A Fools Game

This will be part I of my series on the NBA and statistical approaches to quantifying the value to players. This post is mostly an introduction to the debate, if you are already are well versed in it then skip to part II.

This post is a bit off the usual topics, but the NBA is one of my outside interests. There was a point in my life where I watched a lot of basketball, but now I spend much more time reading about it then actually watching games. I find the idea of constructing teams in way so that they can succeed to be fascinating, because basketball, much more than other sports, is like a game of chess.

Football and Baseball are more of a team sport, as individual players’ results depend greatly on who surrounds them. While a great basketball player can’t win by himself, he can still play great and get huge stats. A great quarterback is useless if his front line lets him get sacked or no one catches his balls. A great batter doesn’t gets less turns at bat and less runs per home run if all his team mates strike out.

As a result, some teams tend to myopically focus on finding superstars (who are easier to identify) over constructing a good team. There is a debate, however, between the “stats nerds” over the best system to quantify how good a given player is, with the two most famous being John Hollinger’s Player Efficiency Rating (PER) and David Berri’s Wins Produced (WP). These models are constructed very differently (PER theorizes a formula and plugs box stat scores into it, whereas WP regresses game results onto the box score formula to see how much each contributes to winning) and produce incompatible results.

Most prominent is that PER places a much higher value on volume shooters with low efficiency (Kobe Bryant, Carmelo Anthony, Allen Iverson, etc) than Berri’s system does.  This makes PER much more consistent with the conventional wisdom, which has historically valorized these players and showered them with MVP awards. The statisticians respond that the low correlation between payroll sizes and team wins creates an a priori reason to believe that the conventional wisdom is wrong.

I don’t think that is quite correct. Teams can only spend higher that the soft cap by utilizing various exceptions for low and mid-level players or resigning their current players for more (although they still can’t go higher than the max individual contract0. It is universally acknowledged that the max individual contract far underpays the true superstar, and most of the low-efficiency superstars would still command something close to the max contract under Wins Produced. Therefore, I think the more likely explanation is that bad teams overpay role players with these exceptions, and that is what creates the negative correlation with wins.

In part II I will deal with the main point of dispute, whether the “shot creation” of these low efficiency scorers makes them superstars.

1. Relaxing the Efficient Market Hypothesis: If people in finance (or a particular subset of people in finance) are more likely to predict a change in Monetary Policy than the general public, then they will purchase T-Bills ahead of time and essentially be given free resources from the Fed. Not only does this transfer resources away from other people (although if it succeeds in raising NGDP it still might be a Pareto improvement?), but it encourages productive people to become bankers instead of doctors.

Note that you don’t have to completely abandon EMH for this to be true. Maybe people can’t usefully predict natural market movements, but some can predict Fed induced ones.

This is probably an acceptable distortion though, since we are basically paying Bankers to transmit monetary easing from the central bank to the rest of the economy. Also, if the transition to NGDP targeting was made, in theory, it would only be a one time distortion.

In this sense, it is a reason to move to explicit rule-based monetary policy over the current informal, ad hoc approach.

2. If you hold full EMH, then you can model the unexpected increase in the price of T-Bills as an ex post facto change in the initial distribution of the economy. The people that happened to be holding the T-Bills find out that they actually have more wealth than they thought they did. While basic Econ 101 Edgeworth Boxes will tell you that this shouldn’t matter for efficiency, it does make vertical inequality worse if you assume that the rich hold a greater proportion of T-bills.

This, by the way, is the argument that some MMTers make as to why QE might hurt the economy. It transfers money to the rich, and they are less likely to spend than the poor people that lose out.

I also sense that this is why mainstream liberals oppose monetary policy, because it seems like a giveaway to bankers.

Once again, however, these effects would disappear or reverse in the long run.

I, as a matter of personal taste, despise driving and would be much happier if I could move to any city in America and get away with solely using public transportation. This country would have been better off over the past century if we had invested in trains and buses, but it’s too late. Liberals need to get over their obsession with high-speed rail, since it will be obsolete long before we can build it. Driverless cars are coming (probably before the decade is over). And they are going to change everything.

When deciding whether to take a train or drive, trains have the following advantages from the self-interested perspective of the passenger:

1. They’re faster (sometimes). Trains have their own dedicated lane, so they don’t have to negotiate with other cars and mostly are immune to traffic. Fewer drivers per passenger, means there is less room for human-error induced accidents, so we don’t need speed limits.

2. They’re cheaper for occasional use. Cars have high fixed costs to buy and let it sit in your garage.

3. You don’t have to drive. When on a train, you can sleep, do work, read, etc. Many people avoid driving long distances for this reason. It also makes their morning commute wasted time.

Driverless cars will resolve all these issues:

1. They will reduce congestion and allow you to avoid it. We will allow them to drive much faster once we realize how incompetent humans were at driving. For intercity travel, they’ll link up as convoys for super efficient and fast train-like movement.

2. No more fixed costs: Think Zipcar, but one that comes to pick you up wherever you are. People won’t have to own cars anymore, because they will be effectively shared. Even if they do, parking will be cheaper/easier, large families could more easily share one car, and they could be rented out.

3. No more driving, duh. Think of the things you could do.

A lot of people like to say that these are decades away. They are wrong. We already have the technology, but we need to pass laws, lower costs and get enough people to adopt. That will take time, but driverless cars are not all or nothing. With self-parking cars and smart cruise control, aspects of driverless cars are already on the commercial market. In the near future, they will begin to gradually cut into the advantages of trains. Plus, they have these HUGE advantages:

1. They are not limited where they can start and end.

2. Governments don’t have to spend hundreds of billions of dollars to build the infrastructure.

Plus, most importantly. Who cares if its a few decades before they show up? Most high speed rail projects take almost that much time too.

In part 1, I explained that a huge part of Bernanke’s (and other central banks’) conservatism is due to the fear that their successors will change course. Here I will deal with the anticipated reaction by market monetarists: the Chuck Norris effect.

It might go something like this: If NGDP targeting (or any new regime) is successful, that will guarantee its permanence, since no new FOMC would mess with a booming economy. It’s a perfect tautology: It’s credible because it works, and it works because it is credible. The theoretical “rightness” of NGDP targeting makes the markets know that it will work, their reaction to the expectation IS it working, and the improved economy guarantees political support.  Therefore, it is only a question of economics and not politics.

Maybe. But the lack of epistemological humility reminds me too much of the Bush administration saying “we will be greeted as liberators.” Supposedly, decades of liberal timidity had arbitrarily hamstrung us from doing what was clearly within our power and, once it became clear that we could bring stability and freedom to wherever we wanted, both Iraqis and the world would bow to the inevitability of Pax Americana.

Sumner likes to say that no central bank has ever failed to create inflation when it has tried. But I wonder if the perfect track record is due to the fact they rarely deviate from what the consensus view of economists prescribes. The cases where central banks get ahead of the consensus sometimes end badly, but these are retroactively labeled “dysfunctional” countries, where the proper institutional norms to set the correct target don’t exist. Once again, this is tautological. What makes these central banks “dysfunctional,” in the first place, is the lack of consensus-following institutions that create regime stability. 

I like to think of the market as a pretty girl in a bar and the Fed as a guy trying to flirt with her. The Fed knows he must appear supremely confident to tame the temptress, but she has an almost supernatural ability to sniff out fear. Even the smallest self-doubt will cause her to pounce in a test of his convictions (Is that really you best line? Why are you talking to me?), which of course makes him more nervous, spiraling out of control into an orgy of awkwardness as he comicly apologizes for “intruding” and slinks back to the same dark, lonely corner from where he came.

The expectations channel might make the Fed incredibly powerful, but it is directionally neutral. The awesome power of central banks requires their cautiousness in making only small changes in the right direction when there already exists a foundation of intellectual and political support.

There’s a lot to like in Krugman’s recent post on why Bernanke hasn’t lived up to the hope created by his wild-eyed academic years, but I think he is missing (or not explicitly saying) a really important part of it.

Bernanke is not going to Fed Chairman forever and, even if he could be, the FOMC’s composition will be substantially different in the relatively near future, even with only normal turnover. Any new targeting regime has to be credible not just for Bernanke and this FOMC, but for all future FOMCs. This is why critics are wrong to say that Bernanke is “caving” to Republican crazyness: if the GOP is ideologically opposed to his new target, when they eventually and inevitably return to power, they will reverse it. Any expected doubts would clearly limit the ability of the Fed to reach it’s target.

And where does that leave us, if we abandon the old, ad hoc  Taylor-rule regime and the new NGDP target fails, not because of theoretical problems, but because Romney wins the election and promises to fire Bernanke? Without the inertia of the old regime, there might be a free-for-all to determine the future of the Fed. ANY replacement would be uncredible, because no one would know how long it would last and that doubt would devolve into a self-fulfilling prophecy that dooms any replacement.

This is why central bankers see a value in the status quo. Sure, most agree that it is not ideal, but once it loses it’s de facto advantage, everyone and their mother will see this as their one shot at getting their dream enacted (Ron Paul?). It might be failing, but it is certainly better than letter inflation get out of control or abandoning the dual mandate for an implicit ceiling (but not floor) on inflation. This is why Sumner is right that the Fed sets policy at the consensus views of economists. Those economists influence the debate and ensure there is the critical mass of support needed for credibility.

But if the new target worked, wouldn’t the next FOMC support it? I’ll deal with that objection in part 2.

It is well established that self-identified liberals make up a much smaller portion of the electorate than self-identified conservatives. Whether this outnumbering is due to “liberal” becoming a dirty word is an interesting question, but not very relevant for some issues. If to the fact that self-identified liberals make up about 20% of the electorate and conservatives vote overwhelmingly Republican, then Democrats must overwhelmingly win self-identified moderates to be successful.

It is in this contest I read a fairly old report from the Brookings Institution:

I think most of us intuitively understood this, but I have never seen actual poll data. Moderates slightly side with Democrats on Economic issues and with Republicans a little bit more on national security issues. The primary reported problem that moderates have with the GOP is on social issues.

This is the best explanation for why Santorum might be a worse general election candidate than Romney (his social conservatism is a turn off for the swing group), but I don’t buy it. The Brooking’s report also looks at poll data on the 2004 and 2008 elections, finding that all three groups made up roughly the same shares of both elections. Obama won and Kerry lost, because Obama increase the Democrat’s share of both conservatives and liberals by 5% and moderates by 6%. Confusingly, the authors interpret this tiny 1% difference as evidence that Obama better appealed to moderate, but to me it just shows how the economy and other issues causes a relatively proportional shift in the electorate.

Bush was perceived to be substantially more socially conservative than McCain, and yet there is no noticeable effect on the ideological distribution of the shift towards democrats. I think this actually has far more explanatory power for Romney’s superior performance among moderates in the primaries, where the partisan “leaning” of moderates doesn’t have the same effect that it does in the general election. It also explains why the majority of the Democratic party ignores the protests of the liberal base on national security/civil liberty issues.

There are always endless debates about whether a particular policy change would “increase” or “decrease” the deficit. Increase and decrease are, of course, relative terms dependent upon whatever baseline you compare them too. When judging the impact of a particular policy change (say raising marginal income tax rates on the rich), it is perhaps unavoidable that we have some confusion over what is the proper baseline. But when discussing the overall budget that a politician proposes, there is no reason to have this problem. When rooting for our team, we like to say “well it is an improvement from before” and “those decisions were already made, I am just choosing to fix some of them,” but it is very easy to judge the overall outcome of a budget. A good example of that is Keith Hennessey’s recent post here.  The key is that he doesn’t look to whether Obama’s budget increases or decreases the deficit relative to his previous budget or Bush’s proposed budget, but he simply looks at how the deficit will change in the future if all the President’s proposals are adopted.

But just because that is the most accurate way of representing what outcome the President is proposing, doesn’t mean it is the most useful way. In an ideal world, we would design our policies with respect to the overall landscape. Consider these two choices:

(a) The government should pay for healthcare reform subsidies by raising some taxes and making cuts to Medicare. Therefore, that action is deficit neutral.

(b) The government should spend X amount, and raise Y revenues. X-Y should equal 3%, so that is the size of our deficit in the future. As a part of that X spending, the government should provide subsidies for health insurance.

Now, there are benefits and disadvantages to each of these. Option (a) makes sense from the perspective of an Economist: we should do all those government actions where the marginal benefits is greater than the marginal cost. In order to determine that, you need to compare each aspect in isolation to the next best alternative. Option (b), from this perspective, commits to some arbitrary deficit number and then just picks the best way to reach that number. If the marginal benefit on a program outweighs the marginal cost of going above a 3% deficit, then we probably should accept a higher deficit.

From Hennessey’s point of view, on the other hand, this mistakes a fundamental problem of political economy. There are many things that seem to provide a net marginal benefit at the time, but when we lose sight of the big picture, we end up spending more than we should. Those targets may be arbitrary, but in order to avoid a Tragedy of the Commons spending spree, we need to start with some number. Option (b) allows us to do this.

For me, both these answers are too idealistic. Maybe it would be better to design an overall budget holistically, but no one politician has the power to do that. The budget is the product of competing interests, and, for all practical purposes, many undesirable things become entrenched due to the nature of politics. Obama’s ideal budget certainly looks very different than the one that he proposed, but Congress will support the budget based on whether they think it improves from the reference point that they have (the previous budget).

A: Imagine you own a house that has been in your family for generations. You inherited this house when it was worth $100,000 dollars, which coincidentally enough is exactly how much it is worth to you. You have no desire to live in this house, but its history is such that people are willing to pay so that they can come and look at it like a museum. You’re good at math, so you calculate the present value of all future cash flows and find out that it is equal to $100,000 dollars.

Sam, the guy living next door to the house, realizes that your inattentiveness is resulting in a poorly run operation. He calculates that if he owned the house, then the present value of all future cash streams would be worth $130,000. Negotiations ensue and you both settle on $115,000 dollars. Capital gains taxes are 50%, so you pay the government $7,500 and pocket $7,500 in profits. Sam also makes a profit, but his is $15,000.

But wait! You have a much more profitable job that prevents you from negotiating with Sam. So you hire a real estate agent for 10% of the sale price, or $11,500 dollars. That exceeds your potential profit of $7,500 so you are not willing to sell for $115,000 and ask for a higher price. Sam realizes that, in order to make the house-museum worth $130,000, he would have to devote all his time for a year. He has another job offer that will pay him $15,000 over the same time period. He therefore rejects any offer over $115,000.

The capital gains tax has made the economy less efficient. Even though that capital would be more valuable in Sam’s hand, it has reduced you potential profit by so much that there is no deal. This is a tax on the transfer of capital.

B. You are a hot shot lawyer working at Law Firm A that pays you $100,000, because that is the marginal profit you generate for the firm. The strange thing is though that this firm hasn’t updated its computers in over a decade, and the constant crashes slow you down considerably.Law Firm B has the latest and greatest computer systems, and, if you worked there, then you would be 30% more productive and produce a profit of $130,000 for the firm. They offer you a salary of $115,000 dollars, which would leave the firm with a net profit of $15,000.

The government has a 50% income tax. You obviously accept this offer, because that would still leave you with a profit of $7,500 over your current job.

But wait! Law Firm B is in a different city, and in order to accept the job you would have to would have to suffer $15,000 dollars in moving expenses. If there were no income tax, you would break even and be willing to accept the offer. But because of it you continue to use your labor at a place where it is less productive.

This is a tax on the transfer of human capital.

Reihan Salam informs me that, if I want to gain a “better understanding of how ‘the other side’ thinks about this election,” I should go take a look Yuval Levin’s 7000 word essay because he makes the completely original claim that… wait for it… conservatives think that the Welfare state will bankrupt America. Really??? Does Salam really think liberals don’t get that.

Truthfully, Levin makes a much more interesting point. He is trying to push back against the liberal tendency to ignore the cultural and moral implications of the welfare state and insist on telling conservatives that they are bad at math. This is an argument that we have made here on this blog before, and liberals need to realize that conservatives are not just “clinging to their guns and religion” if they ever hope to win an argument with their crazy, gun-toting grandfather. I’ll save you the time of reading Levin’s entire essay and post the heart of it here:

It would begin not from the assumption that capitalism is dehumanizing, but rather from the sense that too many people do not have access to capitalism’s benefits. It would start not from the presumption that traditional practices and institutions must be overcome by rational administration, but rather from the firm conviction that family, church, and civil society are the means by which human beings find fulfillment and are essential counterweights to the market. It would reject the notion that universal dependence can build solidarity, and insist instead that only self-reliance, responsibility, and discipline can build mutual respect and character in a free society. It would seek to help the poor not with an empty promise of material equality but with a fervent commitment to upward mobility.

The modern liberal who has spent his entire political career in a dark corner hiding from the ghost of Ronald Reagan would protest that we don’t hate God, family or charity, but Levin is more right that wrong here. The welfare state certainly does act as a substitute for those other institutions that use to provide economic security, or as he puts it:

…the institutions of the welfare state are intended to be partial substitutes for traditional familial, social, religious, and cultural mediating institutions, their growth weakens the very structures that might balance our society’s restless quest for prosperity and novelty and might replenish our supply of idealism.

This is the second major failing of this vision of society — a kind of spiritual failing. Under the rules of the modern welfare state, we give up a portion of the capacity to provide for ourselves and in return are freed from a portion of the obligation to discipline ourselves. Increasing economic collectivism enables increasing moral individualism, both of which leave us with less responsibility, and therefore with less grounded and meaningful lives

This is why I always say that the whole idea of a bleeding-heart liberal doesn’t make sense in a post-Reagan world. The Democratic party is the free market party that stresses the importance of incentives driving human behavior, whereas Republicans refuse to concede the existence of The Tragedy of the Commons. As if this makes one hypocritical, Republicans are always harping on Warren Buffet for paying the legally required tax rate and Al Gore for consuming lots of a cheap resource.

They talk in the language of economics, but really for them its about personal responsibility and morality. If you think emitting carbon dioxide is wrong, then you should stop using products that cause it. And the mere fact that most people don’t choose to drive hybrids is somehow considered proof the we should Drill, Baby, Drill. The modern Democratic party instead recognizes that individual is sometimes incentivized to do things that are not in society’s interest, and the proper role of government is to step in and properly align incentives through market mechanisms like Cap and Trade. How is God, the family or charity going to serve as a “counterweight” to that market failure?

Most important to recognize is the Levin says economic collectivism enables, not requires moral individualism. Exaggerated controversies over contraception mandates aside, nothing in the modern welfare state prevents people from attending church and starting a traditional family, it just makes it unecessary. Women are getting married later and having sex with more partners, because they no longer have to rely on finding a man to provide for them, since they know, if nothing else, they will get a social security check to feed their 20 cats when they are 65. When large portions of society are no longer afflicted by poverty and the threat of violence, they feel less compelled to turn to a Church to give them bread and Commandments that (unsuccessfully) shame people into peace.

Allow it is political suicide to say this, liberals should celebrate the obsolescence of these institutions, because centuries of human history have shown that they failed to do anything to prevent life from being nasty, brutish, and short. The old “familial” and “social” institutions that Levin yearns for did nothing but force people to find security in those that were most similar to themselves. This constrains, not broadens, ethics because a required corollary of trusting those similar and close to you is fearing those different and far from you. How many wars between Churches do we have to witness before we stop trying to preserve the institution of God and merely develop a personal relationship with Him?

This is why liberals should laugh at people who drive hybrids and shun smelly hippies that chain themselves to trees. Occupy Wall Street shouldn’t be trying to convince Harvard graduates to join the Peace Corps instead of Goldman Sachs, they should be lobbying Congress to make finance a less attractive career. A true liberal recognizes that ethics does not require one to act ethically, but to design institutions to make ethical outcomes likely. When you rely on appeals to personal responsibility to prevent people from acting unethically, you reward those who are most unethical because they no longer have any competition in pursuing their self-interest (once again, see Tragedy of the Commons).

Indeed, the welfare state should be celebrated for alleviating economic insecurity, because history has shown it is the best way to make people stop killing each other. That security might breed a complacency that makes growth slightly slower, but, until conservatives come up with a better idea than already existing tax deductions that encourage people to have kids and donate to charity (as Levin suggests), I will take that tradeoff any day of the week.

A thought recently occurred to me that I have been unable to find addressed anywhere in the literature on Capital Gains taxation. I’ll pose this as a question who think that we shouldn’t have a capital gains tax:

Why should we tax gains to human capital more than gains to physical capital?

It is very common to look at education or other means of skill-building as investing in your own “human capital.” And it makes sense too: if I spend 200,000 to go to college, that will increase my potential productivity, and I can receive a higher price when I sell my human capital. Now of course you can’t sell your human capital (i.e. I will work for you till I die or you sell me for X dollars) due to the 13th amendment, but working for a salary is identical to renting out physical capital.

Seen as this way, having a differential between capital gains and income taxes may encourage investment in physical or financial capital, but it will discourage investment in human capital. I am less likely to pay the tuition for University or a vocational college and get taxed at 35% income + 12% Medicare/Social Security on the marginal income I can earn, when I could use that money to buy corporate stock that only gets taxed at 15%.

Now, you might say that tuition and public education are already subsidized, but so is saving/investing in physical capital. Most people save through taxfree IRAs and 401ks, and their companies are incentivized to give them matching contributions. This issue is about taxing returns to capital not the purchase of capital. This also eliminates any concerns about “double taxation.” Corporate profits being taxed affects incomes just as much as it affects stock prices.

The only other argument that I can think of is that physical capital is somehow more desirable than human capital. It’s not more durable, because most people have longer careers than physical capital lasts (your stock purchase may a company finance the upgrading of their computers. How long before they need to upgrade again?). Human Capital also has greater flexibility. The skills that one learns at college can make one more productive at many different types of jobs, while a car-building robot can only build cars.

The obvious conclusion to this line of thinking is that we should switch away from all taxes on income (Personal, Corporate and Capital) towards consumption taxes and Pigouvian taxes on externalities. But given that we do tax income, does anyone have a reason that we should tax returns to physical capital at a lower rate than returns to human capital?