead, Economics would do well to focus on the question of “what is going on here?”. The idea is that every individual, community, or organisation has a personal narrative: a story they are trying to pursue. Assessing “what is going?” on simply refers to the process of collecting information, analysing how this new information influences our dream, and adapting accordingly.
The authors begin by distinguishing risk (a lack of knowledge that can be parameterised) and uncertainty (a lack of knowledge that we cannot parameterise and about which we know little). The study of risk provides us with a variety of disciplines linked to statistical decision making and problem solving. The study of uncertainty is just as important (or more) to decision-making but also much less tractable to statistical analysis. Sometimes admitting that we simply do not know can be the best thing we can do.
The authors then take the reader on a voyage of finance, behavioural economics, politics, policy studies, and decision-making fields to demonstrate the fallaciousness of the majority of the contemporary economic and financial modelling, the potential gravity of its consequences, and what is possible to achieve if we were to embrace radical uncertainty. Their critiques challenge the pretentiousness of social scientists and policymakers in their certainty to be able to predict and control an uncertain future, which is rather fitting in the current coronavirus epidemic.
Given the rapid rise of computing power, the direction Economics is going in, and the current socioeconomic climate, this is certainly a book worth reading. My only major criticism is that the book could have been 100 pages shorter without losing depth, but I suppose academics will always be academics.
March 1, 2021
Mervyn King y John Kay presentan un gran libro. Una lectura muy recomendada, para economistas y científicos sociales en general es una llamada al conocimiento práctico, a pensar en explicaciones no en axiomas. Pará tomadores de decisiones es lectura obligatoria.
Intenta llevar las ideas de Taleb y su incerto al lenguaje de economistas y tomadores de decisiones para que entiendan lo poco que sabemos y por lo tanto lo frágil de nuestro entendimiento y el enorme rol de la incertidumbre en muchos aspectos de la vida.
Es un llamado a dejar de pensar exclusivamente en modelos, si usarlos como herramientas, pero entender sus límites. Pensar abductivamente, preferir lo que funciona a lo que nos gustaría. Y darle valor a las narrativas como herramientas de toma de decisiones, la heuristica.
Las personas, ni la naturaleza buscan optimizar, buscan resolver problemas, buscan la solución que funciona no "la mejor". Es un libro que en general tiene el propósito de decirnos que no debemos engañarnos a nosotros mismos y que somos la persona más fácil de engañar.
Es un libro que nos recuerda que lo que tenemos que formar es un buen juicio y eso requiere narrativas robustas sobre lo que puede salir mal, sobre como los planes pueden no salir como planeamos, pero sabiendo eso, que el mundo no es estacionario podemos prepararnos tomando decisiones que nos presenten un mayor numero de opciones y evitando aquellas alternativas que las reducen.
Intenta llevar las ideas de Taleb y su incerto al lenguaje de economistas y tomadores de decisiones para que entiendan lo poco que sabemos y por lo tanto lo frágil de nuestro entendimiento y el enorme rol de la incertidumbre en muchos aspectos de la vida.
Es un llamado a dejar de pensar exclusivamente en modelos, si usarlos como herramientas, pero entender sus límites. Pensar abductivamente, preferir lo que funciona a lo que nos gustaría. Y darle valor a las narrativas como herramientas de toma de decisiones, la heuristica.
Las personas, ni la naturaleza buscan optimizar, buscan resolver problemas, buscan la solución que funciona no "la mejor". Es un libro que en general tiene el propósito de decirnos que no debemos engañarnos a nosotros mismos y que somos la persona más fácil de engañar.
Es un libro que nos recuerda que lo que tenemos que formar es un buen juicio y eso requiere narrativas robustas sobre lo que puede salir mal, sobre como los planes pueden no salir como planeamos, pero sabiendo eso, que el mundo no es estacionario podemos prepararnos tomando decisiones que nos presenten un mayor numero de opciones y evitando aquellas alternativas que las reducen.
March 31, 2022
Fantastic work comprehensively summarising the history of probability theory intertwined with discourse into how probabilities and uncertainty are inter-related with other disciplines. Written in a very accessible style, it is an excellent "textbook" into uncertainty for anyone, including those whose probabilities expertise is limited to high school knowledge.
My rating for this book is 4 stars, as though I found the reading of multiple historic examples of what the authors consider good decision making fascinating, I was missing the actual toolkit (or an approximation of such) of how to approach decision making in radically uncertain present or future (which the cover suggested the book will present). Also I found it at times quite repetitive. Key message that the authors deliver over and over again in various iterations is: Real world problems are radically uncertain and any kind of attempts to model it will never be perfect; so decision making should be based, first of all, on narrative derived from a specific socio-economic context, as humans are a product of collective intelligence and social norms - numbers and models are mostly only useful to support such narrative.
On balance, I would still read it again and I recommend it to others as a very sobering and sarcastic critique of our over-reliance on models and theories instead of focusing on finding out "What is going on here?".
My rating for this book is 4 stars, as though I found the reading of multiple historic examples of what the authors consider good decision making fascinating, I was missing the actual toolkit (or an approximation of such) of how to approach decision making in radically uncertain present or future (which the cover suggested the book will present). Also I found it at times quite repetitive. Key message that the authors deliver over and over again in various iterations is: Real world problems are radically uncertain and any kind of attempts to model it will never be perfect; so decision making should be based, first of all, on narrative derived from a specific socio-economic context, as humans are a product of collective intelligence and social norms - numbers and models are mostly only useful to support such narrative.
On balance, I would still read it again and I recommend it to others as a very sobering and sarcastic critique of our over-reliance on models and theories instead of focusing on finding out "What is going on here?".
February 24, 2021
Very interesting approach to discussing decision making under uncertainty. Would highly recommend.
November 4, 2020
One thing that's become to clear to me in the Age of Coronavirus is just how uncomfortable most people are with uncertainty; more than once I've had friends and acquaintances evince exasperation at conflicting opinions among public health experts. Along similar lines, anyone making any kind of meaningful business decision knows how much guessing and hoping goes into one. In Econ 101 it might appear trivial to set the right price...in the real world it is anything but. So I was eager to see what John Kay (Dean of Oxford's Business School and columnist for the Financial Times for 25 years) and Mervyn King (the head of the Bank of England during the Global Financial Crisis of 2008) had to say about uncertainty.
I generally prefer concision when reading these kinds of books. But when I set down to read Radical Uncertainty I made my peace with what promised to be a long, discursive read: it is over 500 pages long and it was clear from the beginning that this was going to be the kind of book where the authors include all kinds of tangential anecdotes, histories, and stories.
Unfortunately, like some other reviewers I found it quite repetitive. Their main point is made in the Preface:
They also point out in the Preface that "normal" people (as opposed to economists) find their point obvious and may "thinks we are flogging a dead horse" but that "We hope that general readers will nevertheless enjoy the spectacle of the flogging".
I was reminded of the phrase I learned from Andrew Lo's Adaptive Markets: Financial Evolution at the Speed of Thought: "it takes a theory to beat a theory". I think that is what is lacking here. They continually remind us that typical economic conceptions of decision-making are, on the surface, silly. The closest they come is their concept of a "reference narrative", which seems similar (identical?) to Robert Shiller's "narrative economics", but even that is a fairly vague and ill-defined concept.
But I'm not sure I needed to see that message repeated over and over again for hundreds of pages. It did indeed feel like they were flogging a dead horse -- not something I really needed to see done at such length. You may enjoy this book more than I did it enjoy their multitude of examples: Annie Duke the poker champion, the HMS Gloucester shooting down an Iraqi Silkworm missile aimed at the ship in 1991, the Bay of Pigs invasion, Robert McNara's handling of the Vietnam War, the collapse of Northern Rock in 2008, Peak Oil claims from the 1950s, Malthus's population doom predictions, and many more. The vast array is somewhat fascinating and a testament to the author's broad knowledge but I'm less sure that the sheer number of examples really deepened things substantially for me.
It is a shame that I found their book so long-winded because I think that I basically agree with nearly everything they write. Look at the headings of their final chapter on adapting to radical uncertainty: "Non-stationarity" (that is, the world is full of shocks and abrupt shifts), "Humans are social animals" (not solitary rational machines), "The importance of narrative", "Challenging narratives" (how to avoid groupthink and yes-men), "Collective intelligence and communicative rationality" (we make better decisions in groups). All of those sound like good maxims! But they strike me as a bit too trite, especially as the culmination of 500+ pages on "decision-making behind the numbers".
I generally prefer concision when reading these kinds of books. But when I set down to read Radical Uncertainty I made my peace with what promised to be a long, discursive read: it is over 500 pages long and it was clear from the beginning that this was going to be the kind of book where the authors include all kinds of tangential anecdotes, histories, and stories.
Unfortunately, like some other reviewers I found it quite repetitive. Their main point is made in the Preface:
Business people, policy-makers and families could not even imagine having the information needed to determine the actions that would maximise shareholder value, social welfare or household utility. Or to know whether they had succeeded in doing so after the event.
They also point out in the Preface that "normal" people (as opposed to economists) find their point obvious and may "thinks we are flogging a dead horse" but that "We hope that general readers will nevertheless enjoy the spectacle of the flogging".
I was reminded of the phrase I learned from Andrew Lo's Adaptive Markets: Financial Evolution at the Speed of Thought: "it takes a theory to beat a theory". I think that is what is lacking here. They continually remind us that typical economic conceptions of decision-making are, on the surface, silly. The closest they come is their concept of a "reference narrative", which seems similar (identical?) to Robert Shiller's "narrative economics", but even that is a fairly vague and ill-defined concept.
Real households, real businesses and real governments do not optimise; they cope. They make decisions incrementally. They do not attain the highest point on the landscape, they seek only a higher place than the one they occupy now.
But I'm not sure I needed to see that message repeated over and over again for hundreds of pages. It did indeed feel like they were flogging a dead horse -- not something I really needed to see done at such length. You may enjoy this book more than I did it enjoy their multitude of examples: Annie Duke the poker champion, the HMS Gloucester shooting down an Iraqi Silkworm missile aimed at the ship in 1991, the Bay of Pigs invasion, Robert McNara's handling of the Vietnam War, the collapse of Northern Rock in 2008, Peak Oil claims from the 1950s, Malthus's population doom predictions, and many more. The vast array is somewhat fascinating and a testament to the author's broad knowledge but I'm less sure that the sheer number of examples really deepened things substantially for me.
If neither Magnus Carlsen (in 2019 the world champion) nor Deep Blue can play a perfect game of chess, it stretches the imagination to suppose that ordinary people and businesses could optimise the game of economic life.
It is a shame that I found their book so long-winded because I think that I basically agree with nearly everything they write. Look at the headings of their final chapter on adapting to radical uncertainty: "Non-stationarity" (that is, the world is full of shocks and abrupt shifts), "Humans are social animals" (not solitary rational machines), "The importance of narrative", "Challenging narratives" (how to avoid groupthink and yes-men), "Collective intelligence and communicative rationality" (we make better decisions in groups). All of those sound like good maxims! But they strike me as a bit too trite, especially as the culmination of 500+ pages on "decision-making behind the numbers".
September 13, 2020
Economists have two major shortcomings. The first one is addressed in this book, the second one is demonstrated by it.
First of all, there is the unnerving urge to stick probabilities and mathematics onto everything. It was something that mesmerized my when I was a student. During my undergraduate years I enjoyed the apparent beauty of formulating a target function, adding constraints and optimizing it using appropriate techniques. The outcomes from this exercise magically explained the world as we know it. Or so we thought. Later, as a graduate student and especially during my brief stint as a Ph.D.-candidate, I came to dread the shortsightedness of most people in the field. They were all formulating as many hypotheses as possible to simply capture human behaviour in a number, putting all their faith in the significance of the results of linear regressions, without ever questioning their relevance. Often times they had even forgotten the reasoning behind a mathematical method and applied it almost mechanically. The result is, as would be expected by anyone not too closely involved, a lot of effort being wasted on working through the maths while too little time and attention are spent on actually asking the right questions (which in the book is called 'What is actually going on here') or explaining what is really happening. This is a point that the book continuously makes, and rightly so. It is high time the profession recognized its shortcomings and the fact that it is slowly sliding towards irrelevance.
Secondly - and more fundamentally - there is very little of real interest that economists have to tell. The 'deep insights' are no more than theoretical constructs that bear little relevance to reality and are only meaningful within the context of some arcane model. Kay and King suffer from the same disease. Their core idea is good - it even made me buy this book - but they spend far too long on repeating it over and over again. The authors drone on for close to 350 pages (plus an annex) without adding much in terms of ideas or development of their thesis. The same anecdotes return all the time and each chapter feels as if it is simply rephrasing the previous one. This is disappointing - based on the synopsis and the reputation of the authors, I had hoped for more insights, thoughtful reasoning and convincing arguments. True, there are sections - especially the ones about the financial crisis, its origins and its aftermath - that offered truly interesing historical (first-hand) accounts of what had actually happened. They are however limited to no more than maybe twenty pages before we are back to yet another round of stories about the Bayesian Dial, Barack Obama ordering the strike on Osama Bin Laden or the musings of some scholars during an obscure dinner event. I had expected far more enlightenment than this from people who spent a lifetime working at the forefront of public policy .
All in all the book is a disappointment. Both intellectually (the dearth of ideas) and textually (it simply is far too long). I hesitated between two or three stars; and settled for the latter because of the interesting historical passages and the merit of the initial thesis. If the authors had developed it in a ten or twenty page manuscript it could have been a seminal text; now it feels as another run-of-the-mill ghost written book that tries to capitalize on the reputation of an author. A pity.
First of all, there is the unnerving urge to stick probabilities and mathematics onto everything. It was something that mesmerized my when I was a student. During my undergraduate years I enjoyed the apparent beauty of formulating a target function, adding constraints and optimizing it using appropriate techniques. The outcomes from this exercise magically explained the world as we know it. Or so we thought. Later, as a graduate student and especially during my brief stint as a Ph.D.-candidate, I came to dread the shortsightedness of most people in the field. They were all formulating as many hypotheses as possible to simply capture human behaviour in a number, putting all their faith in the significance of the results of linear regressions, without ever questioning their relevance. Often times they had even forgotten the reasoning behind a mathematical method and applied it almost mechanically. The result is, as would be expected by anyone not too closely involved, a lot of effort being wasted on working through the maths while too little time and attention are spent on actually asking the right questions (which in the book is called 'What is actually going on here') or explaining what is really happening. This is a point that the book continuously makes, and rightly so. It is high time the profession recognized its shortcomings and the fact that it is slowly sliding towards irrelevance.
Secondly - and more fundamentally - there is very little of real interest that economists have to tell. The 'deep insights' are no more than theoretical constructs that bear little relevance to reality and are only meaningful within the context of some arcane model. Kay and King suffer from the same disease. Their core idea is good - it even made me buy this book - but they spend far too long on repeating it over and over again. The authors drone on for close to 350 pages (plus an annex) without adding much in terms of ideas or development of their thesis. The same anecdotes return all the time and each chapter feels as if it is simply rephrasing the previous one. This is disappointing - based on the synopsis and the reputation of the authors, I had hoped for more insights, thoughtful reasoning and convincing arguments. True, there are sections - especially the ones about the financial crisis, its origins and its aftermath - that offered truly interesing historical (first-hand) accounts of what had actually happened. They are however limited to no more than maybe twenty pages before we are back to yet another round of stories about the Bayesian Dial, Barack Obama ordering the strike on Osama Bin Laden or the musings of some scholars during an obscure dinner event. I had expected far more enlightenment than this from people who spent a lifetime working at the forefront of public policy .
All in all the book is a disappointment. Both intellectually (the dearth of ideas) and textually (it simply is far too long). I hesitated between two or three stars; and settled for the latter because of the interesting historical passages and the merit of the initial thesis. If the authors had developed it in a ten or twenty page manuscript it could have been a seminal text; now it feels as another run-of-the-mill ghost written book that tries to capitalize on the reputation of an author. A pity.
June 22, 2020
As a member of the generation of economics students nurtured on Savage, Friedman, Lucas, et al., reading books like this excellent work elicits gasps of “oh, the wasted hours”, “that’s what I argued when I almost failed the test,” and the like.
Decision making under uncertainty based on axiomatic reasoning may have uses in “small world” (class room) situations as a pedagogical tool, but as the authors ably demonstrate - the reality (of the fundamental uncertainty) of our complex and dynamic world where people confront unique events and process their reactions through personal biases defy even the most clever thinkers to simplify the workings of the economy to a set of rules or evaluate decisions through (recklessly) subjective probabilities.
This well written and reasoned book is well worth the time and effort.
Decision making under uncertainty based on axiomatic reasoning may have uses in “small world” (class room) situations as a pedagogical tool, but as the authors ably demonstrate - the reality (of the fundamental uncertainty) of our complex and dynamic world where people confront unique events and process their reactions through personal biases defy even the most clever thinkers to simplify the workings of the economy to a set of rules or evaluate decisions through (recklessly) subjective probabilities.
This well written and reasoned book is well worth the time and effort.
October 13, 2020
Not everything can be neatly quantified and sometimes we just have to deal with less than perfect information.
The point at the heart of this book is a good one, if a lot less radical than the authors imagine, and if the book had stopped after part 2, it would have been really good.
Unfortunately, it doesn't. What we get instead is a series of anecdotes, some of which manage to actually undermine the authors' core point, and which are repeated and restated to the point that I utterly lost the will to continue.
The point at the heart of this book is a good one, if a lot less radical than the authors imagine, and if the book had stopped after part 2, it would have been really good.
Unfortunately, it doesn't. What we get instead is a series of anecdotes, some of which manage to actually undermine the authors' core point, and which are repeated and restated to the point that I utterly lost the will to continue.
December 13, 2020
If you can hold the paradox, it's a good book on many counts but I did not enjoy it much.
What was good: Wide-spanning account of uncertainty from many perspectives (history, science, models, human cognition, economics)
What I did not enjoy much: if you read Gigerenzer, Klein, Taleb, Tetclok, Snowden, and others, there is really not much new (for me) in this book, but it took me through 433 pages of "not much new" and many anecdotes. Valuable read, still, depending on what else have you read on the subject.
3.5/5 for me
What was good: Wide-spanning account of uncertainty from many perspectives (history, science, models, human cognition, economics)
What I did not enjoy much: if you read Gigerenzer, Klein, Taleb, Tetclok, Snowden, and others, there is really not much new (for me) in this book, but it took me through 433 pages of "not much new" and many anecdotes. Valuable read, still, depending on what else have you read on the subject.
3.5/5 for me
December 13, 2020
This was probably the most thought-provoking book I've read this year.
When I say that, I don't mean it was the kind of book that made me reconsider everything I believed, though it did some of that. Rather, nearly every page had me pausing to consider and mentally debate the book's arguments, which were always interesting even when I ultimately came down against the authors. (Of course, I have a larger-than-normal interest in epistemology and uncertainty, so your mileage may vary...)
The main point of the authors if to respond to an intellectual movement that claims to be foregrounding uncertainty in their analysis. This movement, inspired by Bayesian statistics, tries to get away from narrative-driven beliefs by instead quantifying the likelihood of those beliefs, and constantly updating those likelihoods in response to new information. The authors argue this movement is misguided, and actually does the opposite of what its backers claim it does — by quantifying risk, they say, people are applying false precision to things that are actually radically uncertain — impossible to quantify.
"Reasonable uncertainty is uncertainty which can be removed by looking something up (I am uncertain which city is the capital of Pennsylvania) or which can be represented by a known probability distribution of outcomes (the spin of a roulette wheel). With radical uncertainty, however, there is no similar means of resolving the uncertainty — we simply do not know."
Put another way, it is the difference between "risk" and "uncertainty" where "risk" means "unknowns which could be describe with probabilities" and "uncertainty" which can't. Today, the authors argue, we tend to treat uncertain things as if they are actually risks that can be precisely quantified. They give the example of national security advisers meeting with President Barack Obama in 2011, giving their assessments of whether Osama bin Laden was actually in a compound in Abbottabad, Pakistan — one advisor said there was a 95 percent chance bin Laden was there, while another said 80 percent and another 40 percent. Obviously these percentages are completely different from, say, the 50 percent chance that a fair coin will come up heads.
But this example also brings up one of the problems with the book: it's rather over-focused on issues inside the field of economics (and adjacent areas), and the author's arguments against various forms of probabilistic reasoning run into more issues when they move past critiquing over-quantified economic models and move to day-to-day decision-making.
To return to the prior example, in a very literal sense, the statement that there was a 95 percent chance Bin Laden was in Abbottabad is meaningless. Either he was there or he wasn't; it wasn't like you could raid the compound 20 times and expect to find Bin Laden 19 times. But this wasn't a case where the only options for belief were "he's there," "he's not there," and "we don't know." One can believe it is "more likely than not" that something is true, that evidence suggests something but doesn't prove it. Saying "95 percent" may not have any solid statistical basis, but isn't it a perfectly fine synonym for "almost certain"? To be sure, we need to make sure not to take that 95 percent estimate too seriously, as a real, empirical probability. But at a certain point, applied to real life and not economic models, this argument becomes a straw man.
Another favorite straw man argument the authors use is to mock the idea that actual people making real decisions have a "Bayesian dial floating over their heads" — a reference to the model of Bayesian statistics, which starts with a "prior probability" that something is true and then updates that probability based on evidence. Real decisions, they say, are based on narratives, not statistical models. Again, this is an argument that is obviously true in a very narrow sense — as they demonstrate, even professional economists and statisticians usually don't use their probabilistic methods for making life decisions — but falls down a bit when taken a little more loosely. It's perfectly possible to approach life in a pseudo-Bayesian sense, starting with your belief about what is the case, and updating it as you learn more, even if you're not actually constantly performing Bayesian math in your head like an imaginary person in an economic model.
But even if many of their arguments fall apart a bit when applied to real life and not to economics, this is still a helpful book for lay readers. Their emphasis on knowing when to say "I do not know" and the value of asking "What is going on here?" are well-taken. And their targets aren't just straw men — over-quantified economic models are real, and used as the basis for all sorts of hugely consequential decisions. (Among other things, they cite the bank models before the housing bubble burst in 2007-8, for which the collapse of the housing market allegedly involved "25-standard deviation moves several days in a row." As they note, "our universe has not existed long enough for there to have been days on which 25 standard deviation events could occur"; the problem was the models' assumptions, inputs and algorithms were wrong, and considered an event that actually did happen as basically impossible.) I think their points are made too strongly for laypeople's purposes, and are too focused on economics rather than daily life, but they're still well-taken.
When I say that, I don't mean it was the kind of book that made me reconsider everything I believed, though it did some of that. Rather, nearly every page had me pausing to consider and mentally debate the book's arguments, which were always interesting even when I ultimately came down against the authors. (Of course, I have a larger-than-normal interest in epistemology and uncertainty, so your mileage may vary...)
The main point of the authors if to respond to an intellectual movement that claims to be foregrounding uncertainty in their analysis. This movement, inspired by Bayesian statistics, tries to get away from narrative-driven beliefs by instead quantifying the likelihood of those beliefs, and constantly updating those likelihoods in response to new information. The authors argue this movement is misguided, and actually does the opposite of what its backers claim it does — by quantifying risk, they say, people are applying false precision to things that are actually radically uncertain — impossible to quantify.
"Reasonable uncertainty is uncertainty which can be removed by looking something up (I am uncertain which city is the capital of Pennsylvania) or which can be represented by a known probability distribution of outcomes (the spin of a roulette wheel). With radical uncertainty, however, there is no similar means of resolving the uncertainty — we simply do not know."
Put another way, it is the difference between "risk" and "uncertainty" where "risk" means "unknowns which could be describe with probabilities" and "uncertainty" which can't. Today, the authors argue, we tend to treat uncertain things as if they are actually risks that can be precisely quantified. They give the example of national security advisers meeting with President Barack Obama in 2011, giving their assessments of whether Osama bin Laden was actually in a compound in Abbottabad, Pakistan — one advisor said there was a 95 percent chance bin Laden was there, while another said 80 percent and another 40 percent. Obviously these percentages are completely different from, say, the 50 percent chance that a fair coin will come up heads.
But this example also brings up one of the problems with the book: it's rather over-focused on issues inside the field of economics (and adjacent areas), and the author's arguments against various forms of probabilistic reasoning run into more issues when they move past critiquing over-quantified economic models and move to day-to-day decision-making.
To return to the prior example, in a very literal sense, the statement that there was a 95 percent chance Bin Laden was in Abbottabad is meaningless. Either he was there or he wasn't; it wasn't like you could raid the compound 20 times and expect to find Bin Laden 19 times. But this wasn't a case where the only options for belief were "he's there," "he's not there," and "we don't know." One can believe it is "more likely than not" that something is true, that evidence suggests something but doesn't prove it. Saying "95 percent" may not have any solid statistical basis, but isn't it a perfectly fine synonym for "almost certain"? To be sure, we need to make sure not to take that 95 percent estimate too seriously, as a real, empirical probability. But at a certain point, applied to real life and not economic models, this argument becomes a straw man.
Another favorite straw man argument the authors use is to mock the idea that actual people making real decisions have a "Bayesian dial floating over their heads" — a reference to the model of Bayesian statistics, which starts with a "prior probability" that something is true and then updates that probability based on evidence. Real decisions, they say, are based on narratives, not statistical models. Again, this is an argument that is obviously true in a very narrow sense — as they demonstrate, even professional economists and statisticians usually don't use their probabilistic methods for making life decisions — but falls down a bit when taken a little more loosely. It's perfectly possible to approach life in a pseudo-Bayesian sense, starting with your belief about what is the case, and updating it as you learn more, even if you're not actually constantly performing Bayesian math in your head like an imaginary person in an economic model.
But even if many of their arguments fall apart a bit when applied to real life and not to economics, this is still a helpful book for lay readers. Their emphasis on knowing when to say "I do not know" and the value of asking "What is going on here?" are well-taken. And their targets aren't just straw men — over-quantified economic models are real, and used as the basis for all sorts of hugely consequential decisions. (Among other things, they cite the bank models before the housing bubble burst in 2007-8, for which the collapse of the housing market allegedly involved "25-standard deviation moves several days in a row." As they note, "our universe has not existed long enough for there to have been days on which 25 standard deviation events could occur"; the problem was the models' assumptions, inputs and algorithms were wrong, and considered an event that actually did happen as basically impossible.) I think their points are made too strongly for laypeople's purposes, and are too focused on economics rather than daily life, but they're still well-taken.
September 4, 2020
One of the best and clearest introductions to the distinction of small world puzzles vs large world mysteries and the complete nonsense of trying to apply models and the bayesian dial to solve everything under conditions of radical uncertainty.
December 16, 2020
The book is very important for economics, but many of the ideas are already common to cognitive scientists, so little new.
However I'd rate their talk at LSE about the same topic with five stars, for me it condensed the most important parts and was therefore highly valuable.
However I'd rate their talk at LSE about the same topic with five stars, for me it condensed the most important parts and was therefore highly valuable.
April 7, 2020
The book's title is a no brainer in the crazy pandemic world of today. This tour de force questions the assumption behind economic models that humans behave rationally. Instead, begin by asking: "What's going on here?" In many cases, the best you can do is come up with possible alternative explanations or options, and then to construct a reference narrative, a story that expresses realistic expectations that is robust and resilient. Then take this narrative and use trial and error to help navigate a world of radical uncertainty. The best narratives may not be completely realistic, but are credible, coherent, and sufficiently good approximations for the purpose in hand.
Many economists misuse statistics, and particularly probability. Prob works for small world structured games of chance with completely specified and unchanging rules like flipping a coin or spinning a roulette wheel. For most things in the larger world, there is only "subjective probability", which isn't really probability at all. For example: who's going to win the Kentucky Derby? Briskit has a 90 percent chance. This isn't probability, its an opinion that Briskit is a strong contender. It's not objective based on frequency and math, its subjective. Using math models using subjective probabilities and assumed values may provide valuable insights into real worlds, but does not describe them. Economists accept that markets are incomplete, yet believe that people behave as if they were maximizing their subjective probabilities of utility. This is "utterly ridiculous". There is no stable structure of the world such that we could learn from past experience and from this predict future behavior. Engineers produce models that get spacecraft safely to the moon and back. Economic models are useful for framing arguments, but can't accurately predict stock prices or GDP growth.
Even if objective probability is possible to specify, it only works with a credible model that works in the real world, not just in the small world of stylized description. For example, political opinion polls can go off the rails for many reasons. Was the random sample taken from a representative population? Did people answer the questions honestly? And so forth.
Small world models (comparative advantage; asymmetric information; invisible hand; rational, efficient markets) are fictional narratives that deliver broad insights, but not accurate details like who is going to win a close election.
Another misuse is the greater focus of many models on normal statistical distribution, rather than power law distribution. The former leads to a bell curve, the latter to a logarithmic curve hugging the x and y axis, where the nth most frequent item (eg degree of change in stock price, severity of earthquakes) appears with a frequency 1/n times the most frequent item. Power law takes better account of black swan events, while bell curves can make you forget them.
Models apply a common template to disparate situations. They fill in gaps in knowledge by inventing numbers, often in vast quantities. They assume, usually without justification, that underlying processes are stationary. They don't account for uncertainty. They prevent meaningful public consultation because their complexity makes them impossible to understand by laymen. They consider how you would make a decision if you had perfect data about the past and future; very few of the relevant data are known. The solution: make them all up. They don't lead to evidence based policy, but policy based evidence. The purpose of models is provide a superficially objective basis to justify what policy makers have already decided.
On the other hand, useful models identify the key factors in the simplest possible way. They put a value on each factor, and are flexible in considering options. Such models can help decision makers, but don't substitute for the human judgement needed at the end to make a decision. The test of a good model is whether it helps make better decisions. A good economist is like a good dentist or plumber: s(he) fixes things. A good model is like good literary fiction or a doctor's diagnosis: it provides a credible narrative to help understand the context in which a decision has to be made. It frames the problem, and provides credible data.
Moving forward takes a good strategy and good leader. A good strategy isn't a wishlist; it makes choices that are robust and resilient to uncertainty. It reviews the options, framing and data that may come from models and selects what to do next, how to do it, and how to know when its done.
Being a leader means that you have the responsibility to lead, not that you are necessarily more knowledgeable than anyone else. Effective leaders have opinions, but allow them to be challenged and debated by others. Getting the full range of expert views is key to achieving quality results. In the decision to get Bin Laden, Obama asked his experts and they gave a range of subjective probabilities that the mission would succeed: from 20-95 percent. He listened to all and their reasons, figured out "what's going on here", and made the call.
Many economists misuse statistics, and particularly probability. Prob works for small world structured games of chance with completely specified and unchanging rules like flipping a coin or spinning a roulette wheel. For most things in the larger world, there is only "subjective probability", which isn't really probability at all. For example: who's going to win the Kentucky Derby? Briskit has a 90 percent chance. This isn't probability, its an opinion that Briskit is a strong contender. It's not objective based on frequency and math, its subjective. Using math models using subjective probabilities and assumed values may provide valuable insights into real worlds, but does not describe them. Economists accept that markets are incomplete, yet believe that people behave as if they were maximizing their subjective probabilities of utility. This is "utterly ridiculous". There is no stable structure of the world such that we could learn from past experience and from this predict future behavior. Engineers produce models that get spacecraft safely to the moon and back. Economic models are useful for framing arguments, but can't accurately predict stock prices or GDP growth.
Even if objective probability is possible to specify, it only works with a credible model that works in the real world, not just in the small world of stylized description. For example, political opinion polls can go off the rails for many reasons. Was the random sample taken from a representative population? Did people answer the questions honestly? And so forth.
Small world models (comparative advantage; asymmetric information; invisible hand; rational, efficient markets) are fictional narratives that deliver broad insights, but not accurate details like who is going to win a close election.
Another misuse is the greater focus of many models on normal statistical distribution, rather than power law distribution. The former leads to a bell curve, the latter to a logarithmic curve hugging the x and y axis, where the nth most frequent item (eg degree of change in stock price, severity of earthquakes) appears with a frequency 1/n times the most frequent item. Power law takes better account of black swan events, while bell curves can make you forget them.
Models apply a common template to disparate situations. They fill in gaps in knowledge by inventing numbers, often in vast quantities. They assume, usually without justification, that underlying processes are stationary. They don't account for uncertainty. They prevent meaningful public consultation because their complexity makes them impossible to understand by laymen. They consider how you would make a decision if you had perfect data about the past and future; very few of the relevant data are known. The solution: make them all up. They don't lead to evidence based policy, but policy based evidence. The purpose of models is provide a superficially objective basis to justify what policy makers have already decided.
On the other hand, useful models identify the key factors in the simplest possible way. They put a value on each factor, and are flexible in considering options. Such models can help decision makers, but don't substitute for the human judgement needed at the end to make a decision. The test of a good model is whether it helps make better decisions. A good economist is like a good dentist or plumber: s(he) fixes things. A good model is like good literary fiction or a doctor's diagnosis: it provides a credible narrative to help understand the context in which a decision has to be made. It frames the problem, and provides credible data.
Moving forward takes a good strategy and good leader. A good strategy isn't a wishlist; it makes choices that are robust and resilient to uncertainty. It reviews the options, framing and data that may come from models and selects what to do next, how to do it, and how to know when its done.
Being a leader means that you have the responsibility to lead, not that you are necessarily more knowledgeable than anyone else. Effective leaders have opinions, but allow them to be challenged and debated by others. Getting the full range of expert views is key to achieving quality results. In the decision to get Bin Laden, Obama asked his experts and they gave a range of subjective probabilities that the mission would succeed: from 20-95 percent. He listened to all and their reasons, figured out "what's going on here", and made the call.
August 28, 2022
"Radical Uncertainty" to książka, która ni z tego, ni z owego, wpływa całkowicie i całościowo na sposób postrzegania procesów decyzyjnych widywanych w przedsiębiorstwach, finansach i bankowości, czy nawet w podejściach polecanych w różnych książkach. Dlaczego?
Otóż John Kay i Mervyn King zebrali potężny stos przykładów i dowodów, pokazujących że rozumowanie statystyczne aplikowane do rzeczywistego świata zawsze daje wyniki odstające od rzeczywistości lub w ogóle do niej nieadekwatne. Wszystkie modele bazują na pewnych założeniach i uproszczeniach, a te nie zawsze są prawidłowe - a wręcz można powiedzieć, że wybitnie często okazują się błędne. Dlatego przypisywanie "prawdopodobieństwa" do różnych scenariuszy możliwych wydarzeń w różnego rodzaju prognozach jest mydleniem oczu.
Cała idea opiera się na tym, że o ile statystyka i rachunek prawdopodobieństwa bardzo dobrze sprawdzają się w naukach ścisłych czy tzw. zamkniętych światach - jak poker - o tyle w świecie nauk społecznych czy ekonomicznych są często narzędziem w rzeczywistości nieoferującym tego, co zdaje się oferować. Co gorsza, teorie budowane na błędnych wnioskach z tego narzędzia zbyt często trafiają do osób, które nie rozumieją ich błędów.
Objawia się to w ten sposób, że o ile inżynier w oparciu o teorię grawitacji jest w stanie opracować sterowanie sondą kosmiczną w taki sposób, by w przeciągu kilkunastu lat utrzymywała prawidłową trajektorię, o tyle ekonomista nie jest w stanie przewidzieć trendów ekonomicznych i z jakimkolwiek stopniem pewności powiedzieć, że dnia tego i tego wybuchnie kryzys ekonomiczny.
Ale to nie tak, że ekonomiści są głupi. Chodzi raczej o to, że gdy korzystają z modeli szacujących prawdopodobieństwo różnych scenariuszy, muszą brać poprawkę na to, że prawdopodobieństwo na to, że model jest prawidłowy, wcale nie jest jedynkowe. Na domiar złego, każda prognoza wpływa na rynki, więc przewidzenie daty kryzysu może być błędne dlatego, że poczynione przygotowania do tej daty zneutralizują kryzys. Co robić, jak żyć?
Nie chcę się tu zanadto rozpisywać, ale idąc dalej, autorzy rozwijają mnóstwo poszczególnych wątków. Jak podejmować decyzje bez tych skomplikowanych obliczeń statystycznych? Dlaczego to narracje, a nie liczby, działają na nas skuteczniej? Jak tworzyć narracje? Czym są dobre i złe narracje? Czy narracje w ogóle muszą odzwierciedlać rzeczywistość?
Cóż, "Radical Uncertainty" to trudna i dająca do myślenia lektura. Ale po czasie zaczynam rozumieć, że autorzy zawarli tu sporo prawdy i mądrości. Można to określić odniesieniem ekonomii do praktyki. Dlatego gorąco ten tytuł polecam.
Warto przeczytać również:
1. Hans Rosling - "Factfulness"
W tej książce Rosling rozprawia się z błędnymi przekonaniami, bazując na danych. Obala w ten sposób narracje, które narosły w popkulturze.
2. Marcin Popkiewicz - "Świat na rozdrożu"
Jeżeli jest jakaś książka, z którą kłóci się "Radical Uncertainty", to jest to zdecydowanie książka Popkiewicza, prognozująca możliwe scenariusze dotyczące zmiany klimatu i wpływu tego zdarzenia na gospodarkę. Popkiewicz niejednokrotnie podkreśla, że ekonomiści zbyt wiele rzeczy tłumaczą swoimi modelami, ale sam bazuje na innych modelach. Po lekturze obu tytułów, głęboko zastanawiam się nad tym, jakie jest prawdopodobieństwo że scenariusz zahamowania wzrostu gospodarczego przedstawiany przez Popkiewicza jest prawidłowy.
3. Erwin Schrödinger - "Czym jest życie?"
Może nie stricte związana z tematem, ale ta książka znanego fizyka kwantowego w interesujący sposób przedstawia fizykę jako naukę o podłożu statystycznym i często o niej myślałem, czytając "Radical Uncertainty".
Otóż John Kay i Mervyn King zebrali potężny stos przykładów i dowodów, pokazujących że rozumowanie statystyczne aplikowane do rzeczywistego świata zawsze daje wyniki odstające od rzeczywistości lub w ogóle do niej nieadekwatne. Wszystkie modele bazują na pewnych założeniach i uproszczeniach, a te nie zawsze są prawidłowe - a wręcz można powiedzieć, że wybitnie często okazują się błędne. Dlatego przypisywanie "prawdopodobieństwa" do różnych scenariuszy możliwych wydarzeń w różnego rodzaju prognozach jest mydleniem oczu.
Cała idea opiera się na tym, że o ile statystyka i rachunek prawdopodobieństwa bardzo dobrze sprawdzają się w naukach ścisłych czy tzw. zamkniętych światach - jak poker - o tyle w świecie nauk społecznych czy ekonomicznych są często narzędziem w rzeczywistości nieoferującym tego, co zdaje się oferować. Co gorsza, teorie budowane na błędnych wnioskach z tego narzędzia zbyt często trafiają do osób, które nie rozumieją ich błędów.
Objawia się to w ten sposób, że o ile inżynier w oparciu o teorię grawitacji jest w stanie opracować sterowanie sondą kosmiczną w taki sposób, by w przeciągu kilkunastu lat utrzymywała prawidłową trajektorię, o tyle ekonomista nie jest w stanie przewidzieć trendów ekonomicznych i z jakimkolwiek stopniem pewności powiedzieć, że dnia tego i tego wybuchnie kryzys ekonomiczny.
Ale to nie tak, że ekonomiści są głupi. Chodzi raczej o to, że gdy korzystają z modeli szacujących prawdopodobieństwo różnych scenariuszy, muszą brać poprawkę na to, że prawdopodobieństwo na to, że model jest prawidłowy, wcale nie jest jedynkowe. Na domiar złego, każda prognoza wpływa na rynki, więc przewidzenie daty kryzysu może być błędne dlatego, że poczynione przygotowania do tej daty zneutralizują kryzys. Co robić, jak żyć?
Nie chcę się tu zanadto rozpisywać, ale idąc dalej, autorzy rozwijają mnóstwo poszczególnych wątków. Jak podejmować decyzje bez tych skomplikowanych obliczeń statystycznych? Dlaczego to narracje, a nie liczby, działają na nas skuteczniej? Jak tworzyć narracje? Czym są dobre i złe narracje? Czy narracje w ogóle muszą odzwierciedlać rzeczywistość?
Cóż, "Radical Uncertainty" to trudna i dająca do myślenia lektura. Ale po czasie zaczynam rozumieć, że autorzy zawarli tu sporo prawdy i mądrości. Można to określić odniesieniem ekonomii do praktyki. Dlatego gorąco ten tytuł polecam.
Warto przeczytać również:
1. Hans Rosling - "Factfulness"
W tej książce Rosling rozprawia się z błędnymi przekonaniami, bazując na danych. Obala w ten sposób narracje, które narosły w popkulturze.
2. Marcin Popkiewicz - "Świat na rozdrożu"
Jeżeli jest jakaś książka, z którą kłóci się "Radical Uncertainty", to jest to zdecydowanie książka Popkiewicza, prognozująca możliwe scenariusze dotyczące zmiany klimatu i wpływu tego zdarzenia na gospodarkę. Popkiewicz niejednokrotnie podkreśla, że ekonomiści zbyt wiele rzeczy tłumaczą swoimi modelami, ale sam bazuje na innych modelach. Po lekturze obu tytułów, głęboko zastanawiam się nad tym, jakie jest prawdopodobieństwo że scenariusz zahamowania wzrostu gospodarczego przedstawiany przez Popkiewicza jest prawidłowy.
3. Erwin Schrödinger - "Czym jest życie?"
Może nie stricte związana z tematem, ale ta książka znanego fizyka kwantowego w interesujący sposób przedstawia fizykę jako naukę o podłożu statystycznym i często o niej myślałem, czytając "Radical Uncertainty".
Displaying 1 - 30 of 100 reviews