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.

Addressing Prosecutorial Overreach and Intimidation

In the US, the justice system relies on plea bargains and summary judgements to keep the gears of justice working. If every arrest required a jury trial the costs would be significantly higher.

From the prospective of the suspect, whether or not to suspect a plea deal depends on the following equation that compares the cost of both options:
P = Probability of conviction
S(trial) = Trial sentence
S(plea) = Plea sentence
V = Psychic value of being found innocent rather than guilty
C = Cost of mounting defense

1. P*(S(trial)) + C - (1-P)*V > S(plea)

2. P*(S(trial)) + C - (1-P)*V < S(plea)
In scenario 1, it makes sense to take a plea deal. In scenario 2, it makes sense to fight the plea deal. In theory, the probability of conviction should vary with the defendant's knowledge of whether or not they actually committed the crime and many unnecessary trials can be avoided.

The issue comes because prosecutors realize early in their careers that they can make their time much easier by massively inflating the sentencing that defendants would face if the case went to a trial. Accepting a 2 year prison term can make more sense when the other option is a chance at being sentenced 30 years to life if the case goes to trial. This can be the case even if the defendant thinks they are actually innocent. This was famously done in the case of Aaron Schwartz (Who was actually guilty, but of a relatively minor crime with a poorly written statute that allowed for multiple felony charges), in which the stress induced from this overreach helped drive a young programmer and activist to suicide.

As Orin Kerr notes:
"Yes, the prosecutors insisted on jail time and a felony conviction as part of a plea. But it is not particularly surprising for federal prosecutors to use those tactics. What’s unusual about the Swartz case is that it involved a highly charismatic defendant with very powerful friends in a position to object to these common practices."
The ability of prosecutors to bully defendants into plea deals doesn't mean the whole system needs to be revamped, but ways to reduce this imbalance of power should be addressed.

Some potential methods to prevent this:

1. Reduce maximum sentences for various crimes. This would be a very labor intensive process and would be difficult for politicians to back more reasonable sentencing guidelines without getting accused of being soft on crime. But there is a potential for the political environment to turn against the government in security matters. If proponents properly the frame the issue as one of preventing government misconduct there is a slim hope that this can be addressed.

2. Rather than fix individual measures, adopt a broad method such as forcing concurrent sentencing to be adopted over sequential sentencing for crimes of specific type would mean that prosecutors wouldn't be able to threaten quite as much prison time in order to force innocents to plead guilty.

California has a rule like this for single offenses where a single act or omission can only be punished in one manner.  Broadening this rule to include many individually committed crimes would prevent prosecutors from being able to throw the book at suspects. The caution would be in preventing scenarios where a criminal can commit one significant crime and then feels safe the they can commit any number of less significant crimes without many consequences. 

3. Have plea deals cap the maximum potential time served. If a plea deal is offered by the prosecution, the maximum prison time sought by the state (or equivalent punishment) can only be 5 times the time of the offenses mentioned in the plea deal offered. If significant cooperation in other investigations is involved as a condition this cooperation can be made equivalent to X years served for the purpose of the calculation.

This way, people like Michael Milken might not have been incentivized to agree to the charges which led to him serving 22 months in prison because the prosecutor was trying to use RICO to put him in prison for life*. Once the 22 months are on the table the size of the stick the prosecutor is using changes from life to under 10 years.  Because judges determine the sentencing they would have to be involved in signing off on plea agreements with pre-determined sentences, which might cause some additional complications. Maybe this should be reserved for cases defined as "high profile" when the prosecutor is also facing political pressure for a successful resolution.

There are probably other ways to make prosecutors more accountable. The path of a high profile victory against a public enemy towards a career in politics is an unfortunate one. But it isn't just the high profile cases that matter, it's the potential abuse of anyone who is being charged with trumped up charges in order to make them more pliable to a plea deal. Right now the United States has a system where if there is a target on your back, there will be laws that you violated and the prosecutor can add them up until you agree to a plea deal. 

It's a less than optimal system that badly needs to be addressed.


*Guiliani was the overreaching prosecutor in the Michael Milken case. Milken only agreed to the charges which he thought would not lead to prison time after prosecutors threatened to go after his brother.

Summers and the Fed

Bernanke's replacement seems to be the media story of the hour, even though this story probably won't be resolved until the fall. It is a strange type of discussion where people on opposite sides of the political spectrum like John Taylor and Diane Feinstein are on the same side of the debate - both prefer Janet Yellen.

At this point it looks like the media and pundit backlash makes the idea of a Larry Summers Fed chairmanship somewhat remote, but it would lead to a potentially interesting situation. The Fed chairman doesn't have dictatorial powers - if the FOMC board chooses to vote against the chairman the majority has power to set policy. While this is unheard of in our lifetimes in the US and in countries like Japan, it happens all the time in the UK.  Mervyn King, the former Governor of the Bank of England, was outvoted many times.

Given the well documented ability of Larry Summers to anger people around him, his chairmanship could lead to a situation where the chairman is not the person who decides where monetary policy is going next.  The market already has a hard enough time figuring out what the Fed is going to do when one person is nominally in charge. It would be interesting to see how inefficient the market gets if Summers takes over yet Federal Reserve officials don't follow his lead as closely as he might like/expect.

Given my occupation, that's a good reason for me to hope that Summers becomes the next Fed chair.

Assorted Links

1. The UK is thinking of letting a company with links to China's government censor their internet exposure. This is another example of the basic rule of thumb that when people are told to think of the children the policies proposed are going to be pretty asinine.

2. Maybe someone is jealous of all of the attention Carlos Danger has been getting. Side note: I do hope that Anthony Weiner stays in the race, politics is a lot funnier with him around.

3. Welfare has spillover effects onto future generations.

4. An interesting theory on why women like conspicuous and relatively ugly luxury handbags. Key quote: "When women felt jealous, they drew designer logos that were twice the size of those in the other conditions."

Psychic Hedging in UK Betting Markets

In the UK it is legal to bet on many things. And with the royals expecting a baby, many people were betting.  This isn't a market I would want to be a part of - it seems almost obvious that some people will know the actual information and betting against people who hold the winning hand isn't usually a profitable activity. From the WSJ, we learn the odds were heavily in favor of the baby being a girl, despite 1.07 boys being born to every girl.

Bookmakers are offering bettors a chance to predict the child's gender. And a girl is the overwhelming favorite, with payouts of just £4 on a £7 bet this past week. A spokesman for the couple announced last month that the duke and duchess don't know the gender.

The betting markets turned out to be quite wrong - it was a boy. So what happened? For one thing, it seems like there wasn't a sizeable information leakage (Or maybe there was, but people didn't believe Harry enough to bet on his information). If someone made money betting on the sex of the baby, they made it without impacting the wider market. So the betting market was wrong for one of two reasons - psychic hedging and/or misinformation.

Psychic hedging is betting on an outcome that will make the person unhappy. So if a sports fan bets against their favorite team, either their favorite team will have won or they'll have money. In either situation they'll be happy. In this case, there is presumably a preference for a new king. Not only are male leaders more generally preferred, the birth of a girl would involve a messy procedure of changing the law to allow the girl to retain her place in line.  This preference for a psychic hedge would be combined with misinformation and misinterpreted market signals - betting on the expensive side of this type of trade is the easiest way to presumably follow other traders who might have inside information - to skew the odds even more in favor of a girl than could be explained by psychic hedging alone.

If insiders did make money off of these betting markets they got away with it quite handily this time.

Evil CEOs vs Bad Incentive Systems

A lot of people seem to think that most CEOs are actively evil and that this explanation can be used to explain many of the problems in our economy.  A recent example of this fallacy comes from a ritholtz.com post looking into the way appraisers that didn't inflate the value of houses were seemingly systematically blacklisted. 

"The appraisers’ petition was done over the course of seven years.  Even if we assumed, contrary to fact, that the CEO did not originate the plan to inflate the appraisals the CEOs knew that they were making enormous numbers of fraudulent “liar’s” loans with fraudulent appraisals.  It is easy for a CEO to stop pervasive fraudulent lending and appraisals.  Where appraisal fraud was common it was done with the CEO’s support."
I've met a few CEOs, and while they may care much more about the bottom line than the average person they aren't looking to break the law in order to gain a small advantage. The risk usually isn't worth the reward.

In most of these cases there is a much simpler explanation: The CEOs presided over an incentive system that rewarded misbehavior on a small scale without doing anything to directly endorse the behavior.  In the case mentioned above, those who originate the most loans will be bringing in the most revenue. There will be no comeuppance to catch the overconfident employees who bend the rules because an overly optimistic appraisal will be justified by rising prices a year from now. The more cautious employees will stagnate compared with those who are willing to bend the rules.   And so the people who get promoted during a boom will be the revenue generators (the rule benders) and the practice will become more widespread as time goes on.

All that this requires is that the CEOs set up a system to encourage and promote the people who are generating business. They don't have to have anything untoward on their mind. 

Grand conspiracies are fun to talk about. It would be nice to have a few specific people to blame for the problems in our world because that implies we can solve those problems by targeting the people at fault. But the truth of the matter is that the incentive system set up by the people running the show is at least as important as the moral character of the people in charge and probably more important.

Assorted Links

1. This looks like an effective way to do cardio.

2. What we know about heart disease is probably wrong.  Note: While inflammation is likely to be more significant than current treatments emphasize, there are people skeptical of this doctor. 

3. This was a sad day for the UC system - this latest development suggests that no major problems will get fixed. 

4. Most people look at experimental economics as undermining market fundamentalists because of the rational expectation roots of the Chicago School. Here is a take which suggests that when rational expectations is taken too far it eventually undermines market fundamentalism. 


This Era's Pet's.com

My friend Ben posed an interesting question - what will be the Pet's.com investments of this era?  He made the point that it isn't obvious to everyone at the time that these companies are ignoring economic realities - many people were convinced to invest in the company and many others decided to work for it before it was exposed as hopeless. 

Some people already had an answer to that: Groupon. It's possible that Groupon will be representative of the issues with start ups in this era, but Groupon's rise and fall didn't occur at the tail end of a bubble. Consumer centric tech start ups have gotten less popular, but there are still many other things to consider.
  • Today's era "bubble" investing is occurring largely away from the public markets. For better or worse, normal investors aren't getting involved in tech start ups until much later in the game. This means there are less high profile cases where individuals lose money, which is why people were so surprised by Facebook's public market performance.
  • VCs are relatively secretive about their return on individual investments, especially failed investments. Many failed companies turn in acqui-hires for undisclosed amounts. When these companies shut down or change directions without an exit their profiles are either scrubbed of investors or are never completed in the first place to avoid damaging the reputation of their investors (Which is silly since all VCs invest in companies that do badly). This means there are less high profile failures.
  • The old economy still has lots of inefficiencies, so there is still lots of room for smart programmers to team up up with people who know industries to fix inefficiencies. This trend can be summarize by anything from smart enterprise where new start ups are going to replace old inefficient software to software eating the world. Either way, it means that there will be many more success stories for every failure. And with all of the successful companies it seems silly to focus on the failures. No one is making fun of Bebo these days, unless it is to point out the elitism and hypocrisy of the founders' new clubhouse.
But even given that, there are obviously a lot of companies with overly optimistic business models and it might be easy to look back and point out some obvious misconceptions. Here are my candidates:

1. Generally ignoring that some users are more valuable than other users is insane.  Looking to 10th graders for investment advice should seem a lot sillier in 10 years later. Some users are much less likely to turn into cashflow down the road and companies that are engaging just those users shouldn't be valued at a premium.

2. It's crazy to assume that user networks are going to remain stable over a 10 to 20 year timeframe. Facebook's valuation is dependent on it not being Myspace.  They bought themselves some time by buying Instagram, but they are going to have to do a lot of really useful things to justify their $62 billion dollar valuation. (The valuation makes more sense if it is seen as an option).  In ten or twenty years, people might decide that the right way to value a social network is more similar to analyzing an active oil well, which we all know will predictably run out of oil (in the case of social networks, eyeballs and engagement) while there are numerous ways to maximize its yield in the meantime.

3. Counting on merely acquiring active users and then being acquired in turn. It worked for Tumblr, but eventually the companies acquiring these users might start to figure out that it isn't worth it.  It's important to remember that during the last bubble there were many companies that were basically worthless that ended up getting acquired before the bust - the acquisition of useless companies even made a few people into billionaires. These companies are successes for those who were able to build and sell them, but in 10 years the idea that a blogging platform was worth a billion dollars should look ridiculous.

4. The idea that users acquired for one reason can be transitioned to other purposes. Digg found out this doesn't work the hard way when it tried to turn its news aggregator into a social network. Yahoo might soon find this out with Tumblr. Just because you have users on your platform interacting with each other doesn't mean that they will want to interact in ways that could be profitable for the platform owner. 

5. That government regulation can be ignored and a business can be built before addressing the fact that the business is illegal. Hopefully these companies still succeed, but companies like Uber and Airbnb and newer companies like Sidecar were built by ignoring current regulations and hoping that by the time they were really challenged they'd have enough public good will to change laws in their favor.  In the future, incumbents might be quicker to challenge fledging start ups in their sector that are technically illegal.

6. Dependable low cost part time labor isn't ubiquitous in a healthy economy. Many companies, including the Webvan type companies of this era, are dependent on low cost part time laborers to make their system work. Without these available workers, the companies would be unable to source workers to deliver food/wash a car/clean a house/drive a car at a cost that makes sense for both the customer and the worker when the company itself is taking a cut of the earnings and paying for insurance.  Whether this turns out to be a misconception has a lot to do with the future economic growth as well as immigration and health insurance laws.  

7. Bitcoin. This topic deserves a separate post, but generally companies that invest in bitcoin start ups when the start ups' success is generally dependent on the price of bitcoins going up just seems crazy - especially compared to the strategy of just buying bitcoins. 

8. Sharing less durable goods (most products outside of housing) drastically increases the wear and tear because a significant minority of participants will treat the shared objects they are renting like they don't own them - because they don't.  This is more likely to emerge as a problem after more sharing based companies attain scale, because as of now many participants in the sharing economy feel more affinity for the people they are sharing with/from because they are all early adopters. 

With the exception of number 1, it would be good if most of these ideas don't turn out to the misconceptions of this era. Number 6 seems the least likely to be a misconception - if anything the gig economy seems likely to be a growing trend enabled by a growing number of market places for short term work.

As Ben pointed out in his post, the idea that companies like Webvan and its ilk were extremely valuable in 1999 wasn't silly to a large group of investors.  And some potential misconceptions would have been absolutely wrong in the past decade - Youtube, which was losing money on each user, is now quite profitable for Google. Facebook, whether or not its current valuation is justified, is bringing in enough significant revenue from its user base that it makes my 2006 discussion with an early investor on how he could convince Mark to just sell to Yahoo for $1 billion because $300 million is a lot of money seem very ridiculous in hindsight. 

And there is still a lot of innovation in the personal technology space that is just waiting for a company to come along and execute efficiently.  Wearable computing means forgetting people's names will be easier to avoid in the future. And the time period when you don't need a wallet, just a phone, and you'll be able walk in and out of a restaurant without dealing with a bill the same way this can be done for cabs today shouldn't be more than 5 or 10 years away. Progress isn't going to stop, but some of our current ideas about what that progress will be are likely to be very wrong right now.

What other ideas about start ups of this era will be seen as crazy 10 years from now?

A few more links

1. Robin Hanson examines why betting isn't more common.  All of these issues he brings up seem relevant, but the introduction of money into the equation seeming rude and highlighting power differences outside the conversation is probably the most important.  Maybe that's why my friends won't make many bets with me... or maybe that's why I don't have more friends.

2. This ghost city, an attempt to replicate Manhattan, in China will be relatively easy to find, or at least recognize.  The story is funny because it combines many of the common worries about China: It covers their over investment in real estate, willingness to directly copy the west and the increasing problem of bad debt.

3. An interesting perspective on the pre-access to consumer confidence survey data provided by Reuters.  In the end, what Reuters is doing is just selling private research. If providing early access to market moving research is disallowed completely that will have some very strange repercussions.  And traders not involved in the HFT game should know enough not to be trading during this time period anyway. (Unfortunately, when I looked at execution algorithms provided by major I-banks they didn't specifically step away from the market, but they would usually stop trading when volume dried up as they often do before economic releases.)

4. Matthew Yglesias almost endorses mercantilism.  The idea that when it comes to mercantilistic policies the public choice critique is generally better than the Ricardian/comparative advantage critique is an important one.

Puffery

Standard and Poor's is going for an interesting defense.

S&P said in its request to dismiss the case that the government can’t base its fraud claims on S&P’s assertions that its ratings were independent, objective and free of conflicts of interest because U.S. courts have found that such vague and generalized statements are the kind of “puffery” that a reasonable investor wouldn’t rely on.

And yet rating agencies are still a key part of financial regulations