Relationships seem to be more important in Chinese business than they are in American business. This is sometimes attributed to cultural differences, but I am not sure this is a satisfactory explanation – indeed it is not really an explanation at all. Instead, I propose that the importance of 关系 or relationships in China can be explained as a rational response to differences in the legal system. More generally, cultural differences can often be best explained as rational responses to different environments.

The US has a very well-developed, perhaps overdeveloped, legal system; China’s legal system is still developing. A business person in the US can write a fairly complex contract and feel confident that he can enforce the contract relatively easily and cheaply. This is not as true in China (although like most things here, the legal system is changing rapidly). Now suppose our US business person has a choice between two suppliers offering similar quality products: one who is a friend, but whose price is higher; one who is a stranger, but whose price is lower. Let us also assume that the businessman has good information about both companies and they are equivalent – he just doesn’t have a personal relationship with the latter. With enforceable contracts, the businessman may feel comfortable in offering the contract to the latter. In China, however, the contract may not be enforceable – the Chinese businessman can not rely on the legal system to prevent the supplier breaching their contract. As a result, he may rely on his personal relationship to guarantee supply. So he will choose to work with his friend rather than the stranger.

If this explanation is true, we should see 关系 become less important as the legal system improves. I think there is some evidence of this. In the more developed parts of China (Shanghai, Shenzhen) it appears 关系 are less important than in less developed areas (say Henan).

关系 are likely to remain important when dealing with the government. It is difficult to sue the Chinese government in China – and even if it were possible, it would likely lead to the loss of future government business.

WARNING: I talk about Fellini’s 8½ in this post, so if you have not seen the film and want to watch it fresh, you may want to skip this post until you have seen it. Having said that, this is not the sort of film for which you can really have spoilers, so probably this post will not affect your enjoyment of the film.

The main purpose of this blog is to look for connections between ideas in apparently unconnected fields. One such connection occurred to me last night after watching Fellini’s 8½ – a movie about a movie director going through a crisis both in his professional and personal life. At one point, the director, Guido, is talking to a friend of is wife’s, Rossella. He asks her when he went wrong, and she, referencing the spirits she believes in, replies “They say you’re free, but you must learn to choose. You don’t have much time. And you have to hurry.” To me this was the central theme of the film: Guido’s problems stem from an inability to choose.

In his personal life, he is unable to choose between the multiple women in his life. When his wife recognizes his mistress, he refuses to admit to his affair even though she knows, and he knows she knows, that he is lying. Why continue to lie when you have been found out? I think he continued to lie because it enabled him to avoid choosing. If he admitted to the affair, he would be forced to make a decision; by continuing to lie, he could avoid that decision.

In his professional life he is making a movie that seems to lack any sort of focus. At one point he says, “ln my movie l have all sorts of things happen. l’m putting everything in – even a tap-dancing sailor.” This does not seem far from the truth – he has screen tests for the part of wife, mistress, etc (paralleling the parts in the film) but also builds a huge 70 meter tall model of a spaceship. Throughout the movie his producer, actresses and others involved with the movie ask him questions which he avoids. Near the end of the movie he is shown being unable to answer a bombardment of questions at a press conference for the launch of the movie.

The common theme is an inability to choose – a reluctance to shut the door on one opportunity in order to focus on another. I can sympathize with this Guido in this problem as it is one I share. Keeping options open – refusing to choose – does have value. In business, the concept of real options allows us to analyze and value this freedom. Therefore, I think it is interesting to consider Guido’s problem from this perspective.

Why might you want to keep options open? From real options perspective, we keep options open so that we can delay a decision until after we have learned some new information relevant to that decision. For example, suppose you have to decide on a location for a new factory. You may want to wait until the local government makes a decision about where to route a new highway. By delaying the decision until after the information arrives you can make a better decision. However, for almost any decision, you will always receive new relevant information if you wait – there is always an option value to waiting – and in the absence of other factors we would never make a decision.

Why might you want to make a decision now? Well that may depend on the particular decision. For example, in the factory placement decision, there is a time cost of money – profit now is worth more than profit in the future – and you may have a time limit on a certain option – someone else might buy the land you were considering for your factory. When you make a decision involves trading the option value of waiting against the costs of waiting.

How does this apply to Guido?

Well the first point is that Guido may feel they do not apply. Real options are usually applied to decisions where there are mutually exclusive choices: you can buy a stock or not; locate a factory in place A or place B. Do Guido’s decisions fit? For example, does he have to choose between his wife and mistress or can he have both? I think it is clear that he would like to try to keep both. Maybe he is delaying the decision not to learn more information, but to avoid the decision entirely. In the case of the movie, trying to include everything (even a tap-dancing sailor) is also an attempt to avoid making a decision. However, the message of the film, I think, is that this strategy has failed for Guido. He must make a decision – avoiding making a decision leads to a worse outcome than making a bad decision. In fact, the decisions may be exclusive: he must choose the wife or the mistress (or he will lose both); he must choose what and who to include and exclude from the movie or he will have no movie. If this is the case, we can consider his decisions using the framework outlined above.

What is the information Guido wants to gain by delaying the decisions about his movie? This is tough to say – perhaps some internal inspiration? It seems there is no obvious reason to believe that by waiting the situation will improved. On the other hand, there are clear costs – both financial to the producer and in terms of uncertainty and loss of morale for all concerned – including Guido. This is probably a situation where a decision needs to be made quickly (as suggested by Rossella’s spirits).

What information would Guido gain by waiting about his relationships with his wife and mistress? Probably nothing. It is true that people and relationships change over time so you will get new information, but once you know someone well (as Guido appears to know both these women), the value of waiting seems relatively low. The cost of waiting is the stress it places on the women and the relationships. Again, in Guido’s position, it seems a quick decision would be beneficial.

I think the main point is that in both his personal and professional life, Guido is paralyzed – he cannot make a decision. I think this paralysis comes from wanting to avoid closing off any options. He wants both his wife and his mistress. He wants his film to have everything. However this refusal, or inability, to make a decision appears to be very costly – instead of keeping all doors open, it may lead to all doors closing. I think Guido can sense the doors closing which explains his panic. However, even at the end of the film, I am not sure he is able to decide. He says he will commit to his wife, but is this another lie?

In this post, I want to discuss an idea I did not discuss in the first part: the convexity of executive / investment banker remuneration.

Executive Remuneration

During the 1990s, there was a striking increase in the share of CEO pay that comes in the form of options and stock.

Figure 1: Options and Stock as a Share of Total CEO Pay 1989-2008

As we can see in figure 1, the share of total CEO pay coming in options and stock increased from about 20% in 1989 to about 80% in 2000. This is a major change and it is worth thinking about how it effects CEO incentives. The data comes from Forbes.

The first effect is to provide a strong incentive for CEOs to increase the stock price of their company. This is a good thing – they should, after all, be working for shareholders. Indeed, this is the main incentive effect for stock.

For options, however, there is a secondary incentive effect which is important. Options (normally call options), the payoff is convex. For a call option, the owner (in this case the CEO) has the right, but not the obligation, to buy the stock at a predetermined price, the strike or exercise price, before a certain date. At expiry, if the stock price is less than this exercise price, the option is worthless. If it is more than the exercise price, the option is worth the stock price minus the exercise price.

Consider an example. Suppose the exercise price is $100 and the option is about to expire. If the stock price is $80, how much is the option worth? Nothing. It gives you the chance to buy the stock for $100, but you can buy the stock in the market for $80, so you would not exercise the option. On the other hand, if the stock price is currently $120, you could exercise the option, and buy the stock for $100, making a $20 profit. The option would therefore be worth $20. We can see this in figure 2.

Figure 2: Convexity of Call Option Payoff

The red line shows the payoff at expiration (what we just calculated). The green line shows the payoff some time before expiry. When there is some time to expiry, there is uncertainty about where we will end up. Therefore we have to average over all the possible future stock prices to get an estimate of the value of the option. Exactly how we should do that is not a simple question and one I will not address here. However, notice that both the lines are convex. Given our previous discussion, this suggest that call options will encourage CEO’s to take risk. The increase in the use of options for remunerating CEOs will increase their desire to take risk.

Think of it this way: Suppose the share price is currently $100. If the CEO pursues a safe strategy, the price is unlikely to change much and the option will be worth very little. On the other hand, if he pursues a risky strategy, then either the share price will fall a lot (which doesn’t reduce the value of the option much) or the it will increase a lot (which greatly increases the value of the option). Essentially, he takes a bet in which he can not lose, but can win.

I believe these incentives are an important part of the cause not only of the excessive risk taken prior to the financial crisis, but also of some of the major fraud cases (Enron, Worldcom) from a few years ago. After all, lying is risky and will also be encouraged by options.

And the LORD God commanded the man, saying, Of every tree of the garden thou mayest freely eat:
But of the tree of the knowledge of good and evil, thou shalt not eat of it: for in the day that thou eatest thereof thou shalt surely die.

Genesis, 2.16-17

Eve took an apple from the Tree of Knowledge.

Eve took the first bite of the apple:

And the universe, with its stars, planets and immeasurable space was revealed to her. Long she searched before she found the Garden of Eden and its small planet. Inside she died a little. The first insight had reduced her world to an astrological speck.

Eve took the second bite of the apple:

And creatures, from the lowliest bacteria to the largest whale were revealed to her. Humans were but one of many in a constantly evolving milieu. Inside she died a little. The second insight had reduced her to a biological species.

Eve took the third bite of the apple:

And her brain, in all its complexity of structure and functioning was revealed to her. Everything about her was revealed, but there was no soul. Inside she died. The third insight had reduced her to a collection of atoms.

Adam returns

Adam found Eve and his ignorance allowed him to weep and to kiss her. God was angry and forbade Adam from eating from the Tree of Knowledge. As Adam slept, God fashioned a clone of Eve from one of his ribs. When Adam awoke he was no longer in the Garden of Eden, and the death of the first Eve appeared as a dream.

Adam had lost access to the Tree of Knowledge, but in that kiss, Adam too had consumed a morsel of the forbidden fruit. And from this seed, over many millenia, man’s knowledge would grow. Adam begat Copernicus begat Darwin begat Fleischbun.

Tree of Knowledge Manifesto, pp 1-2

Questioner: Are you a member of the terrorist group “The Tree of Knowledge”?

Fleischbun: I am not.

Questioner: Do you agree with them – that there is no such thing as a guilty person? That all prisoners should be set free? They draw these conclusions from your work, right?

Fleischbun I disagree with the second proposition. The first is more subtle. I argue that physical processes are deterministic and human beings are physical processes. Therefore, all are actions are deterministic. We do not have free will and cannot change our actions. In that sense, we are not guilty of bad actions any more than an earthquake is guilty of mass murder. For that matter, our great deeds are also deterministic.

Questioner: So why not free the prisoners if they are not guilty?

Fleischbun Whether you want to free the prisoners depends on why you are holding them in prison. There are basically two reasons to do so: to prevent them doing further harm, or to punish them. My argument would eliminate the second reason, but not the first. In other words, we should treat people with an abnormal brain that provides murderous urges (for example) the same way we would treat someone with an abnormal physical condition, say blindness. We do not consider that the blind person should be punished for being blind, but we may want to limit her freedom in some ways – for example, by not letting her drive a car. Similarly, we should not seek to punish the person with the murderous brain, but we may want to limit his freedom in some ways to prevent harm to others. This is not a new idea.

Questioner: But this is crazy liberal talk! I have free will. I can choose to pick up this pen or not. Denying free will defies common sense. It is just an excuse to deny reality.

Fleischbun How do you know you have free will? Psychologists have long known that introspection can be misleading. There are plenty of studies showing that people’s ideas and actions are systematically affected in ways that they do not realize. Advertisers and supermarkets (among others) rely on this.

Questioner: Well perhaps my choice of shampoo may be affected by such tricks, but not my decision to murder.

Fleischbun It is true that there are not experimental studies on murder or even lesser morally dubious actions for obvious reasons. However, there is evidence that by physically altering the brain we can systematically affect people’s choices in hypothetical ethical dilemmas.

Consider this situation: you and several others are hiding from enemy soldiers with a baby. The only way to stop the baby crying is to smother and kill it. If the baby cries, the soldiers will find and kill all of you. Should you kill the baby? Most people have two responses this: an initial emotional revulsion from the idea of killing a baby; and a second more rational evaluation that losing one life is better than losing many. Almost everyone will have both these thoughts and their final choice will depend on the relative strengths of those thoughts.

These two responses (emotional and rational) are triggered by different parts of the brain and yet another part is used to resolve conflicts between them (this has been shown in neuroimaging studies). If we can impair the functioning of one part of the brain (say that dealing with the rational response) we should be able to affect the ethical choices of the person. One study found that if they asked the question to subjects under cognitive load (ie those engaged in activities that required a lot of rational processing) were less willing to kill the baby. The emotional impulse overpowered the weakened rational impulse.

Those people who behave unethically have abnormal brains that exaggerate or suppress certain of the processes we go through when making moral judgments. The are not guilty any more than the blind person, though if not constrained in some way, they may, like a blind driver, be dangerous.

Questioner I can’t accept that everything was determined at the big bang and ever since then the universe (including us) have just been following a predetermined path. What about quantum uncertainty?

Fleischbun Suppose a roll a die. For any individual roll, I do not know which number will come up, but I do know the probability rules. The process is stochastic, but deterministic. The die must follow the rules – it has no free will to change them. And, if it follows the rules, over many roles, it becomes very predictable. Each number will come up 1/6th of the time. Quantum uncertainty is like this. It follows very precise probabilistic rules that, when aggregated to the scale of humans or our brain cells, follow very precise non-probabilistic rules. There is no room for free will.

Questioner So are you saying that we have no control over our lives at all? We must just passively accept the lives determined for in the Big Bang and then die – with no life after death. That is depressing.

Fleischbun The soul represents what we do not understand about the brain – just as ancient gods represented what people did not understand about the natural world. As science has advanced, the role for the soul has become smaller and will eventually cease to exist. Everything we do is determined by our brain so when that dies, so do we. There is no life after death, heaven (or hell) or reincarnation. It is interesting to me that while some ancient ideas (gods, soul) are becoming obsolete, others (such as fate) may be reemerging.

Whether this is seen as depressing will depend on the person. I find it liberating. I no longer feel stress over what to do with my life – it is predetermined – I can just relax and enjoy the journey.

Police Transcript, Questioning of Dr P Fleischbun, 2014/11/14

Baozi (包子, 370-318 BC) was a great, but sadly forgotten sage who was a contemporary of both Mencius (孟子, 372 – 289 BC) and Zhuangzi (庄子, 369–286 BC). In this post, I want to discuss his contributions to different fields and bring some small measure of deserved attention to his work.

Introduction

His philosophy attempted a radical fusion of the two dominant schools of thought of his day (Confucianism and Taoism). Baozi saw these two schools as each capturing half of the truth – Confucianism embraced society, but rejected nature; Taoism embraced nature and rejected society. In addition to seeking a unification of these two themes, Baozi wanted ground philosophy in physical reality – reality accessible to all – rather than the abstractions of the intellectual elite. He believed that philosophy should be explained in terms ordinary people understood and he attempted to create connections with the ideas they were engaged with in their lives. (The desire to build connections is part of the inspiration for this blog).

Unfortunately, very little reliable information has come down to us about his thought. Apparently he wrote nothing down – possibly because he focused on reaching ordinary (illiterate) people, or possibly because he saw writing as static and true learning as dynamic. As a result, we are forced to rely on second-hand records – a process complicated by the hostility with which he was regarded by the dominant intellectuals of the era. No doubt this goes some way towards explaining his current anonymity.

Philosophy

It appears that much of Baozi’s philosophy involved using meat buns as metaphors. Indeed, the current Chinese term for meat bun (包子) probably arose from crude attempts by rival schools of thought to ridicule Baozi’s philosophy by equating it literally (not metaphorically) with a meat bun. One view is that Baozi claimed the world was a giant meat bun. (There is a story, probably apocryphal, that Mencius challenged Baozi to eat a bun made of earth to prove his theory). This was clearly not meant to be taken literally, but as a metaphor. While it is not clear what Baozi meant by this metaphor, the possibilities suggest he may have been a thinker on a par with the great minds of history.

Scientific

The first, and most obvious, implication is that the earth is, roughly, spherical. This idea was first vaguely alluded to by Pythagoras, and subsequently by Plato and Aristotle – the latter a contemporary of Baozi. If Baozi had some scientific reason for his hypothesis, this alone would make him a key figure in intellectual history.

A more subtle, but perhaps even more remarkable, interpretation is that he had in mind a structure of the earth similar to that of a meat bun in which a solid, cool outer crust contains a hot, liquid interior with a hot solid core. This matches, exactly, the current view of the structure of the earth – something which was not universally accepted until the 20th century. To my knowledge, Baozi may have been the first to suggest this idea.

S. G. Chessington, in his controversial, and admittedly now somewhat discredited, work “Plato’s Cave: A History of Metaphor in Science” posited a radical reinterpretation of the meat bun theory. Baozi was not considering the earth, but the universe (a subtle distinction for his era), the universe was not a bun, but made of buns. This viewpoint can be profitably compared with the atomistic theory developed by earlier Indian and Greek philosophers (from one of whom Democritus, we get the modern term “atom”). Those philosophers argued that there must be some indivisible unit of matter. With recent discoveries of a string of subatomic particles, this claim seems questionable. However, Baozi’s bun theory holds up a little better. Explicit in the bun theory is that everything is made from something with a similar structure (the bun), but (unlike the atom) this something has it’s own structure. Furthermore, the structure of the bun is at least suggestive of the early 20th century model of the atom – a nucleus surrounded by an electron cloud. Indeed the quantum nature of electrons (are they particles or waves?) is suggested by the surface of the bun – should we think of it as one surface, or as a collection of flour particles?

Socio-political

The implications of meat-bun philosophy for social and political theory is less clear. The general consensus (formed around an idea first outlined in an anonymous 19th century pamphlet rumored to have been written by the Argentinian social critic Julio Alberdi) was that the spherical meat-bun social theory opposed the pyramidal theory of Confucianism. In Confucianism, power and responsibility flows downwards from the emperor through the various layers of government to the head of households and on down to the lowest household members. Loyalty flows in the opposite direction. In meat-bun theory, an “emperor” (if such a person would exist in a meat-bun society) occupies one point on the surface of the bun, but this is not fundamentally different from any other. Therefore, both power and duty flow to and from every one in society.

This theory is incredibly rich. Both democracy and communism can be seen as distorted forms of the pure theory. In democracy’s case, the equality of people is de-emphasized, while the circle of power is emphasized: leaders rule their subjects but are ultimately dependent on them for their power. In communism’s case, the circle of power is de-emphasized, while the equality of people is emphasized. The failures of each of these political systems may be traced to these distortions. In democracies, the circle of power is undermined by the inequality of wealth and subsequently power among people. In communist countries, the equality of people was undermined by the lack of a circular power structure that led to a concentration of wealth and power within a group of insiders (there are those who refer to Communism as practiced in the 20th and 21st century as Confucian communism – as opposed to meat-bun communism). While no fully-fledged social system devised by Baozi has come down to us, I suspect effort to elaborate and implement his ideas may prove fruitful. He may not only have been ahead of his time, but ahead of ours too.

Neglect

How is it possible that a great thinker such as Baozi could have been so comprehensively white-washed from history (there is not even a Wikipedia page and a Google search turns up only his culinary namesake)?

If you look at the dates posted at the beginning of this article, you will notice that while both Mencius and Zhuangzi lived for 83 years (to the best of our knowledge), Baozi only lived for 52 years. While 52 years was not as short a lifespan then as now, it still suggests he did not die of old age. We know that Baozi was unpopular both with the ruling classes (for whom his political theories represented a threat to their legitimacy) and with the Confucians (whose theories were inconsistent with meat-ball philosophy and whose fortunes were dependent on the current political structure). This combination represented powerful opposition to Baozi and his ideas.

While we know little about his death, contemporary accounts suggest that he was teaching in the state of Qi just around 320 BC. Around this time, Mencius became a Qi official. Though there are several historical references to meat-bun philosophy in Qi between 330 BC and 320 BC, there are none thereafter. Does this suggest some sort of political suppression of his ideas? Might that have contributed to Baozi’s death? We cannot be sure.

What is not in doubt is that meat-bun theories have been almost completely excluded from Chinese intellectual discussion since. Occasionally, there are signs that Baozi was not completely forgotten. For example, some scholars believe that the line 浮云游子意 (roughly translated as “while travelers’ intents are fleeting clouds”) from Li Bai’s 送友人 (Seeing a friend off) originally read 浮云包子意 (“while Baozi’s intents are fleeting clouds”). If we accept this alternative version, the poem changes from an emotional poem on parting with a friend to an ode to Baozi and a thinly-veiled political critique of Confucianism.

Conclusion

It is interesting to speculate how Chinese history might have differed if meat-ball philosophy rather than Confucianism had become the dominant paradigm. Not only would the explicitly Confucian imperial systems have been radically altered, but Chinese communism, in its current form, would be unthinkable. But we can not change history; we can only change the future. By acknowledging the significance of Baozi and his meat-ball philosophy, perhaps we can do just that.

The causes of the financial crisis, while numerous and interconnected, are not complicated. The perceived complexity is, in my view, an attempt by various groups to avoid the financial repercussions of their actions. In these posts, I want to illustrate some of the causes of the crisis and explain why the solutions that have been implemented, while alleviating the problem in the short-run, leave in place the systematic causes of the crises that will almost surely lead to similar, or worse, problems in the future. The ideas here are not necessarily original, but they are important enough to bear repetition.

Part 1 looks at mortgages, securitization and credit default swaps.

Part 1: Mortgages, securitization and credit default swaps

Securitization

In the ‘good old days’, banks lent money to people. If the people repaid the money, the bank made a profit; if they did not, the bank lost money. As a result, the banks carefully evaluated the credit of loan / mortgage applicants.

But there was a problem. Banks had long-term assets (loans), but short-term liabilities (deposits). This created risk. For example, if short-term interests rates went up relative to long-term interest rates, the gap between the interest rate the banks received from loans and that which they paid on deposits would fall. And so would their profits. What they needed was a way to turn their loans in to a short-term asset – like cash.

Securitization solved this problem. The banks sold the loans to a financial intermediary that packaged the loans together and sold shares in them to investors. The bank received cash; the investors received a share of the future loan repayments. This solved the banks balance sheet problem.

But there was a problem. When selling the loan, the bank also transferred the credit risk to investors. Now if a borrower defaulted, the bank did not lose money – the investor did. Investors believed these securitized loans (known as mortgage backed securities or MBS) were low risk (see below for why) and so were eager to buy them. Provided they could find borrowers, the banks could always sell the loans and make money. The credit risk didn’t matter. So banks became increasingly aggressive in competing for borrowers regardless of credit risk. The sub-prime mortgage market boomed.

Meanwhile, investors were buying more and more securitized mortgages. Why were they not worried about the risk? Two reasons: first, their models underestimated the risk; second, they bought insurance (credit default swaps). Let’s consider each in turn.

Underestimating risk

There are two types of risk: idiosyncratic and systematic. In terms of mortgages, an idiosyncratic risk is something that affects one (or a few) borrowers. For example, a borrower may suffer a health problem that causes them to default on their loan. This does not affect the ability of other borrowers to repay. Systematic risk is something that affect a significant proportion of all borrowers. For example, the financial crisis. Idiosyncratic risk is not a problem for investors as they can diversify that risk away. If I own shares in 10,000 mortgages, maybe a few people get sick and cannot repay, but the others will be OK, so I can accurately predict what repayments I will receive. Systematic risk is a problem. If there is a major recession and many borrowers default, the repayments I receive will drop sharply. I can not diversify this risk by buying more loans because it affects everyone. Therefore, idiosyncratic risk is fine; systematic risk is not.

So when determining their risk, investors need a way to estimate systematic risk. One way to do that is to look at history – how volatile have average repayment rates been in the past? Prior to the financial crisis, this risk looked low. Even if we considered a possible increase in the unemployment it seemed default rates would not increase greatly.

These estimates were inaccurate. Partly this was because the economic downturn was sharper than expected so there were more unemployed, etc. But not only that, regardless of their financial situation people became more likely to default. When house prices are rising, they will always be higher than outstanding mortgage debt. In this case, it does not make sense to default (you could always sell the house to pay off the debt). However, when house prices are falling sometimes the value of the house will be less than the money owned on the mortgage. Then it may make sense to default. Why pay off a debt worth $500,000 when you can just give the bank a house worth $300,000 and walk away? Well it is a little more complex because of the effects on credit rating, but it may still make sense. In other words, even some people who could repay may choose not to. The combination of these effects was not foreseen by models that had based on default rates during a period where house prices were rising.

Credit default swaps

Credit default swaps (CDS) are basically insurance. An owner of a mortgage-backed security (MBS) can pay a regular sum to a counterpart and, if the borrowers default, the counterpart will cover the losses. The credit risk is transferred to the counterpart – for a fee.

While default rates were low, offering this insurance was a great business. Investment banks and firms like AIG received the ‘insurance premiums’ but hardly ever had to pay anything. These departments made huge profits. In fact, the business was so profitable, CDS’s were sold to people even if they didn’t have an MBS. The investment banks and other sellers of CDS’s also underestimated the risk involved for the same reasons we have discussed.

The total size of the credit default swap market is uncertain as it is unregulated, but around the time of the financial crisis estimates put it about $30-50 trillion. The total annual GDP of the US is about $14 trillion. You can see how quickly the firms who had provided the CDS insurance would get into financial trouble when they had to start paying out on the defaulting mortgages they had insured.

Summary of part 1

In this post, I sketched out how the securitization of mortgages combined with credit default swaps and poor risk management were partly to blame for the financial crisis. Securitization meant that banks did not have to worry about credit risk and could (and did) lend to increasingly risky borrowers (the sub-prime credit market). Meanwhile, investors were comfortable buying these securities because they underestimated the systematic risk and / or insured themselves with credit default swaps. Prior to the financial crisis, investment banks and insurance firms made a huge amount of money by providing insurance (credit default swaps) against mortgage defaults and also underestimated the risk involved. Therefore, they took on excessive default risk. As local banks had been lending to more risky borrowers, a downturn in the economy and housing market caused mortgage defaults to increase sharply, leading many financial firms to (or close to) bankruptcy.

This is just one part of the story, but perhaps the most important. As you can see, it is really not that complicated. In future posts I will discuss some other aspects of the financial crisis, why they have not been addressed and what we should do about it.

In my post about the CCTV speech competition, I talked about how evaluation by penalizing mistakes will encourage conservatism, while evaluation by rewarding success will encourage risk taking. In my post on lying, I mentioned that lying is a gamble and people will be more willing to undertake the gamble when the potential losses are small and the potential gains are large. Both these ideas are specific instances of a more general idea that I want to talk about in this post: how the concavity (or convexity) of incentives can affect behavior. This idea is very simple yet gives great insight into many issues – including (as I show in this post) creativity.

We can consider the basic idea with the aid of some diagrams.

Figure 1: Concave and convex incentives

In figure 1, the red line is concave; the blue line is convex. For our purposes, we can say that with a concave function the slope decreases as we move to the right and with a convex function the slope increases as we move to the right. The horizontal axis measures performance; the vertical axis measures reward. Both of these labels might require liberal interpretation in some of the examples that follow, but they are the best I can think of right now. Basically, as we move to the right, the outcomes get ‘better’ in some way and the rewards increase.

We are going to be interested in comparing a safe choice with a risky choice. A safe choice guarantees reasonable performance. A risky choice may lead to exceptional performance if the gamble is successful, but may lead to poor performance if it is not. We will consider how the convexity / concavity of the curves affect the decision.

An example might be helpful here. Suppose a worker has a choice between using an existing, standard solution to a problem or trying a new, creative one. The standard solution is the safe choice; the creative one is risky. How will the concavity / convexity of different incentive schemes affect the creativity of the worker?

Figure 2: Concave incentives

Suppose the worker is considering switching from the safe option to the risky one. In the concave case, if his gamble fails, he loses a lot; if his gamble succeeds, he gains little. Unless the probability of success is very high, he will prefer the safe option.

Figure 3: Convex incentives

In the convex case, if his gamble fails, he loses a little; if his gamble succeeds, he gains lot. Unless the probability of success is very low, he will prefer the risky option.

This suggests that if we want someone to take risks, for example to be creative, we should give the convex incentives.

Convex incentives involve rewarding outstanding performance without punishing poor performance. Concave incentives involve the opposite. For example, if we pay people a flat wage with punishment (firing?) for poor performance you would expect workers to avoid risk. Among other things, they are unlikely to be creative. McKinsey was famous for their ‘up or out’ policy where a certain proportion of the worst performing consultants were fired. On its own, this would discourage risk-taking and creativity. On the other hand, they also had various rewards for outstanding performance which would encourage creativity. Tenure for professors reduces the cost to poor performance, so provided there is some (possibly psychological) reward for ground-breaking research, tenured professors have an incentive to take risks and be creative in their research.

It is now always desirable to encourage creativity and experimentation. For example, a factory worker may be part of a larger production system. The success of the system relies on him taking predictable actions and quality control requires his performance to meet some minimum standard. In this case, concave incentives (with punishment for poor performance) may be appropriate. Military organizations may be similar. Essentially when reliability, speed or cost are paramount considerations, concave incentives might be best. When experimentation and creativity is important, convex incentives may be the way to go.

The Chinese government would like to move the country from ‘making’ to ‘creating’. Instead of being a low-cost, low-margin manufacturer, it would like to develop high-technology, creative, high-margin businesses. One of the things that need to change is the incentive structures. Traditional Chinese organizations tend to be more likely to punish failure or deviance than to reward creativity and success. This is may work for a manufacturing company competing on speed, reliability and cost. It may be less suitable for more creative industries. I am not sure how easy this will be. The traditional approach is fairly well entrenched from schools onwards. A Chinese student’s best strategy for success will normally involve memorizing the teacher’s / textbook’s ideas with few mistakes, rather than thinking creatively. And of course creative thinking in certain areas may be frowned on by certain powers. There are other barriers to this transition that I will discuss in future posts.

There are many, many other applications of this basic idea – some of which I may discuss elsewhere. But let me justify the title of this blog by connecting the ideas to a couple of the existing posts.

As I mentioned in the CCTV speech competition post, I think the reward scheme for this competition tends to reward mistake-free, but fairly standard speeches, rather than surprising, quirky or innovative speeches. As this posts shows, penalizing mistakes rather than rewarding success will have this effect.

Lying is risky, so concave payoffs will tend to encourage truth-telling, while convex payoffs will encourage lying. In that post (and in my thesis), I showed how politicians who are relatively popular may be less likely to lie than less popular ones. This suggests that popular politicians face more concave incentives and this does seem to be the case. We can see that in the following diagrams.

Figure 4: Relationship between popularity and reelection probability

This diagram illustrates a simple assumption: the politician needs 50% or more fo the vote to be reelected. The analysis is not affected by more realistic assumptions.

Figure 5: Popular politician’s reelection function is concave

This diagram illustrates the reelection function for a politician with 65% popularity prior to the scandal. It is concave across the points relevant for his decision on whether to lie or tell the truth. As in the lying post, assume that the politician loses 10% support if he is known to have been involved in the scandal and 10% support if he is known to have lied about it (again the specific numbers are not important).

Lie & not caught → 65% popularity → reelected
Tell truth → 55% popularity → reelected
Lie & caught → 45% popularity → not reelected

Over these three points, the reelection function is concave and so he will avoid the risk of lying and tell the truth.

Figure 6: Unpopular politician’s reelection function is convex

The reelection function for a politician with 55% initial support is convex.

Lie & not caught → 55% popularity → reelected
Tell truth → 45% popularity → not reelected
Lie & caught → 35% popularity → not reelected

Over these three points, the reelection function is convex and so he will take the risk of lying.

There are many other applications of this basic idea. Any time a decision is taken involving uncertainty or risk, the convexity / concavity of the payoff function will matter. That includes many, perhaps most, decisions we make in life. Some of these applications I will discuss in other posts.

On Saturday, I was one of the questioners in the Shanghai heat of the CCTV English public speaking contest. Afterward, I was unexpectedly asked for my comments on the performances. I gave a few ideas, but as is often the case when I give unprepared comments, I subsequently thought of some other things I could have mentioned. In this post, I will give more considered thoughts.

Here is the format of the competition:

1. Prepared speech on the topic “Culture smart or science intelligent?”
   3 minutes
2. Impromptu speech on some unseen topic
   1 minute to prepare, 3 minutes to speak
3. Q & A based on impromptu speech
   2 minutes (not counting time taken by questioners)

First, I should say that the quality of speeches given in a foreign language was outstanding. I shudder to think what an equivalent foreign language speech competition in the UK or US would look like. The prepared speeches were very polished and I know that the competitors and coaches spent a lot of time and effort in preparation. Really it was an honor to be involved with this. None of what follows is intended to detract from this. I can’t provide much advice on scoring better in the competition – I am sure all the competitors and coaches are more expert than me on this topic – so here I will just comment on different ways of thinking about public speaking: competition versus ‘real life’, China versus the West.

As I have said, the prepared speeches were well-polished with good use of language. Most speakers also made an effort to generate emotion and emphasis, but sometimes it came across as a little artificial. In other words, it looked like a prepared performance rather than a natural response to the question. I feel that audiences tend to switch off if they feel they are just watching someone recite a memorized speech, but feel more engaged if they think they are hearing the person’s real ideas. Even within this competition, I felt that the audience paid more attention to the impromptu speeches than the prepared speeches – despite their lower quality according to the metrics of the competition. If people aren’t paying attention, then, no matter how great your speech is otherwise, it is useless.

Now, in the case of this competition contestants will have to prepare to do well and I am not suggesting otherwise. However, I think it is worth spending some time making the speech look natural even if it is prepared. Barack Obama might represent a good model of how to do this. I am not sure if this advice would be helpful for this competition, but it is useful for real life.

Related to this, some of the students were very nervous and it affected their ability to perform well. This is a common problem in public speaking and some other situations where a person has to perform under pressure. Why is it that somebody can talk in an engaging, humorous and intelligent way with their friends, but freeze on stage? Fear of failure. When a person is with her friends, she does not worry about making mistakes – if she stutters or forgets what she is saying, it is fine. When she speaks in public, however, she worries about what might go wrong and focuses on avoiding those mistakes. This has two effects. First, focusing on what might go wrong tends to cause her to make those same mistakes. Secondly, in trying to avoid those mistakes, she becomes very conservative in the way she talks. She avoids humor, for example, because it is too risky. As a result her speech becomes dull… and she still makes mistakes.

How can nervousness be overcome? I think the key is to be comfortable with failure. Michael Jordan, someone well known for performing well under pressure, said

I’ve missed more than 9000 shots in my career. I’ve lost almost 300 games. 26 times, I’ve been trusted to take the game winning shot and missed. I’ve failed over and over and over again in my life. And that is why I succeed.

and

I can accept failure, everyone fails at something. But I can’t accept not trying.

and

I never looked at the consequences of missing a big shot… when you think about the consequences you always think of a negative result.

and

I’ve never been afraid to fail.

All these quotes are about accepting failure. It is accepting failure that allows you to succeed.

Applying this to public speaking, if we worry about making a mistake (don’t forget what you want to say, don’t stutter, don’t mispronounce a word) rather than communicating with the audience (what do I want them to know, how can I make this interesting and enjoyable for them) it is unlikely we will give an excellent speech. We may or may not avoid failure, but we will never really succeed.

Following on from this point, I am not sure I agree with the evaluation methods used in this competition. I think they focus too much on penalizing mistakes and not enough on rewarding speakers who are interesting, entertaining and enlightening. I had a look at the guidelines for judges and actually you could interpret them either way, but I feel the way they are interpreted has the effect of encouraging formulaic, mistake-free but uninspiring speeches. For example, points are awarded for pronunciation which, beyond the comprehensibility, I do not think is important. Another example, points are awarded for emotion and emphasis (which are important) but good marks seemed to be awarded for speeches where every sentence was emphasized. This is unnatural and ineffective – if every sentence is emphasized, none are. If you bold every word in a paragraph, it is an unbolded word is emphasized. Change = emphasis. You can change anything – volume, tone, speed, body language, etc – but it is the change not the level that indicates emphasis.

One of the ideas I constantly force on my poor students is the idea of convexity and concavity in incentives. This topic requires several entire posts to cover, but let me outline how I think it affects this type of competition. If people are punished for pronunciation or grammar mistakes but not rewarded for inspirational speech, they will focus on avoiding the mistakes and not take the risks necessary to deliver truly entertaining speeches. The incentives encourage a conservative approach.

Finally, I should admit that what constitutes a good speech may be specific to a culture (or even an individual). I dislike most Chinese speeches I have seen. To me, they seem stilted, unnatural and often filled with empty praises or exhortations. I can’t help but prefer Barack Obama (or even an inferior Western speaker) to the typical Chinese political (or other) speech. I taught a class on communication and used Obama’s speeches as an example. I think my Chinese students enjoyed his speeches, but some felt that they seemed more entertaining than serious. I suppose I don’t see these as mutually exclusive. It is true that Obama’s speeches are mostly style, with very little substance, but that is true of almost all political speeches in any country. If the content is worthless, I may as well be entertained!

Can we predict when people are likely to lie? As an economist, my approach to this question is to think about incentives – people will lie when they have an incentive to do so. In this post, I extract one of the ideas from my PhD thesis which addresses this issue.

The risk of lying

We can think of lying as a gamble and telling the truth as a safer choice. This is the basis of lie detector tests that look to detect signs of stress from people taking the risk of lying. What exactly is being risked? When we lie, we risk our reputation for honesty (or what I will call credibility). This might be directly costly (we will be fired for being dishonest), indirectly costly (as in the story of the boy who cried wolf), or it may be a psychological cost (we don’t like to be seen as dishonest). For our purposes it doesn’t really matter as long as there is some cost to being thought of as dishonest. If we lie, we may gain something if we are not caught in the lie (otherwise why even consider lying?) but we lose credibility if we are caught. This is the gamble.

Lying and incentives

Who would be likely to lie? Someone who gains a lot if they are not caught but does not lose much if they are caught.

Example: Politician covering up a scandal

Suppose a politician has been involved in a scandal. Does he admit to it? Or does he cover it up and hope no one discovers the truth?

First, we need to know what the politician wants. I am going to assume he wants to maximize his chances of reelection.

Second, we need to know how the probability of reelection is affected in the following three cases:

1. Tell truth
2. Lie & caught
3. Lie & not caught

If these payoffs vary in some measurable way, we might be able to predict when a politician is more likely to lie and when he is more likely to tell the truth.

Suppose the politician loses 10% support if he has been involved in the scandal, and a further 10% support if he tried and failed to cover it up. If his current support is X%, his post scandal support will be

1. Tell truth = X-10
2. Lie & caught = X-20
3. Lie & not caught = X

How will the incentives to lie vary with differing levels of initial support?

X > 70%

Whatever the politician does, he will remain above 50% and get reelected. We cannot predict what he will do based on this model.

70% > X > 60%

If the politician tells the truth, he will remain above 50% for sure, but if he lies and is caught, he will slip below 50% and could lose the election. Therefore, we would expect him to tell the truth.

60% > X > 50%

If politician tells the truth, he will slip below 50% for sure, but if he covers up the scandal and is not caught, he would stay above 50%. Therefore, we would expect him to lie.

50% > X

Whatever the politician does, he will remain below 50% and is unlikely to get reelected. We can not predict what he will do.

Of course, the 10% numbers are arbitrary, but the logic of this argument suggests that their should be some region in which probability to come clean on a scandal increases with popularity.

In my PhD thesis, I tested this theory with data from the House Bank Scandal that occurred in the US House of Representatives in 1991/1992. This scandal was useful for statistical analysis because many politicians were involved and so statistical tests could be carried out. The idea presented in this post was supported to a statistically significant level. Interested readers can read a more rigorous presentation of the theory and the empirical testing in the first chapter of my thesis.

Conclusion

A person’s propensity to lie will depend on the relative size of the benefit if they are not caught and the cost if they are. In some cases these benefits and costs may vary systematically with observable variables and so we can predict that certain people will be more likely to lie than others. I applied this to politicians, but it could be applied to others. For example, certain incentive schemes for employees or executives may encourage lying and fraud. I may return to this in a later post.

What is intelligence?

I think of intelligence as a combination of memory and creativity. Memory is our store of ideas, experiences, etc; creativity is the ability to combine those ideas in new ways. If either of these are missing, then our ability to solve problems or react to new situations would be restricted.

Existing computers might have a huge store of “memories” but they can not use those memories creatively. They can solve certain sets of problems, but can not address others at all. Conversely, someone who has no memories, may still be able to solve problems creatively, but the range of problems they could solve would be limited because of their lack of knowledge – no language, maths skills, etc.

So to be able to solve a wide range of unanticipated problems requires both memory and creativity. This seems a fair definition of intelligence.

How connections determine meaning

We can think of the mind as a collection of ideas, experiences, emotions, memories, etc connected in a network. If when we do any thinking, we jump from idea to experience to an emotion, etc. Anyone who has tried meditation in which this jumping is to be stopped and focus maintained on one object or idea know how hard it is. Everything we think and feel happens in this network.

The network determines meaning for us. For example, suppose John has a pet Labrador as a pet. If he sees a Labrador, he is likely to feel happy. On the other hand, Jane has was bitten by a Labrador when a child. Upon seeing a Labrador, she may feel fear. This example is about emotion-based meaning, but we can also have idea-based meaning. For example, a farmer in a developing country may look at a horse as a tool of a trade. A book-maker may see a horse in terms of track record, probabilities, profit and loss. A zoologist may see a horse as a bio-mechanical system. An artist may see a horse as a representation of power and grace. The meaning of ‘horse’ depends on what other ideas it connects to for that individual.

What is memory?

Connections are memory. How do we remember something? In my view, memories are normally recalled as part of the jumping from idea to idea that I discussed earlier. In other words, we are more likely to remember something if we are currently thinking about a related idea or memory. I rarely think about anything from my school days as there is little in my current life connected to those days. However, sometimes I talk about our school days with my brother and one memory from that time triggers another in a chain reaction. Some times things that I may not have remembered on my own are triggered by some thing he says.

This can represent a problem in some cases. For example, suppose we learn something in school. We may be able to remember this thing in a school setting (e.g. an exam) but not normally recall it when in an outside setting. In face there is a lot of evidence that students abandon there ‘book learning’ when they leave the classroom and revert to ‘common sense’ rules of thumb. This suggests that effective education should look to establish links between the ideas taught and there applications in the real world.

Aside from context, ease of memory can be influenced by the number of connections to other ideas and the strength of those connections. For example, you are unlike to forget what your mother looks like because you probably have many memories of her in different situations and some of those will be emotionally powerful. On the other hand, remembering something you learned in a class may be hard because you only heard it once and it made no strong impression.

Again this has implications for how to make sure you can remember things. You need to ensure you connect the target idea to as many other ideas and memories as possible and to make those connections strong through emotional power, vividness or repetition.

What is creativity?

I see creativity as connecting existing ideas in new ways. For instance, taking an idea that would normally be seen in one context and placing it in another, we may create something new. We might think that Ang Lee was creative when he placed homosexuality in a cowboy movie. Neither homosexuality or cowboys were new topics for movies, but the combination was.

In this way, creativity makes new connections in our mind and thus changes what things mean for us. One way to think of the value of art, a creative exercise, is that it changes how we see the world.

Creativity is hard not so much because it is difficult to make new connections, but because it is easier to use existing old connections. In the preface to The General Theory of Employment Interest and Money, the economist John Maynard Keynes said

The difficulty lies, not in the new ideas, but in escaping from the old ones, which ramify, for those brought up as most of us have been, into every corner of our minds.

This implies that creativity maybe hindered by experience that entrenches existing ideas in our minds. The great creative achievements in science, art, philosophy, business or other fields are normally the work of people in their twenties or thirties. For example, Einstein was 26 when he published his ideas about special relativity and developed general relativity between the ages of 28 and 36 (source Wikipedia). Shakespeare appears to have written most of his plays between the ages of 26 and 40 (source Wikipedia). Bill Gates was 20 when he founded Microsoft, 25 when he created MS-DOS and 30 when Microsoft published the first version of Windows (source Wikipedia). Of course, there are exceptions. Immanuel Kant published the Critique of Pure Reason when he was 57 and other important works still later (source Wikipedia) – a counter-example I welcome as I approach the critical cutoff point! However, it does seem that 20-40 is the optimal age for major breakthroughs – old enough to have the necessary knowledge, but not old enough to be trapped by it.

Intelligence as intellectual networking

In this post, I have tried to show

1. Intelligence is a combination of memory and creativity
2. Memory is the existence of connections between ideas in our mind; and creativity is the creation of those connections

I want to use this blog as a kind of online idea network to create and remember connections between different ideas that I may have. Maybe it will help me become a little bit more intelligent!

Some of the ideas on memory were influences by the ideas of Vera Birkinbihl that I came across in an (English!) tape course by Paul Scheele.