The Market's Exit Interview

On most days, CEOs can probably pretend that what is happening to the stock is beyond their control.  But some days the stock is all about them.

Today, Microsoft's stock price is reacting to the news that Steve Ballmer is going to retire within the next 12 months.  The market capitalization of MSFT is currently $285 billion. Yesterday it was about $270 billion. Microsoft's jump in stock price means that people think whoever gets picked to replace him will make Microsoft approximately $15 billion more valuable than it would have been if Ballmer remained at Microsoft indefinitely.

It could be worse, Microsoft is only up 5.8% at the moment. When Carly Fiorina left HP the shares rose 7.5%. And the leaving executives will own large amounts of the stock that is appreciating in value in response to their leaving. So this hit to their pride is balanced by a boost to their pocketbook and this is before whatever ridiculous severance they've negotiated for themselves is also taken into account.

Edit: It appears Ballmer doesn't have an additional retirement package, so he's going to have to content himself with the over $700 million he made on his stock holdings from the market's reaction to his impending retirement. 

Disclosure: I talk about a lot of stocks. It should be assumed that I might own any mentioned in my blog posts either directly or indirectly and I'm not trying to get the reader to buy or sell anything since they should do their own research.

Regarding Useless Ladders

Tim Harford had an interesting op-ed last week and it's now outside of the FT's pay-wall.  In the op-ed he looks at the reasons behind the increase in inequality.  It covers a lot of the same ground as Mankiw's paper, Defending the One Percent.  After reviewing the reasons why globalization and technology have caused winner take all markets to become the norm in many areas, the analysis turns to the increasing correlation of intergenerational wealth in America and the UK and how this amounts to the rich pulling the ladder up behind them.

He focuses mostly on the higher cost of status markers.  
"The well-off feel that they must strain to prevent their children from slipping down the income ladder. The poor see the best schools, colleges, even art clubs and ballet classes, disappearing behind a wall of fees or unaffordable housing."
In a wealthy society, the market for status does not get more efficient - more money is chasing the same number of spots and the signal of money being spent is often mistaken for quality. Take these competitions to their logical extreme and we end up with 400 dollar sessions designed to help 4 and 5 year olds get into the top preschools*. The solution to inequality is not to make it easier to attain the same status markers. Those status markers are functionally useless ladders. It has been attempted with education, and throwing more money into student aid and loans has helped drive education costs higher, saddling students with debt that is far in excess of what they can handle given their economic opportunities. 
No matter what you do, there are still a limited amount of spots at the top. Subsidizing university education for everyone just makes going to a top university or getting additional degrees in relevant fields more important. If everyone takes ballet classes, then ballet classes won't matter as a status marker anymore. The total costs to signaling go up, but the winners will be by and large the same people. The op-ed goes on to note how high inequality countries are self perpetuating.
The painful truth is that in the most unequal developed nations – the UK and the US – the intergenerational transmission of income is stronger. In more equal societies such as Denmark, the tendency of privilege to breed privilege is much lower.
The comparison of Denmark's lower intergenerational transmission of income to the US completely misses Denmark's significantly lower variation from both a genetic and cultural standpoint. 
The population of Denmark is so much more homogenous than that of the United States that bringing them into the discussion is basically a non-sequitur.  Not only is the US significantly more diverse, the United States has also been segregating itself along educational lines. In the US, universities have increasingly acted as genetic and cultural sorting mechanisms. In 2011, 71 percent of college graduates married other college graduates and it should be obvious that children of these couples face many advantages. The school districts they attend are better because of their involvement. Their children go on to outperform children from other backgrounds, reducing intergenerational wealth mobility. 
Expensive private schools and extracurricular activities do make life very difficult for the intelligentsia who are members of the upper middle class but who lack the income to support that membership for their children. But they aren't the heart of the problem - the concern is about creating opportunities for the least well off rather than making sure those who are well off are able to see their children fail. The heart of the problem lies in cultures ravaged by the war on drugs and in the broken school systems the war on drugs has helped create. The obvious solution is to stop the crazy war on drugs.  Otherwise, getting rid of crazy occupational licensing laws and other rent seeking structures in society, creating frameworks for apprenticeships and finding ways to match students with good teachers who have time to teach without the presence of disruptive students are all potential paths towards a much better future. These issues and many others come well before those looking at the big picture should start even worrying about the rising cost of ballet lessons and art classes.
*Unfortunately, this isn't really the logical extreme. Things will get much crazier.

Companies have been underpaying for bugs for a while

This Forbes article on the Zero-Day bug market is quite interesting. It highlights that there are people (mostly consisting of government intelligence organizations) willing to pay hundreds of thousands of dollars for specific exploits. Exploits in Adobe reader sell for between 5k and 30k, while IOS exploits sell for between 100k and 250k. The buyers? Western spy agencies. What about the companies themselves?

 Google typically offers a maximum of $3,133.70 for information about the most complex flaws in its software, a number that’s meant to spell out “elite” in hacker slang.

Perhaps there are good reasons for companies to refuse to participate in a market that revolves around exploiting their mistakes. And maybe the brokers who deal in these exploits aren't actually allowed to sell the exploits to them, for fear of being accused of blackmail.  Or maybe companies don't want to have to pay market rates for a problem they have identified and are resolving.

Still, there are ways around all of these problems. In the end this seems to be a subject that most companies are too prideful to treat properly. Something that used to be offered for free or quite cheap is getting expensive and companies are caught flatfooted. But some companies are finally starting to accept reality, according to a recent New York Times study of zero day exploits Microsoft recently raised its top bounty to $150,00. Facebook, when it is willing to make payouts, has made payments as high as $20,000.

Suppliers who used to provide these services for free are starting to realize they should be paid at the same time more buyers are entering the marketplace. Combine this with high profile bounty rejections and more hackers will figure out that there are usually other ways to monetize their findings. Given the supply and demand dynamics in this market it is unlikely that prices are going anywhere but up.

Facebook follows Paypal in trying to turn white hats black

A lot of companies have bounty programs. Sometimes they try to find silly reasons for refusing to pay out. Paypal didn't want to pay a programmer under 18 years old who found a bug. This is a bad idea because it encourages those who have found bugs to figure out other ways to make money with those bugs. Today, it's Facebook that has decided they also want to create perverse incentives for hackers.  Their security team ignored a bug report until the researcher used the bug to post on Zuckerberg's wall.

Joshua also informed Shreateh that he would not be receiving a bug reward for reporting the exploit because he violated the site's terms of service. "We do hope, however, that you continue to work with us to find vulnerabilities in the site," he wrote.

There are two interpretations of what is going on here. The first is that Facebook does not want to encourage people to embarrass the company or harass their CEO/Founder.  But adding to this is that the Facebook employees who are in charge of this program that decided to not listen to the researcher decided to not pay out the $500 due to their general embarrassment.

Either way, it should be noted that exploits that allow for spamming Facebook walls are worth significantly more than $500.  In terms of cost, $500 is a very small amount for Facebook and they are used to spending much more in order to keep employee morale high. By not even paying these relatively small payouts, Facebook is encouraging people who can find exploits to find other ways to monetize their findings. They are also decreasing the morale of their large free workforce. If Facebook is going to refuse to honor their pay outs unless people follow the rules exactly they should make the payout significantly larger.

For the individual involved in each case the publicity the media gives them is worth more than the payout they missed out on, but the opposite is the case for the companies involved.

The Politically Connected, Banks and the FCPA

I've already talked a little about the inherent disadvantages that US and other developed market economy based companies face due to the Foreign Corrupt Practices Act (FCPA). Companies that can't or won't bribe officials in corrupt economies will be at a large disadvantage compared to companies who will make these bribes. In order to win business they will have to be orders of magnitudes more efficient than their local competitors - which is an advantage that is hard to sustain for long periods of time.

I thought companies had an easy way around it - hire locally connected employees. The relationships these employees have are likely to be based on favors rather than explicit bribes. Hiring these connected people will open doors and prevent the developed market based companies from being shut out of deals or from being taken advantage of. A new story concerning JP Morgan in China highlights how this strategy is becoming a risk for the banks.

...with the U.S. Securities and Exchange Commission (SEC) investigating whether JPMorgan's Hong Kong office hired the children of China's state-owned company executives with the express purpose of winning underwriting business and other contracts...

The facts of the case aren't out yet. It might be the case that people within the relevant JP Morgan departments weren't careful enough to keep incriminating facts out of email and an angry employee became a whistleblower. But if the SEC is merely applying the same criteria to enforcing the FCPA as they do to insider trading (if the correlations look like insider trading they will go after you) companies will have a more difficult time working in emerging markets.

If developed market companies can expect a harder time competing in developing markets this is good news in the long run for developing market equities. Considering the underperformance of EM equities this year they need all of the help they can get.

Board games and the "Luck Time" Measure

So let's assume a new entity, gamer-economus. Gamer-economus plays board games hoping to defeat their opponents by demonstrating their skill differentials across a variety of games. They prefer to minimize the time spent on games where achieving victory is beyond their control.  Luck Time is a new metric designed for the gamer-economus in many of us.

Luck time = L*T

L = Luck component. If a game is 50% luck, the completely dominant player will win 75% of the time. This is a number generated by a lot of hand waving, because as we know from chess even high skill games have lower skilled players being higher skilled players occasionally.

T = Average time to completion.

Combined, the Luck Time measure gives us a sense of how much time Gamer-economus "wasted" on a game that was in the end decided by a dice roll.


In "flip the fair coin" - the game is dominated by luck. L=1 It's also a quick game, so T approaches zero.

In chess or go - the game is dominated by skill. Lack of skill may cause unpredicted lucky situations to arise, but a computer with enough time foresee these situations.  The better player may not win every game if the skill gap is close enough of if the losing player learns from their mistakes, but any time the game is played skill generally determines the outcome. L approaches 0. T varies.

In the above games, the LT is close to zero. In "flip a coin" it is because the time played is close to zero (a match of "flip a coin" to 100 would be pure luck time), while in chess or go LT approaches zero because L is close to zero.

Applying this analytical method to games in-between is where it becomes interesting. If backgammon and Settlers of Catan both have similar amounts of lucky events that even out over the game to the same extent, then the two games can still be quite different in the luck time measure.  This is because the time variable between the two games is vastly different. Settlers can last over an hour, while one game of backgammon is often over in 10 minutes.  Given these assumptions, settlers has six times as much Luck Time as backgammon. 

If a single game of backgammon was expanded into a match, the time would go up while the luck would go down significantly. This is because there are additional mechanics in the match, such as "doubling" that make the longer match more skill intensive. The LT might actually decrease and the average luck time per point of the match definitely decreases. 

But despite their similar reliance on dice rolling, backgammon and Settlers have very different mechanics. Luck Time is more relevant in games where the core mechanic is similar. A shorter amount of Luck Time is a core reason for why many gamers prefer The Resistance over Battlestar Galactica. They are both "spot the traitors" games, but The Resistance finishes each game in between ten and thirty minutes, while Battlestar can take multiple hours. There are reasons to play Battlestar, but it would be aimed at the gamer who enjoys themes and different mechanics for their own sake. 

Some games take a while to learn but once the optimal strategy is learned there isn't much variation in how it is applied. The first games played would have low levels of Luck Time, but once everyone figures out the game's strategies the LT would be much higher if there aren't enough subtleties in the game or if the subtleties have a small impact relative to the luck of the game.  This is a problem with many board games in general, and it is one of the reasons board gamers switch between many different games instead of repeatedly playing just one or two. 

What about Kingmaking?

One slight modification to Luck Time would be Luck Time without Kingmaking and Luck Time plus Kingmaking (Total Wasted Time). Kingmaking occurs when a player who is not going to win gets to use the mechanics of the game to essentially pick the player who ends up winning. Sometimes its occurrence emerges as part of the strategy (the player who was mean earlier will be punished), but sometimes it is pure social dynamics unrelated to the game itself. These social dynamics are part of why Settlers of Catan is a decent tournament game but a subpar general social game.

Luck Time = L*T
Total Wasted Time = L*T + Km*T

Km = Percent of time the winner is picked capriciously by another player

Munchkins is an example of a game that has both a low skill ceiling as well as significant king making. Once everyone knows how to play, the question becomes if players will let someone get to level 10 before them. It has high Luck Time, but even higher Total Wasted Time - especially when particularly aggressive play at lower levels can make the game drag on for hours.

When thinking about games, there are many factors to consider. The style of the game, how fun the mechanics are, the level of the skill ceiling, how fun the theme is, how easy the game is to learn and how long the game takes should all be taken into account. But even if gamers don't think of themselves as a gamer-economus, the potential for high amounts of Luck Time and Total Wasted Time is a metric that ought to be explicitly considered.

One way to frame a debate

Scott Sumner makes an interesting point.

Taken from what the Keynesian commentators are saying, if the US economy continues on its current course of a slow recovery it means that extreme government budget deficits during downturns are no problem and people were stressing out too much over them. If the economy stalls it means austerity really was bad and quite dangerous.
People like it when their pundits are confident, not accurate. So this little trick is probably more for the pundits themselves, since it is easier to be confident if accuracy is assured no matter what the result.

The New and Improved Experience Machines

 "Suppose there were an experience machine that would give you any experience you desired. Superduper neuropsychologists could stimulate your brain so that you would think and feel you were writing a great novel, or making a friend, or reading an interesting book. All the time you would be floating in a tank, with electrodes attached to your brain. Should you plug into this machine for life, preprogramming your life's experiences?"

 -The Experience Machine, Robert Nozick

Nozick uses the hypothetical scenario as an argument against the idea of pleasure as the only intrinsic value. And he points out that the conflict over psychoactive drugs is partially because those opposed to drugs see them as local experience machines, as they believe the drugs disconnect the user from the world to give them pleasure. (And the drug users might defend themselves by suggesting they are seeing a deeper reality.) 

What Nozick found most disturbing about his experience machine was that it would live life for its user - things would occur in a preprogrammed fashion. Modern industry has come up with an alternative solution: Video games.

There are many genres of video games. MMO's, or massively multiplayer online games currently seem to be the top contender for potential experience machines.  The player might control an avatar in a medieval Tolkien-like setting or choose a game where their avatar exists in a world of warring space empires. In these games they get to interact with others and if they avoid large mistakes they will be rewarded within the game for the time they've spent playing. These games are quite immersive and many people have to get help to stop playing them.

Like all thought experiments, this one only works if the reader agrees with the author. Many people might choose to hook themselves up to an experience machine. In high school, the majority of my peer group reacted positively to Huxley's dystopian vision in a Brave New World. The same mindset could easily lead to the complete embrace of more and more immersive video games.

The immersiveness of video games is being taken further by companies such as Oculus Rift, who are creating virtual reality headsets that apparently work quite well. This video of people’s reactions to a guillotine simulator displays how powerful virtual reality can be. Other developers, such as Riot, are analyzing the behavior of their millions of players in order to figure out how to make their games even more addictive (and how to generate more micropayments from the addicted players).

Meanwhile, video games are rising in status and visibility.  In the competitive game genre, the third International Dota 2 tournament recently finished. The prize pool was over $2.8 million dollars. This is still a long way away from the purse sizes of other competitions, but top professional gamers are making more money than biathlon competitors and it is likely that they will pass chess players soon. Their respectability is growing along with the prize pool.

Today's children are growing up in a world where computers are a normal part of their lives – most two year olds will know how to play with an Ipad. They will have never known a world where there isn’t an alternate digital escape, and that escape will seem normal.  It might be possible to argue that today’s children have it much worse than children in the past. Their time is often excessively managed by their helicopter parents, who are worried that without many scheduled extracurricular activities the children will fall off of the path to the Ivy League and success. Society has also become overprotective of children – many schools are so concerned with childhood safety that they ban games like tag and parents don’t let kids wander off alone because they’ve watched too many episodes of Law & Order.  At the same time, the people designing alternate worlds for these children have gotten very good at generating simulations which create dopamine rewards for prolonged interactions with the game. Digital escape gets more appealing for each generation.

In Japan there is already a term for those who have given up interacting with society and who do not leave their rooms for months on end, hikikomori. It's not always related to video games, cultural expectations, low opportunities for employment and psychological issues play a large role. But the availability of immersive video games and the Internet certainly increases the opportunity cost of interacting with the real world.  

It will be interesting to see how society reacts to the increased immersion and addictiveness of video games in the coming years.  Right now games are seen as scary by the professional worriers who worry that gamers will act out the more violent scenarios from games. Merely making a bad joke about going crazy can put a kid in jail.  But the real risk is not that gamers will take the game into the real world – it’s that many of them will decide that they don’t like being in the real world at all.  

It’s interesting to compare television and novels to video games. Both can be immersive technologies, but the immersion is much higher for games.  To show this, try to talk to a person playing in the middle of reading a book, watching TV, or playing a video game. The person reading a book can put the book down without too much trouble. The person watching TV is a little farther away. The person playing the game might not even realize you are talking to them. Another difference between television watching and gaming is that games require mental exertion. If a gamer spends 34 hours a week playing games, equivalent to the time the average American spends watching TV, they may be too tired to do other intellectual activities. 

The question of whether or not gaming or television is more harmful has yet to be settled. Gaming provides more intellectual stimulation - and yet outside of reality television watchers, dedicated gamers are widely acknowledged to be lower status than television watchers. This probably has more to do with the type of people who seek out the experience machine being low status in society to begin than it does with social ostracization being used as a tool to prevent people from choosing to disconnect from reality. The latter reason is likely to become more important as the experience machines become more appealing and more addictive.

And yet thanks to a generation of gamers growing up, getting jobs and making money, gaming is enjoying more social acceptance than ever. Perhaps the millennial generation is proving that an interest and hobby of playing video games does not preclude most of them from interacting with the world. But video games are getting better, much better. We can't necessarily say the same thing for the opportunities the children growing up today. Both youth unemployment and student debt levels are near all time highs. The one positive that really stands out is that they have option of plugging themselves into new and better designed experience machines. 

I’ll be going to PAX at the end of this month… to investigate these trends more closely.

Survivorship Bias Applied to Movies

The NYT's Catherine Rampell presents an interesting graphic that compares movies by their return of their production budget.

The problem is that this chart has massive survivorship bias. These numbers were taken from a data set:

"...this data set, which includes movies made from 2003 to 2012 with at least $2 million in domestic box office returns."

Eric Falkenstein's The Missing Risk Premium references a study by Art Devany which shows us that while R-rated movies have high average gross returns but their risk is also much higher. Below is a chart of movie gross return and volatility from 1985 through 1996.

Chart from DeVany and Walls (1999)

Given this information, the original chart is about as accurate as the following study: "I'm going to look at the return of micro cap stocks compared to other stocks, but I'm only looking at the ones that didn't go bankrupt."  Even though we are excluding all bankrupt stocks it's a forgone conclusion that the returns of the micro cap stocks will be significantly higher than any other type of stock. 

The original movie chart isn't devoid of information - it does suggest that globalization is friendlier to NC-17 movies than movies of other ratings.  We've known for a while that the consumption of blue movies is popular in the developing world and this is more evidence.

Accurate Chinese Data

Getting accurate data on events within China can be quite frustrating. Thus, it was somewhat amusing to me to see how the Dui Hua Foundation, cited in an article by the Economist, gathered data on the number of executions in China. 

In 2012 a deputy minister of health cited the decline in executed prisoners as a reason for a shortage in organs available for transplant in China.
This highlights both the disturbing fact that China's government is still killing a lot of people (And there may arise cases where a prisoner is killed more because they are compatible than because they guilty) as well as how hard it is to get accurate statistics on events occurring within China.

The lack of good data from China is highly relevant right now because they are attempting to cool down their overheating shadow banking credit markets prior to financial reform. The main signs of what is happening is coming from their interbank market, where interest rates are spiking.
Tightening financial conditions well into a boom without causing a crash is difficult, and without data it might be hard for foreign observers to tell if things are going wrong until it is too late to react.