‘The curious task of economics is to demonstrate to men how little they really know about what they imagine they can design.’
‘Cross the river by feeling for stones.’
– attributed to Deng Xiaoping
The electric toaster seems a humble thing. It was invented in 1893, roughly halfway between the appearance of the light bulb and that of the aeroplane. This century-old technology is now a household staple. Reliable, efficient toasters are available for less than an hour’s wage.
Nevertheless, Thomas Thwaites, a postgraduate design student at the Royal College of Art in London, discovered just what an astonishing achievement the toaster is when he embarked on what he called the ‘Toaster Project’. Quite simply, Thwaites wanted to build a toaster from scratch. He started by taking apart a cheap toaster, to discover that it had over four hundred components and sub-components. Even the most primitive model called for:
Copper, to make the pins of the electric plug, the cord, and internal wires. Iron to make the steel grilling apparatus, and the spring to pop up the toast. Nickel to make the heating element. Mica (a mineral a bit like slate) around which the heating element is wound, and of course plastic for the plug and cord insulation, and for the all important sleek looking casing.
The scale of the task soon became clear. To get iron ore, Thwaites had to travel to an old mine in Wales that now serves as a museum. He tried to smelt the iron using fifteenth-century technology, and failed dismally. He fared no better when he replaced bellows with hairdryers and a leaf-blower. His next attempt was even more of a cheat: he used a recently patented smelting method and two microwave ovens, one of which perished in the attempt, to produce a coin-sized lump of iron.
Plastic was no easier. Thwaites tried but failed to persuade BP to fly him out to an offshore rig to collect some crude oil. His attempts to make plastic from potato starch were foiled by mould and hungry snails. Finally, he settled for scavenging some plastic from a local dump, melting it down and moulding it into a toaster’s casing. Other short cuts followed. Thwaites used electrolysis to obtain copper from the polluted water of an old mine in Anglesey, and simply melted down some commemorative coins to produce nickel, which he drew into wire using a specialised machine from the RCA’s jewellery department.
Such compromises were inevitable. ‘I realised that if you started absolutely from scratch, you could easily spend your life making a toaster,’ he admitted. Despite his Herculean efforts to duplicate the technology, Thomas Thwaites’s toaster looks more like a toaster-shaped birthday cake than a real toaster, its coating dripping and oozing like an icing job gone wrong. ‘It warms bread when I plug it into a battery,’ he told me, brightly. ‘But I’m not sure what will happen if I plug it into the mains.’ Eventually, he summoned up the courage to do so. Two seconds later, the toaster was toast.
The modern world is mind-bogglingly complicated. Far simpler objects than a toaster involve global supply chains and the coordinated efforts of many individuals, scattered across the world. Many do not even know the final destination of their efforts. As a lumberjack fells a giant of the Canadian forest, he doesn’t know whether the tree he topples will make bed frames or pencils. At the vast Chuquicamata mine in Chile, a yellow truck the size of a house growls up an incline blasted into the landscape; the driver does not trouble himself to ask whether the copper ore he carries is destined for the wiring of a toaster or the casing of a bullet.
The range of products, too, is astounding. There are a hundred thousand or so distinct items in an ordinary Wal-Mart. Eric Beinhocker, a complexity researcher at the McKinsey Global Institute, reckons that if you were to add up all the different sizes and shapes of shoes, shirts and socks, the different brands and flavours and sizes of jams and sauces, the millions of different books, DVDs and music downloads on offer, you would find that a major economy such as New York or London offers over ten billion distinct types of product. Many of these products were undreamt of when the toaster was first invented, and millions of new ones appear every month. The complexity of the society we have created for ourselves envelops us so completely that, instead of being dizzied, we take it for granted.
I used to view this sophistication as cause to celebrate. Now I am less sure. Certainly, this complex economy produces vast material wealth. Not everyone gets a share, but far more people today enjoy a high material standard of living than at any time in history; and, notwithstanding the occasional recession, the wealth continues to grow more quickly than it ever used to. The process that produces this wealth is near miraculous, and the job is far harder than we tend to acknowledge. Alternative systems, from feudalism to central planning, have attempted the same task and been consigned to the history books.
Yet the Toaster Project should give us pause for thought. Because it is a symbol of the sophistication of our world, the toaster is also a symbol of the obstacles that lie in wait for those who want to change it. From climate change to terrorism, fixing the banks to ending global poverty, there is no shortage of big policy problems out there. They are always up for debate, yet we never seem to move any closer to a solution. Humbler problems in business and everyday life also tend to conceal the same unexpected complexity as the Toaster Project.
This is partly a book about those problems. But more fundamentally, it’s a book that aims to understand how any problem – big or small – really gets solved in a world where even a toaster is beyond one man’s comprehension.
The toasting problem isn’t difficult: don’t burn the toast; don’t electrocute the user; don’t start a fire. The bread itself is hardly an active protagonist. It doesn’t deliberately try to outwit you, as a team of investment bankers might; it doesn’t try to murder you, terrorise your country, and discredit everything you stand for, as a terrorist cell or a group of insurgents in Iraq would. The toaster is merely an improved way to solve an old problem – the Romans loved toast – unlike the World Wide Web or the personal computer, which provide solutions to problems we never realised we had. The toasting problem is laughably simple compared to the problem of transforming a poor country such as Bangladesh into the kind of economy where toasters are manufactured with ease and every household can afford one, along with the bread to put into it. It is dwarfed by the problem of climate change – the response to which will require much more than modifying a billion toasters.
Such problems are the stuff of this book: how to fight insurgents who, of course, fight back; how to nurture ideas that matter when so many of those ideas are hard even to imagine; how to restructure an economy to respond to climate change, or to make poor countries rich; how to prevent rogue investment bankers from destroying the banking system again. These are complex, fast-moving problems in a complex, fast-moving world. I will argue that they have far more in common with each other than we realise. Curiously, they also have something in common with the more humble problems we face in our own lives.
Whenever such problems are solved, it is little short of a miracle. This book is about how such miracles happen, why they matter so much, and whether we can make them happen more often.
We’re proud of the change we’ve brought to Washington in these first hundred days, but we’ve got a lot of work left to do, as all of you know. So I’d like to talk a little bit about what my administration plans to achieve in the next hundred days. During the second hundred days, we will design, build and open a library dedicated to my first hundred days … I believe that my next hundred days will be so successful I will be able to complete them in 72 days. And on the 73rd day, I will rest.
This was President Obama addressing the White House Correspondents’ Dinner, traditionally a venue for a joke or two, a few months after a tidal wave of hope and high expectations had swept him into power in November 2008. It seems a long time ago now, but Obama’s joke cut close to the bone even then: people were expecting too much of one man.
We badly need to believe in the potency of leaders. Our instinctive response, when faced with a complicated challenge, is to look for a leader who will solve it. It wasn’t just Obama: every president is elected after promising to change the way politics works; and almost every president then slumps in the polls as reality starts to bite. This isn’t because we keep electing the wrong leaders. It is because we have an inflated sense of what leadership can achieve in the modern world.
Perhaps we have this instinct because we evolved to operate in small hunter–gatherer groups, solving small hunter–gatherer problems. The societies in which our modern brains developed weren’t modern: they contained a few hundred separate products, rather than ten billion. The challenges such societies faced, however formidable, were simple enough to have been solved by an intelligent, wise, brave leader. They would have been vastly simpler than the challenges facing a newly elected US president.
Whatever the reason, the temptation to look to a leader to fix our problems runs deep. Of course, a leader doesn’t have to solve every problem by himself. Good leaders surround themselves with expert advisers, seeking out the smartest specialists with the deepest insights into the problems of the day. But even deep expertise is not enough to solve today’s complex problems.
Perhaps the best illustration of this comes from an extraordinary two-decade investigation into the limits of expertise, begun in 1984 by a young psychologist called Philip Tetlock. He was the most junior member of a committee of the National Academy of Sciences charged with working out what the Soviet response might be to the Reagan administration’s hawkish stance in the Cold War. Would Reagan call the bluff of a bully or was he about to provoke a deadly reaction? Tetlock canvassed every expert he could find. He was struck by the fact that, again and again, the most influential thinkers on the Cold War flatly contradicted one another. We are so used to talking heads disagreeing that perhaps this doesn’t seem surprising. But when we realise that the leading experts cannot agree on the most basic level about the key problem of the age, we begin to understand that this kind of expertise is far less useful than we might hope.
Tetlock didn’t leave it at that. He worried away at this question of expert judgement for twenty years. He rounded up nearly three hundred experts – by which he meant people whose job it was to comment or advise on political and economic trends. They were a formidable bunch: political scientists, economists, lawyers and diplomats. There were spooks and think-tankers, journalists and academics. Over half of them had PhDs; almost all had postgraduate degrees. And Tetlock’s method for evaluating the quality of their expert judgement was to pin the experts down: he asked them to make specific, quantifiable forecasts – answering 27,450 of his questions between them – and then waited to see whether their forecasts came true. They rarely did. The experts failed, and their failure to forecast the future is a symptom of their failure to understand fully the complexities of the present.
It wasn’t that expertise was entirely useless. Tetlock compared his experts’ responses to those of a control group of undergraduates, and the experts did better. But by any objective standard, they didn’t do well. And the return on expertise was distinctly limited. Once experts have acquired a broad knowledge of the political world, deeper expertise in a specific field doesn’t seem to help much. Predictions about Russia from experts on Russia were no more accurate than predictions about Russia from experts on Canada.
Most accounts of Tetlock’s research savour the humbling of the professional pundits. And why not? One of Tetlock’s more delicious discoveries was that the more famous experts – those who spent a lot of time as talking heads on television – were especially incompetent. Louis Menand, writing in the New Yorker, enjoyed the notion of bumbling seers, and concluded, ‘the best lesson of Tetlock’s book may be the one that he seems most reluctant to draw: Think for yourself’.
Yet there is a reason why Tetlock himself hesitates to draw that conclusion: his results clearly show that experts do outperform non-experts. These intelligent, educated and experienced professionals have insights to contribute – it’s just that those insights go only so far. The problem is not the experts; it is the world they inhabit – the world we all inhabit – which is simply too complicated for anyone to analyse with much success.
So, if expertise is of such limited help in the face of our complex, ever-changing human society, what can we do to solve the problems we face? Perhaps we should look for clues in the success story we’ve already encountered: the amazing material wealth of modern developed countries.
In 1982, just a couple of years before Philip Tetlock began his painstaking examination of expertise, two management consultants, Tom Peters and Robert Waterman, concluded their own detailed study of excellence in business. In Search of Excellence was published to great acclaim and launched Peters’s career as one of the world’s most recognisable management gurus. The two authors, working with their colleagues at McKinsey, used a mixture of data and subjective judgement to settle on a list of forty-three ‘excellent’ companies, which they then studied intensively in a bid to unlock their secrets.
Just two years later, Business Week ran a cover story entitled ‘Oops! Who’s Excellent Now?’ Out of the forty-three companies, fourteen, almost a third, were in serious financial trouble. Excellence – if that was what Peters and Waterman really found when they studied the likes of Atari and Wang Laboratories – appears to be a fleeting quality.
It seems strange that so many apparently excellent companies could find themselves in deep trouble so quickly. Perhaps there was something uniquely silly about Peters and Waterman’s project. Or perhaps there was something uniquely turbulent about the early 1980s – In Search of Excellence was published during a severe recession, after all.
But perhaps not. The ‘who’s excellent now?’ experience is reinforced by a careful study from the economic historian Leslie Hannah, who in the late 1990s decided to trace the fortunes of every one of the largest companies in the world in 1912. These were corporate giants that had survived a merger shakedown over the preceding few years and typically employed at least ten thousand workers.
At the top of the list was US Steel, a gigantic corporation even by today’s standards, employing 221,000 workers. This was a company with everything going for it: it was the market leader in the largest and most dynamic economy in the world; and it was in an industry that has been of tremendous importance ever since. Yet US Steel had disappeared from the world’s top hundred companies by 1995; at the time of writing, it was not even in the top five hundred.
Next on the list was Jersey Standard, which these days continues to prosper under the name Exxon. General Electric and Shell were also in the top ten both in 1912 and in 1995. But none of the other top-ten titans was in the top ten by 1995. More remarkably, none of them was even in the top hundred. Names such as Pullman and Singer recall a bygone age. Others, such as J&P Coats, Anaconda and International Harvester, are barely recognisable. It is hard to imagine just how large and powerful these companies once were – the closest parallels would be the likes of Microsoft and Wal-Mart today – and how permanent their success must have seemed. And while it could be said that Pullman and Singer suffered from being market leaders in declining industries, their fate was not inevitable. Singer made sewing machines, but Toyota’s origins as a manufacturer of looms were no more promising. Other former titans, such as Westinghouse Electric, Cudahy Packing and American Brands, were in the same dynamic industries as the rare success stories General Electric and Procter & Gamble. Yet they failed.
Just as Philip Tetlock’s experts have proved less capable than we tend to think, in the face of a complex world, these great companies are more transient than we realise. Ten of Hannah’s top hundred had vanished within a decade; over half disappeared over the next 83 years. The lesson seems to be that failure is fundamental to the way the market creates sophisticated and wealthy economies. But perhaps what Peters, Waterman and Hannah found merely reflects the fact that if you start at the top, the only way is down. What happens when we look at survival rates in young, dynamic industries?
The answer is that failure rates are even higher. Consider the early printing industry. The printing press was invented by Johannes Gutenberg, a man who changed the world utterly, and produced the celebrated Gutenberg Bible in 1455. But the Gutenberg Bible was a ruinous project that put him out of business. The centre of the printing industry quickly moved to Venice, where twelve companies were established by 1469. Nine of them were gone in just three years, as the industry fumbled for a profitable business model. (It eventually found one: printing pre-packaged relief from divine punishment in the form of religious indulgences.)
At the dawn of the automobile industry, two thousand firms were operating in the United States. Around 1 per cent of them survived. The dot-com bubble spawned and killed countless new businesses. Today, 10 per cent of American companies disappear every year. What is striking about the market system is not how few failures there are, but how ubiquitous failure is even in the most vibrant growth industries.
Why, then, are there so many failures in a system that seems to be so economically successful overall? It is partly the difficulty of the task. Philip Tetlock showed how hard it was for expert political and economic analysts to generate decent forecasts, and there is no reason to believe that it is any easier for marketers or product developers or strategists to predict the future. In 1912, Singer’s managers probably did not forecast the rise of the off-the-peg clothing industry. To make things even more difficult, corporations must compete with each other. To survive and be profitable it is not enough to be good; you must be one of the best. Asking why so many companies go out of business is the same as asking why so few athletes reach Olympic finals. In a market economy, there is usually room for only a few winners in each sector. Not everyone can be one of them.
The difference between market-based economies and centrally planned disasters, such as Mao Zedong’s Great Leap Forward, is not that markets avoid failure. It’s that large-scale failures do not seem to have the same dire consequences for the market as they do for planned economies. (The most obvious exception to this claim is also the most interesting: the financial crisis that began in 2007. We’ll find out why it was such a catastrophic anomaly in chapter six.) Failure in market economies, while endemic, seems to go hand in hand with rapid progress.
The modern computer industry is a striking example: the most dynamic sector of the economy has also been the one in which failure is everywhere you look. The industry started with failure: when transistors replaced vacuum tubes as the basic elements of the computer, vacuum-tube manufacturers failed to make the switch. The likes of Hughes, Transitron and Philco took over, before stumbling in turn as integrated circuits replaced transistors, and the baton passed to Intel and Hitachi.
Meanwhile, Xerox, struggling to survive the expiry of its patents on photocopying, established the Palo Alto Research Center (or Parc), which developed the fax machine, the graphical interface that defines all modern computers, the laser printer, the Ethernet, and the first personal computer, the Alto. Yet Xerox did not become a powerhouse in personal computing. Many of the Alto’s successors – including the ZX Spectrum, the BBC Micro and Japan’s MSX standard – were dead-ends in the history of computing. It fell to IBM to produce the direct ancestor of today’s personal computer – only to then unwittingly hand over control of the most valuable part of the package, the operating system, to Microsoft. IBM eventually bowed out of the personal computer business in 2005, selling its interests to a Chinese company. Apple also lost out to Microsoft in the 1980s, despite perfecting the user-friendly computer (although it was later to bounce back selling music, iPods and phones). Microsoft itself was caught unawares by the internet, lost the search-engine war with Google, and may soon lose its dominant position in software altogether. Who knows? Only the most arrogant forecaster would be able to convince himself that he could predict the next twist or turn in this market. The most successful industry of the last forty years has been built on failure after failure after failure.
The humble toaster which so baffled Thomas Thwaites is itself a product of trial and error. The Eclipse of 1893 was not a success: its iron heating element was prone to rust and tended to melt and start fires. The company that marketed it no longer exists. The first successful toaster did not emerge until 1910. It boasted a superior nickel-chrome alloy for the heating element but was still flawed. Most notably, that heating element was exposed, making it a potential source of household fires, burns and electrocutions. It took several decades for the practical and familiar pop-up toaster design to emerge, by which time many manufacturers had quit the business or gone bankrupt.
The market has solved the problem of generating material wealth, but its secret has little to do with the profit motive or the superior savvy of the boardroom over the cabinet office. Few company bosses would care to admit it, but the market fumbles its way to success, as successful ideas take off and less successful ones die out. When we see the survivors of this process – such as Exxon, General Electric and Procter & Gamble – we shouldn’t merely see success. We should also see the long, tangled history of failure, of all of the companies and all of the ideas that didn’t make it.
Biologists have a word for the way in which solutions emerge from failure: evolution. Often summarised as survival of the fittest, evolution is a process driven by the failure of the less fit. Disconcertingly, given our instinctive belief that complex problems require expertly designed solutions, it is also completely unplanned. Astounding complexity emerges in response to a simple process: try out a few variants on what you already have, weed out the failures, copy the successes – and repeat for ever. Variation, and selection, again and again.
We are used to thinking about evolution as something that happens in the natural world – a biological phenomenon. But it doesn’t have to be. Anyone can watch evolution taking place in a digital world, thanks to a graphics expert named Karl Sims. If you’ve ever seen Titanic, or the Lord of the Rings trilogy, or the Spider-Man films, then you’ve enjoyed the work of Karl Sims, who founded the special-effects company GenArts. But in the early 1990s, before Sims turned his attention to the visual-effects business, he produced moving images that are far cruder and yet, in some ways, more remarkable.
Sims wanted to watch evolution in progress. More than that, he wanted to create a virtual environment in which he could set its direction. Sims programmed simulations of settings such as a tank of water, and into them he dropped crude virtual creatures consisting of simple control systems, some sensors and random assortments of articulated blocks. Most of these jumbled creatures sank to the bottom and thrashed about without any great success. A few, however, were able to swim a little. Sims then applied the evolutionary process, instructing his computer to discard the floundering creatures and to create mutations based on the more successful swimmers: variation and selection. Most of the mutations were failures, of course. But the failures were continuously discarded, the occasional successes were allowed to flourish. From the most mindless and random of processes, remarkable results emerged: virtual creatures that resembled tadpoles, eels and rays, along with a number of apparently successful entities that looked like nothing on earth.
On another evolutionary run, Sims rewarded creatures for successfully taking possession of a green cube, in competition with each other. The trial-and-error process of evolution produced a wide range of workable solutions, some obvious, others less so, from ignoring the cube and lunging at the opponent, to making a quick grab for the cube and then dashing off, to simply toppling forward and covering the cube with a heavy slab of a body. Sims was not the designer, nor even the subjective judge of success after the fact: he simply set up an evolutionary environment and recorded what happened. The process he created was entirely blind and stupid: there was no foresight, planning or conscious design in any of the mutations. Yet the blind evolutionary process produced marvellous things.
Why is trial and error such an effective tool for solving problems? The evolutionary algorithm – of variation and selection, repeated – searches for solutions in a world where the problems keep changing, trying all sorts of variants and doing more of what works. One way to think about this quest for solutions is to imagine a vast, flat landscape, divided into a grid of billions of squares. On each square is a document: a recipe describing a particular strategy. Evolutionary theorists call this a ‘fitness landscape’. If the fitness landscape is biological, each strategy is a different genetic recipe: some squares describe fish; some describe birds; some describe human beings; while the majority describe a genetic mush that represents nothing that could ever survive in reality. But the fitness landscape might equally represent recipes for dinner: some produce curries; others produce salads; many produce dishes that are nauseating or even poisonous. Or the fitness landscape might contain business strategies: different ways to run an airline or a fast-food chain.
For any problem, it’s possible to imagine a huge range of potential solutions, each one carefully written down and scattered on this vast landscape. Imagine, too, that each recipe is very similar to its neighbours: two adjacent dinner recipes might be identical save for one demanding a little more salt and the other a slightly longer cooking time. Two neighbouring business strategies might advocate doing everything the same, except that one prescribes slightly higher prices and a bit more marketing.
We’ve been imagining a flat plane stretching in every direction, but now let’s change the picture and say that on our fitness landscape: the better the solution, the higher the altitude of the square that contains it. Now the fitness landscape is a jumble of cliffs and chasms, plateaus and jagged summits. Valleys represent bad solutions; mountain tops are good. In an ecosystem, the latter are creatures more likely to survive and reproduce; in the market place, they are the profitable business ideas; and at the dinner party, they are the tastiest dishes. In our dinner-party landscape, a deep, dark pit might contain a recipe for spaghetti with fish fingers and a jar of curry sauce. From there, the only way is up. Trek in one direction and you might eventually ascend to the soaring peak of Bolognese ragù. Head off in the opposite direction and you might eventually climb to the summit of a Bangladeshi fish curry.
Problem-solving on a contoured fitness landscape means trying to find the high peaks. In dinner-party space, that’s not so hard. But in a biological ecosystem, or an economy, the peaks keep moving – sometimes slowly, sometimes quickly. Pullman and Singer went out of business because the peaks they were standing on suddenly disappeared. The peak that McDonald’s currently occupies has been around for a while, moving slowly as new technologies become available and new tastes develop. The Google peak is very young, and it exists only because of earlier developments, such as the computer and the World Wide Web, just as squirrels exist only because there are trees for them to inhabit. And the Google peak is moving fast, more like a rolling wave than a mountain. At the moment, Google is surfing along, adapting its strategy to stay on or near the crest of the wave. Like surfing itself, this is harder than it looks.
As one peak subsides, others may not be clearly visible. The biological process of evolution through natural selection is entirely blind; finding a corporate strategy may or may not be a more deliberate and far-sighted process, as we shall shortly see. But Tetlock’s research on expertise suggests that, even if other peaks in corporate strategy are sometimes visible, executives see them only fleetingly, through heavy cloud.
We can imagine many ways to search for peaks in this changeable and mysterious landscape. Biological evolution usually moves in small steps, but occasionally takes wild leaps – a single mutation might give a creature an extra pair of legs or totally different skin pigmentation. This combination, along with the culling of failed experiments, works well. Some strategies will cling to a familiar summit as it shifts around; others, by darting off, may find a new peak rising up. The process of evolution strikes a balance between discovering the new and exploiting the familiar very well. In fact, Stuart Kauffmann and John Holland, both complexity theorists affiliated with the multidisciplinary Santa Fe Institute, have shown that the evolutionary approach is not just another way of solving complex problems. Given the likely shape of these ever-shifting landscapes, the evolutionary mix of small steps and occasional wild gambles is the best possible way to search for solutions.
Evolution is effective because, rather than engaging in an exhaustive, time-consuming search for the highest peak – a peak that may not even be there tomorrow – it produces ongoing, ‘works for now’ solutions to a complex and ever-changing set of problems. In biological evolution, solutions include photosynthesis, pairs of eyes and mothers’ milk. In economic evolution, solutions include double-entry book-keeping, supply-chain management and ‘buy one, get one free’. Some of what works seems to be perennial. The rest, such as being a Tyrannosaurus rex or the world’s most efficient manufacturer of VHS video cassettes, is rooted in a particular place and time.
We know that the evolutionary process is driven by variation and selection. In biology, variation emerges from mutations and from sexual reproduction, which mixes the genes from two parents. Selection happens through heredity: successful creatures reproduce before they die and have offspring that share some or all of their genes. In a market economy, variation and selection are also at work. New ideas are created by scientists and engineers, meticulous middle managers in large corporations or daring entrepreneurs. Failures are culled because bad ideas do not survive long in the market place: to succeed, you have to make a product that customers wish to buy at a price that covers costs and beats obvious competitors. Many ideas fail these tests, and if they are not shut down by management they will eventually be shut down by a bankruptcy court. Good ideas spread because they are copied by competitors, because staff leave to set up their own businesses, or because the company with the good ideas grows. With these elements of variation and selection in place, the stage is set for an evolutionary process; or, to put it more crudely, solving problems through trial and error.
This is all rather counter-intuitive, not to say uncomfortable. Many people assume that top corporate executives must be good for something: the shareholders who pay them handsome salaries certainly do; as do the millions of people who buy books purporting to convey the wisdom of successful business chiefs. Tetlock’s experts were almost helpless in the face of the complex situations he asked them to analyse. Are chief executives just as impotent, fumbling around for workable strategies in an impenetrable fog?
That would be what the evolutionary analogy implies. In biological evolution, the evolutionary process has no foresight. It is the result of pure trial and error over hundreds of millions of years. Could that also be true in an economy, despite the best efforts of managers, corporate strategists and management consultants?
A compelling clue comes from the economist Paul Ormerod. Ormerod had been reviewing what the fossil record tells us about extinctions over the last 550 million years – including mass extinctions that make the death of the dinosaurs look almost trivial. That record revealed a clear relationship between the scale of an extinction event and how often such events occur: if the extinction event is twice as severe, it is four times as rare; if it is three times as severe, it is nine times as rare. Eras in which very few extinctions take place are the most common of all. The pattern is very clear, and biologists now have mathematical models that show how a blind evolutionary process, when combined with an ever-changing contest for resources and the occasional asteroid strike, produces this distinctive signature.
Ormerod is a blunt, widely-read iconoclast from Lancashire in northern England, with a taste for disarming fellow economists with their own favourite weapon – mathematics. He decided to take a look at the data for corporate extinctions as well. He studied Leslie Hannah’s statistics on the death of corporate titans and compared them with half a billion years of data from the fossil record. The timescales were different, but the relationship between the size of an extinction event and its frequency proved to be exactly the same. (By far the worst year for the corporate titans was 1968, when six of them ‘died’.) Next, Ormerod turned to a much bigger database of smaller corporate extinctions in the United States, state by state, sector by sector, with thousands of data points describing literally millions of small companies. He discovered the same thing. He cast the net even wider, looking at corporate extinctions across eight other rich countries. He found the same thing again.
Biological extinctions and corporate extinctions share that special signature. This does not prove that the economy is an evolutionary environment and that corporate strategies evolve through trial and error rather than successful planning, but it does offer a big hint. And Ormerod went further, again building on work by biologists. He took a stripped-down mathematical model of biological extinction that produced the tell-tale extinction signature and adapted it to represent corporate life and death. But he added a twist: he changed the rules of his model so that some companies were allowed to be successful planners. These planners were able to adjust their strategies to maximise the advantage they gained from interacting with other companies in the economy; some could do this perfectly, while others had just a tiny edge over a company whose strategy was determined entirely at random.
Ormerod discovered something disturbing: it was possible to build a model that mimicked the real extinction signature of firms, and it was possible to build a model that represented firms as modestly successful planners; but it was not possible to build a model that did both. The patterns of corporate life and death are totally different from reality in the ‘planning is possible’ model, but uncannily close to reality in the ‘planning is impossible’ model. If companies really could plan successfully – as most of us naturally assume that they can, despite what Tetlock tells us about the limitations of expert judgement – then the extinction signature of companies would look totally different to that of species. In reality, the signatures could hardly be more similar.
We should not leap to conclusions based on an abstract mathematical model, but Ormerod’s discovery strongly implies that effective planning is rare in the modern economy. I wouldn’t go so far as to suggest that Apple might as well replace Steve Jobs with a dart-throwing chimpanzee – even though it would certainly liven up Apple product launches. But the evidence suggests that in a competitive environment, many corporate decisions are not successful, and corporations constantly have to cull bad ideas and search for something better.
The same conclusion is suggested by Tetlock’s studies of expert judgement and by the history of ‘excellent’ companies that so often lose their way: we are blinder than we think. In a complex, changeable world, the process of trial and error is essential. That is true whether we harness it consciously or simply allow ourselves to be tossed around by the results.
While trial and error is fundamental to the way that markets work, it makes for a challenging approach to life. Who wants to grope her way to a successful solution, with her repeated failures in full view of the world? Who wants to vote for a politician who takes that approach, or promote a middle manager whose strategy seems to be to throw around random ideas and see what works? Remember that President George W. Bush vowed to ‘stay the course’ while his opponent, John Kerry, lost the Presidential election in part because he had a reputation for changing his mind. Kerry’s fans agreed that ‘flip-flopper’ was an insult, although they felt it was ill-deserved. But if we took trial and error seriously, ‘flip-flopper’ would be a badge of flexibility, worn with pride. A similar attitude prevails in British politics. Margaret Thatcher famously declared, ‘You turn if you want to. The lady’s not for turning.’ Tony Blair was proud of the fact that he didn’t have a reverse gear. Nobody would buy a car that didn’t turn or go backwards, so it is unclear why we think of these as desirable qualities in Prime Ministers. But British voters rewarded Thatcher and Blair for their self-professed lack of adaptability with three general election victories apiece.
But whether we like it or not, trial and error is a tremendously powerful process for solving problems in a complex world, while expert leadership is not. Markets harness this process of trial and error, but that does not mean that we should leave everything to the market. It does mean – in the face of seemingly intractable problems, such as civil war, climate change and financial instability – that we must find a way to use the secret of trial and error beyond the familiar context of the market.
We will have to make an uncomfortable number of mistakes, and learn from them, rather than cover them up or deny they happened, even to ourselves. This is not the way we are used to getting things done.
A railroad foreman named Phineas Gage has the unfortunate distinction of being the world’s most famous victim of a brain injury. In 1848, he was preparing an explosive charge when it detonated unexpectedly, driving his tamping iron – a rod over a yard long and an inch thick – through his cheek, behind his left eye, through his left front brain and out of the top of his head. The rod landed eighty feet away. Astonishingly, Gage survived, but his character was changed radically: previously sober and reliable, he became feckless, stubborn, unable to settle on any plan and prone to yelling obscenities. Along with a chunk of his brain, a particular part of his mind had gone. His friends said he was ‘no longer Gage’.
The Soviet Union is to economics what Phineas Gage is to neuroscience. Neuroscientists study patients with damage to specific regions of the brain because their plight illuminates how the brain is ordinarily supposed to work. In much the same way, economists study dysfunctional economies when attempting to figure out the secrets of healthy ones. It is of course not a new insight that the Soviet system failed, but the unexpected details of why it failed are often glossed over – and they hold an important lesson for our mission to understand how to harness trial and error to solve problems.
The story starts in Russia’s coal-rich Don Basin, north of the Black Sea, in 1901, before the Soviet Union even existed. A twenty-six-year-old engineer named Peter Palchinsky was sent by the Tsar’s government to study the area’s coal mines. Palchinsky gathered reams of data, paying attention to every local detail, and in particular building up a dossier on working conditions. The miners, he discovered, were housed forty or even sixty to a room, stacked in shared wooden bunks like cheap goods in a warehouse. In order to sleep, they had to crawl into position from the foot of the bed because there was no headroom to clamber over their fellows. Toilets and other facilities were rudimentary.
When Palchinsky sent back his findings, his superiors realised that his research was political dynamite: Palchinsky was sent to Siberia to perform less sensitive assignments. Palchinsky and his stubborn streak were inextricably linked. A few years earlier, winning a place at Russia’s top engineering school, he had taken pride that he had based his application on the strength of his exam results, rather than relying on the right connections. In short, Palchinsky was bright, energetic, confident – and almost absurdly honest.
Palchinsky’s early brush with the authorities worked to his advantage. He slipped across the Russian border to work in Western Europe. Palchinsky soaked up knowledge in Paris, Amsterdam, London and Hamburg, making copious notes on the new industries those cities were developing, and paying just as much attention to new ideas in management as in engineering. He wanted to absorb the latest thinking on organising a workforce as well as cutting-edge science and technology. Hungry to understand as much as he could, he became a successful industrial consultant, and was as eager to spread expertise as to gain it.
Incredibly, Palchinsky began writing articles suggesting suitable reforms for the Russian economy, advising the very Tsarist government that had exiled him to Siberia. But that was Palchinsky through and through: he just couldn’t stop telling it the way he saw it. He wrote letters to his wife Nina freely admitting that he had had an affair while travelling in Europe. (She received the news stoically.) When he returned to Russia after receiving a pardon in 1913, he became an influential adviser to the Tsar’s government, and – after narrowly escaping being bayoneted during the revolution – later he advised the Soviet government, too. But his stubborn honesty continued: he refused to join any scientific or engineering organisation that was controlled by the Communist Party, on the grounds that engineering advice should not be distorted by politics. He frequently criticised foolhardy engineering. He even drafted a letter to the Soviet leadership, offering the helpful observation that technology and science were more important than communism; friends begged him not to send it, and he relented.
Yet while Palchinsky’s political antennae were missing, his technical judgement and humanitarian instincts were sharp. He warned against prestige projects: why drill oil wells just for the spectacular ‘gush’ when cheap coal and gas were widely available? He defended small projects that, according to his own painstaking research, were often more efficient than gigantic ones. He defended workers’ rights throughout.
It is easy to forget just how successful the Soviet economy was … for a time. We tend to assume that the planned economy fell apart because it lacked the galvanising force of the profit motive and the creativity of private-sector entrepreneurs. But this does not really make sense: there were many creative people in the Soviet Union, including Palchinsky. It is not immediately obvious why they would lose their creativity merely because they worked for state-owned enterprises. Nor did the Soviet Union lack motivational techniques: in fact, it possessed as great a range of incentives, positive and horrifyingly negative, as any civilisation in history, and deployed them ruthlessly. And the results were initially impressive. So much so that, by the 1950s, many Western experts had concluded that communism – while anti-democratic and cruel – was more effective than capitalism as a way to run an economy.
The Soviet failure revealed itself much more gradually: it was a pathological inability to experiment. The building blocks of an evolutionary process, remember, are repeated variation and selection. The Soviets failed at both: they found it impossible to tolerate a real variety of approaches to any problem; and they found it hard to decide what was working and what was not. The more the Soviet economy developed, the less of a reference point the planners had. The whole system was unable to adapt.
Peter Palchinsky, with his international experience and his painstaking analysis of local conditions, was just the kind of man who could have changed that. He was assigned to advise on two of the most important projects in Stalin’s first five-year plan: the Lenin Dam and Magnitogorsk. The Lenin Dam, on the Dnieper River in modern Ukraine, was the world’s largest when it was commissioned in the late 1920s. Palchinsky was unmoved by its scale. Stalin’s brainchild it might be, but he warned that the river was too slow and, on a flood plain, the reservoir would be huge and would swamp many thousands of homes and much prime farming land. Nobody knew how much, he pointed out, because no hydrological surveys had been carried out; but the reservoir eventually proved to be so large that simply growing hay on the land it had covered and burning it in a power plant would have generated as much energy as the dam did. There was a dry season, Palchinsky admonished, so coal-fired power stations would have to be built and run for three months a year in any case. He advocated a step-by-step approach as the local economy expanded, combining small coal-fired plants with more modest dams. He pointed out that smaller dams would likely be more effective. In every detail, his concerns were later proved correct. But Stalin was not interested: he simply wanted the world’s largest hydroelectric project and gave the order to proceed anyway. The project suffered huge cost overruns and was an economic and engineering disaster, even setting aside the ecological costs, the forced relocation of ten thousand farmers and the appalling labour conditions.
The steel mills of Magnitogorsk, the ‘City of Magnet Mountain’, were if anything more ambitious. The city was to be built in the remote heart of Russia, far to the east of Moscow, but near apparently plentiful iron-ore deposits. It was designed to exceed the entire steel output of the United Kingdom. Again, Palchinsky counselled caution – he wanted more analysis and a step-by-step approach. His old studies of workers’ conditions in the coal mines of the Don Basin led him to worry about the fate of Magnitogorsk’s workers. But he also pointed out the key technical objections to the project, which seemed to be cast from the same mould as the Lenin Dam: it was begun without a detailed study of the area’s geology and without any interest in the availability of the coal needed to fire the mills.
Palchinsky’s warnings were ignored, and again they were horribly accurate. One witness described conditions on the cattle wagons transporting workers to the site: ‘For a day and a half, the door was not even opened … mothers had children die in their arms … From only the wagon in which we travelled, four little corpses were removed. More were carried out from other wagons.’ Over three thousand people died in the first winter of construction work. Promised a garden city, Magnitogorsk’s forced labourers were housed downwind of the blast furnaces. The iron ore ran out in the early 1970s, and then both coal and iron had to be shipped to what were the world’s largest steel mills over vast distances. When the US historian Stephen Kotkin spent time living in the city in 1987, he discovered endemic alcoholism, shortages of almost everything, crumbling infrastructure, ‘almost unfathomable pollution and a health catastrophe impossible to exaggerate’.
What Palchinsky realised was that most real-world problems are more complex than we think. They have a human dimension, a local dimension, and are likely to change as circumstances change. His method for dealing with this could be summarised as three ‘Palchinsky principles’: first, seek out new ideas and try new things; second, when trying something new, do it on a scale where failure is survivable; third, seek out feedback and learn from your mistakes as you go along. The first principle could simply be expressed as ‘variation’; the third as ‘selection’. The importance of the middle principle – survivability – is something which will become clear in chapter six, which explores the collapse of the banking system.
The monstrous moral flaws of the Soviet system are now obvious. The economic flaw was more subtle: its inability to produce variation and selection, and therefore its inability to adapt. Central planners decided what would be built, lulled into a sense of omniscience by having a map or a table of statistics in front of them. Such plans inevitably missed the messy complexities of the situation on the ground, and also produced far too little variation. Almost every apartment in 1960s Moscow had the same iridescent orange lampshade. In Magnitogorsk, there were two types of apartment, named ‘A’ and ‘B’. They were the city’s sole concession to variety.
Above all, feedback is essential for determining which experiments have succeeded and which have failed. And in the Soviet Union, feedback was ruthlessly suppressed.
One icy Leningrad night in April 1928, there was a knock on the door of Peter Palchinsky’s apartment. He was arrested by the secret police and was never seen by his wife again. Over a year later, it was announced that he had been executed. There had been no trial, but a secret police dossier on Palchinsky, unearthed and smuggled out of Moscow many decades later by the historian Loren Graham, documented his ‘crimes’. He was accused of ‘publishing detailed statistics’ and sabotaging Soviet industry by trying to set ‘minimal goals’. In other words, Peter Palchinsky was murdered for trying to figure out what would work, and for refusing to shut up when he saw a problem.
Palchinsky was not alone. Three thousand of the USSR’s ten thousand engineers were arrested in the late 1920s and early 1930s, mostly bound for near-certain death in Siberia. (Palchinsky’s wife, Nina, also met that fate.) Anyone who tried to object to looming technological disasters and to suggest alternatives was denounced as a ‘wrecker’. Palchinsky’s secret execution was unusual – perhaps because, stubborn to the end, he refused to recant. His persecution was not.
The Soviet Bloc began to fall apart in the late 1980s, punctuated with famous events such as the victory of the newly legalised Solidarity movement in the Polish elections of June 1989, and the fall of the Berlin Wall in November of that year. In the heart of the Soviet Union itself, a momentous but less famous revolt was also taking place: the first major strike in Soviet history. In July 1989, a quarter of a million coal miners walked away from their jobs. Part of the protest was about grotesquely dangerous conditions: the death rate for Soviet miners was fifteen to twenty times higher than it was for their American equivalents, with the local pits claiming the lives of over fifty men every month. But the strike was also provoked by simple deprivation: the miners often had no meat or fruit to eat, and few had access to soap or hot water. After risking their lives each day in the suffocating depths, they couldn’t even wash themselves or rest in a comfortable bed. President Mikhail Gorbachev was forced to appear on national television, acknowledging the justice of the miners’ cause and offering substantial concessions. It was a notable moment in the downfall of the Soviet system.
The miners who had walked out and humiliated Gorbachev worked, of all places, in the Don Basin. Sixty years after Peter Palchinsky’s execution, and eighty-eight years after he had initially pointed to the problem of working conditions in the Don coal mines, the Soviet system had still failed to adapt.
The Soviet Union, like poor Phineas Gage, is a grotesquely extreme example. Only the worst dictatorships have exhibited the same pathological immunity to feedback. Yet, in a gentler way, most organisations and most forms of politics have the same difficulty in carrying out the simple process of variation and selection.
Variation is difficult because of two natural tendencies in organisations. One is grandiosity: politicians and corporate bosses both like large projects – anything from the reorganisation of a country’s entire healthcare system to a gigantic merger – because they win attention and show that the leader is a person who gets things done. Such flagship projects violate the first Palchinsky Principle, because errors are common and big projects leave little room to adapt. The other tendency emerges because we rarely like the idea of standards that are inconsistent and uneven from place to place. It seems neater and fairer to provide a consistent standard for everything, whether it’s education, the road network or the coffee at Starbucks. Such uniformly high standards sound tempting: as Andy Warhol once commented, ‘You can be watching TV and see Coca-Cola, and you know that the President drinks Coke, Liz Taylor drinks Coke, and just think, you can drink Coke, too. A Coke is a Coke and no amount of money can get you a better Coke than the one the bum on the corner is drinking. All the Cokes are the same and all the Cokes are good.’
But Warhol found Coke intriguing because it was an exception; and it still is. Producing a sweet, fizzy drink is a static, solved problem. No further experimentation is necessary, and it is perfectly possible to set uniformly high standards in the production of Coca-Cola. (The delivery of Coke to remote parts of the world is another matter, and is a minor miracle of local adaptation.) Ensuring uniformly high standards in more complex situations is much harder: it’s the chief achievement of Starbucks and McDonald’s, and even then the standardisation comes at a price in charm, flexibility and quality.
Running a hospital or a school is another matter altogether. We love the idea that every single one should deliver the same high quality. In the UK, we even have our own catchphrase, the ‘postcode lottery’, to describe the scandal that standards vary from place to place. It is something of a national obsession. We want all of our public services to be like Coca-Cola: all identical, all good. And they can’t be.
If we are to take the ‘variation’ part of ‘variation and selection’ seriously, uniformly high standards are not only impossible but undesirable. When a problem is unsolved or continually changing, the best way to tackle it is to experiment with many different approaches. If nobody tries anything different, we will struggle to figure out new and better ways to do anything. But if we are to accept variation, we must also accept that some of these new approaches will not work well. That is not a tempting proposition for a politician or chief executive to try to sell.
It seems to be equally hard for traditional organisations to deliver the selection component of variation and selection. The difficulty is in selecting what is really working on the ground. Peter Palchinsky was all for taking things step by step, but politicians resist pilot schemes with objective measures of success. This is partly because politicians are in a hurry: they expect to hold on to a role for two to four years, not long enough for most experiments to deliver meaningful results.* Even more politically inconvenient is the fact that half of the pilot schemes will fail – many things do in a complex world – so the pilot will simply produce stark evidence of that failure. This is our fault as much as the fault of our politicians. We should tolerate, even celebrate, any politicians who test their ideas robustly enough to prove that some of them don’t work. But, of course, we do not.
It’s a sad truth that one of the most successful pilot schemes of recent years was implemented not by politicians but by a celebrity chef and a television crew. Jamie Oliver, chirpy Essex boy turned darling of the British middle class, created a national phenomenon in 2005 when he tried to persuade schools to serve healthier meals. Almost accidentally, he created a reasonable approximation of a controlled experiment. He convinced schools in the London borough of Greenwich to change their menus, and then mobilised resources, provided equipment and trained dinner ladies. Other London boroughs with similar demographics received none of these advantages. Indeed, because the resulting television programme wasn’t broadcast until after the project was well under way, they probably knew little about it.
Two economists, Michele Belot and Jonathan James, picked up the data generated by the cheeky chef’s campaign and analysed it, discovering that if primary-school kids eat less fat, sugar and salt, and more fruit and vegetables, they are ill less often and do somewhat better at English and science. These conclusions would be more robust if the trial had been rigorously controlled, but until Jamie Oliver came along, none of the country’s politicians had shown much interest in the experiment. Tony Blair, then British Prime Minister, fell over himself to endorse the campaign. He had been in power for eight years at the time.
If formal experiments hold few joys for traditional leaders, informal feedback will often fail to reach them, too. Few advisers face Peter Palchinsky’s fate, but even so his compulsion to blurt out the truth is rare. There is a limit to how much honest feedback most leaders really want to hear; and because we know this, most of us sugar-coat our opinions whenever we speak to a powerful person. In a deep hierarchy, that process is repeated many times, until the truth is utterly concealed inside a thick layer of sweet-talk. There is some evidence that the more ambitious a person is, the more he will choose to be a yes-man – and with good reason, because yes-men tend to be rewarded.
Even when leaders and managers genuinely want honest feedback, they may not receive it. At every stage in a plan, junior managers or petty bureaucrats must tell their superiors what resources they need and what they propose to do with them. There are a number of plausible lies they might choose to tell, including over-promising in the hope of winning influence as go-getters, or stressing the impossibility of the task and the vast resources needed to deliver success, in the hope of providing a pleasant surprise. Actually telling the unvarnished truth is unlikely to be the best strategy in a bureaucratic hierarchy. Even if someone does tell the truth, how is the senior decision-maker supposed to distinguish the honest opinion of a Peter Palchinsky from some cynical protestation calculated to win a budget increase?
Traditional organisations are badly equipped to benefit from a decentralised process of trial and error. Static, solved problems are ideal for such organisations; as are tasks where generalised expertise counts for much more than local knowledge. But such ‘Coca-Cola problems’ are increasingly rare in a rapidly changing world, which is why – as we shall see – many businesses are beginning to decentralise and strip authority away from managers. In the next chapter, we’ll see how adaptive organisations need to decentralise and become comfortable with the chaos of different local approaches and the awkwardness of dissent from junior staff. We’ll also see the heroic effort required to force a traditional hierarchy to change its mind.
But there is a more fundamental problem here than the right way to design an organisation, because it isn’t just organisations that struggle to acknowledge and adapt to their mistakes. Most individuals suffer from the same problem. Accepting trial and error means accepting error. It means taking problems in our stride when a decision doesn’t work out, whether through luck or misjudgement. And that is not something human brains seem to be able to do without a struggle.
I spent the summer of 2005 studying poker. I interviewed some of the best players in the world, attended the World Series of Poker in Las Vegas, analysed ‘pokerbots’ – poker-playing computers – and chronicled the efforts of highly rational players, such as Chris ‘Jesus’ Ferguson, a game theorist with a PhD who is a world champion and a formidable one-on-one player.
While poker can be analysed rationally, with big egos and big money at stake it can also be a very emotional game. Poker players explained to me that there’s a particular moment at which players are extremely vulnerable to an emotional surge. It’s not when they’ve won a huge pot or when they’ve drawn a fantastic hand. It’s when they’ve just lost a lot of money through bad luck (a ‘bad beat’) or bad strategy. The loss can nudge a player into going ‘on tilt’ – making overly aggressive bets in an effort to win back what he wrongly feels is still his money. The brain refuses to register that the money has gone. Acknowledging the loss and recalculating one’s strategy would be the right thing to do, but that is too painful. Instead, the player makes crazy bets to rectify what he unconsciously believes is a temporary situation. It isn’t the initial loss that does for him, but the stupid plays he makes in an effort to deny that the loss has happened. The great economic psychologists Daniel Kahneman and Amos Tversky summarised the behaviour in their classic analysis of the psychology of risk: ‘a person who has not made peace with his losses is likely to accept gambles that would be unacceptable to him otherwise’.
Even those of us who aren’t professional poker players know how it feels to chase a loss. A few years ago, my wife and I had booked a romantic weekend in Paris. But she was pregnant, and a couple of hours before we were due to catch the train she began feeling sick. She was throwing up into a plastic bag in the taxi on the way to the station. But when I met up with her, she was determined to go to Paris because our tickets weren’t refundable. She didn’t want to accept the loss and was about to compound it.
Being an economist is rarely an advantage in a romantic situation, but this was perhaps an exception. I tried to convince my wife to forget about the tickets. Imagine that the money we had spent on them had been lost for ever, I told her, but also imagine that we stood on the steps of Waterloo station with no plans for the weekend, when somebody came up to us and offered us free tickets to Paris. That was the correct way to think about the situation: the money was gone; and the question was whether we wanted to travel to Paris for no further cost. I asked my wife whether she would accept such an offer. Of course not. She was feeling far too sick to go to Paris. She forced a faint smile as she realised what I was telling her, and we went home. (As if to confirm that we had made the right decision, the nice people at Eurostar refunded our tickets anyway. And a few months later, my wife somewhat more pregnant, we got to Paris in the end.)
The behavioural economist Richard Thaler, with a team of co-authors, has found the perfect setting to analyse the way we respond to losses. He studied the TV game show Deal or No Deal, which is a great source of data because the basic game is repeated incessantly, with similar rules, for high stakes, in over fifty countries. Deal or No Deal offers contestants a choice of between twenty and twenty-six numbered boxes, each containing some prize money, ranging from pennies to hundreds of thousands of dollars, pounds or euros. (The original Dutch version has a jackpot of five million euros.) The player holds one box, not knowing how much money is inside. Her task is to choose the other boxes in any order she likes. These are then opened and discarded. Every time she opens a box containing a token amount, she celebrates, because that means her own mystery box doesn’t contain that low prize. Every time she opens a box with a large prize, she winces, because that reduces the odds that her own box will be lucrative.
All of this is pure chance. The interesting decision is the one that gives the game show its title. From time to time, the ‘Banker’, a mysterious and anonymous figure, calls the studio to offer the contestant cash in exchange for the unknown sum inside her box. Will it be a deal, or no deal?
The psychology of the game is revealing. Let’s take a look at Frank, a contestant in the Dutch version of Deal or No Deal. After a few rounds, the expected value of his box – that is, the average of all the remaining amounts – was just over €100,000. The Banker offered him €5,000 – serious money, but less than 75 per cent of his box’s expected value. He turned it down. Then he received a nasty shock. Frank opened a box containing €500,000, the last big prize remaining. His expected winnings plunged to just €2,508. The Banker’s offer plunged, too – from €75,000 to €2,400. Relative to Frank’s likely winnings, this was a more generous offer than the previous one – 96 per cent of the expected value of playing on – but Frank rejected it. The next round, Frank spurned a Banker’s offer that was actually greater than the average value of the remaining boxes. And in the final round, Frank’s two remaining possibilities were €10 or €10,000. The Banker offered him a more-than-generous €6,000. Frank turned it down. He left the studio with €10. After being stunned by the loss of a guaranteed €75,000 and a decent chance of a €500,000 prize, Frank started taking crazy gambles. Frank had gone on tilt.
Frank’s behaviour is typical. Thaler and his colleagues looked at how people responded to the Banker’s offers immediately after making an unlucky choice, a lucky choice, or a choice that was broadly neutral. They found that the neutral choosers tended to be quite keen to accept the Banker’s deal. Lucky choosers were cocky: they were more likely to turn down the Banker and keep going. But it was unlucky choosers who stood out. They were extremely unlikely to accept an offer from the Banker.* Why? Because if they did, it would lock in their ‘mistake’. If they kept playing, there was a chance of some sort of redemption. The pattern was all the more striking because the Banker tended to make more generous offers to losers – lower in absolute terms, of course, but closer to the average of the remaining boxes. Objectively, players who had just made an unlucky choice should have been more willing to deal than anyone, because they were receiving more attractive odds from the Banker.
Perhaps this is a phenomenon restricted to game shows and the poker tables of the Rio in Las Vegas? No such luck. The economist Terrance Odean has found that we tend to hang on grimly, and wrongly, to shares that have plunged in the hope that things will turn around. We are far happier to sell shares that have been doing well. Unfortunately, selling winners and holding on to losers has in retrospect been poor investment strategy.
All four examples – poker, Paris, Deal or No Deal and share portfolios – show a dogged determination to avoid crystallising a loss or drawing a line under a decision we regret. That dogged determination might occasionally be helpful, but it is counterproductive in all these cases and in many others. Faced with a mistake or a loss, the right response is to acknowledge the setback and change direction. Yet our instinctive reaction is denial. That is why ‘learn from your mistakes’ is wise advice that is painfully hard to take.
We face a difficult challenge: the more complex and elusive our problems are, the more effective trial and error becomes, relative to the alternatives. Yet it is an approach that runs counter to our instincts, and to the way in which traditional organisations work. The aim of this book is to provide an answer to that challenge.
The adaptive, experimental approach can work almost anywhere, so we’ll look at a huge range of problems. We’ll meet the rebellious colonels who risked their careers – and their lives – to change the shape of the war in Iraq; and the doctor whose desperate gamble in a wartime prison camp should serve as an example to the staff of the World Bank today. We’ll discover what the disasters at Three Mile Island and Deepwater Horizon have to tell us about preventing another Lehman Brothers crisis. We’ll learn from a watchmaker, a street urchin, a Wall Street rebel, two aircraft designers and a failing choreographer. We’ll study the corporate strategies of companies from Google to a simple high-street cobbler. We’ll search for solutions to problems from the banking crisis to climate change.
Along the way, we’ll also be learning about the recipe for successfully adapting. The three essential steps are: to try new things, in the expectation that some will fail; to make failure survivable, because it will be common; and to make sure that you know when you’ve failed. Palchinsky would have recognised these steps, but they involve formidable obstacles. To produce new ideas we must overcome our tendency to fall in step with those around us, and overcome those with a vested interest in the status quo. Making failure survivable sometimes means taking small steps, but not always: many innovations emerge from highly speculative leaps, and surviving such leaps is not easy. Nor is it easy to survive a failure in the financial system. And distinguishing success from failure, oddly, can be the hardest task of all: arrogant leaders can ignore the distinction; our own denial can blur it; and the sheer complexity of the world can make the distinction hard to draw even for the most objective judge.
Along the way, I hope that we’ll learn something about how to adapt and experiment in business and in our own lives. Faced with the costs and risks of trial and error, should you and I try to experiment and adapt more than we do? What price would we pay in our quest to succeed?
*Donald Green, Professor of Political Science at Yale, tells me that one question in the social sciences has been thoroughly tested with field experiments: how to get out the vote. So politicians can use rigorous evaluation methods when it suits them.
*Typical contestants accepted the Banker’s offer 31 per cent of the time. ‘Winners’ accepted the offer 25 per cent of the time. ‘Losers’ accepted only 14 per cent of the time, despite receiving objectively more generous offers.