Thoughts from thesis

The gold standard of statistical significance is p<0.05, which means that there’s less than a 5% chance of getting the data you did if your conclusion is false.  But this means that up to one in twenty papers presents an incorrect conclusion!  And if you account for the fact that only interesting stuff gets published, the number of false positives could be a lot higher.

Chew on that, science reporters!

(Update: this article talks about publication bias, claiming that such bias might invalidate up to one in three animal-model studies.)

Pick your inequality

Tyler Cowen:

Wednesday I was in The City and out of The Suburbs. It is always startling how much better looking people are in The City, and how much looks matter there. These same good-looking people are left-wing Democrats to a high degree. They are reveling in the primordial inequality, namely that of looks and social alliances. Inequalities of wealth are of more recent vintage, from an evolutionary point of view. It is interesting how well young Democrats do at this inequality game and with what enthusiasm; only a few southern Republican women can rival them….

I think of The Suburbs as a revolutionary development. Pull people apart from their neighbors and pair them off in separate parcels, so that looks don’t much matter. Enhance some of the new realms for inequality competition, namely jobs and income. Overall lower the level of inequality competition. Subsidize the ugly. Put them in cars.

We all have blind spots when we talk about mental categories like “inequality.”  While lefties are concerned with economic inequalities and social inequalities between particular large groups, inequalities in looks, charisma, and social networks aren’t “real” inequalities to be mitigated.   And while righties like laissez-faire in market competition, they seek to reduce social inequalities through rules that apply to everyone.  Since pretty, well-liked people can get away with more in a state of nature, these moral rules most constrain the behavior of the socially best-off.  Social mores are progressive taxation on status!

There’s yet another form of inequality, that appears to cut across party lines.  This is inequality in influence.  Hippies and Tea Partiers alike decry it (as “elitism”); academics and business leaders support it (as “meritocracy”).  And, despite my pleas, nobody seems to care about temporal inequalities.

One thing that people do care about, however, is inequality of thesis completion, a very serious issue that I should probably go back and start redressing.

Could the healthcare bill be good for medical system innovation?

While I am not a fan of a public option as it is likely to appear, this is not because I am particularly fond of the current mix of Medicare/aid and private insurance companies, either.  Short version: tax credits for employer-provided insurance unfairly ties people to their jobs.  Third-party payment means that patients demand more care (since it’s free to them), doctors perform too many procedures (because they get paid for each), and insurers have an incentive to make things inefficient (the fewer claims they approve, the less they have to pay.)  An excellent but longer summary is here.

Looking at it that way, there really isn’t that much difference between a government-run single-payer system and the system of private insurers we have right now.  Yes, politicians will be incentivized to spend more on rare but photogenic diseases rather than focus on cost-effectiveness and balancing the budget.  And when cuts will be made, they will be affect the politically weak rather than those best able to bear it.  But in the broad strokes, Uncle Sam and United Healthcare will behave similarly in the marketplace, and the flaws in one will likely be replicated in the other.

Nevertheless, there are many problems with the current system, including escalating costs, reduced patient time, and gross underpayment of primary care.  I see innovative medical practices, such as medical homes (where a primary care physician coordinates care for a complex chronic illness) and retainer medicine (where physicians are paid an annual fee, competing on how healthy their patients stay rather than how many procedures they perform), as a way to address these issues by offering better value and demonstrating superior practice methods.  And the main reason I oppose increased government intervention in healthcare is because such interventions always seem to entrench the current model of employer-provided, low-deductible health insurance, and squelch such innovations.

The recently-passed healthcare bill does indeed follow in this tradition.  In particular:

Limiting Health Flexible Savings Account Contributions.  Limits the amount of contributions to health FSAs to $2,500 per year, indexed by CPI for subsequent years.

Promoting Individual Responsibility. Requires most individuals to obtain acceptable health insurance coverage or pay a penalty.

Health savings accounts are a tax-free way to save and pay your own medical expenses out of pocket, thus providing an alternative to health insurance.  And the penalty for not having “acceptable” health insurance is likely to apply to all sorts of innovative medical systems, including those listed above and, most likely, high-deductible plans – low-cost plans that cover only catastrophic care, and which sidestep lots of the bad incentives of the current insurance setup.  Essentially, if you don’t follow the government-approved business model, you get taxed.

So why do I say that the bill could be good for innovation?  Because as I outlined in my previous post, the bill is likely to lead to significant increases in insurance premiums, with healthy folks opting to take the fine and go without health insurance.  In the meantime, they have to get healthcare somewhere!  While these conditions last – and again, they are unlikely to last for long before another round of intervention comes along – these conditions are perfect for nontraditional practice models to expand and prove their effectiveness.  For these few years, there will be droves of patients looking for alternatives to insurance, and if such practices can effectively expand and gain public goodwill, they will become more politically powerful and be much harder to squelch the next time around.

The healthcare bill will mandate adverse selection

After noting that Princeton seniors are busily scrawling away at their theses, the House has graciously released a seven-page summary of the thousands of pages of healthcare bill that passed earlier this week.  Of note is how small the effects will be early on – early provisions focus on things like “cutting waste, fraud and abuse” and and “improving consumer assistance” – things with relatively small budget impacts.  There was significant pressure to keep the cost of the healthcare reform bill under $1 trillion, which explains why the early years focus on relatively small research grants and taxes on various rich folks.

The real kicker starts in 2014:

Reforming Health Insurance Regulations. Implements strong health insurance reforms that prohibit insurance companies from engaging in discriminatory practices that enable them to refuse to sell or renew policies due to an individual’s health status.  Health plans can no longer exclude coverage for treatments based on pre‐existing health conditions.  It also limits the ability of insurance companies to charge higher rates due to heath status, gender, or other factors.

Promoting Individual Responsibility.  Requires most individuals to obtain acceptable health insurance coverage or pay a penalty of $95 for 2014, $325 for 2015, $695 for 2016 (or, up to 2.5 percent of income in 2016), up to a cap of the national average bronze plan premium.

The problem with the first provision is that it basically ensures that adverse selection will happen on a massive scale. If everyone can buy the same level of insurance, regardless of their health condition, the people who are most likely to buy it are those with preexisting conditions who expect to undergo expensive procedures.  As high-cost patients sign on, the insurer must cover these costs by increasing its rates, thus making health insurance a worse and worse deal for relatively low-cost healthy people.  As healthy people drop out, premiums must rise further.  After several iterations of this, the equilibrium is a series of very expensive insurance pools covering the sickest and most costly-to-treat patients, with everyone else priced out of the market.  What this really means is that the winning strategy is not having health insurance while you’re healthy, and then signing up for it when you come down with an expensive ailment – there’s no way the insurance company can turn you down, and they have to foot the bill.

Paul Krugman has argued that if you ban differential pricing based on preexisting conditions, then you must have a mandate to avoid adverse selection.  I agree with this argument – you can’t have one without the other – though I disagree on the aims (he wants both; I’d prefer neither).  But the second provision – theoretically a mandate – imposes a puny penalty for not having health insurance.  The penalty for employers who do not offer insurance are correspondingly puny.  There is no way, with adverse selection operating at full blast, that a penalty of this size will induce healthy folks to sign up for expensive insurance policies that they are virtually guaranteed to lose money on.

Of course this turn of events will hardly go unnoticed.  As David Henderson imaginatively explains, it is unlikely that people will make the connection between the nice-sounding anti-discrimination provision and skyrocketing insurance premiums.  Most likely this will be a target for demagoguery about those Greedy Insurance Companies gouging their customers, and politicians, avoiding blame for their own actions that precipitated these conditions, will turn around and use this populism to push through even more regulation.  Plausibly, this second stage of reform will involve raising the penalties for not having government-approved health insurance, or the creation of a public option, which will look much more attractive after the private insurance market has been trashed.  In this way, although the recently-passed bill is more modest than House’s initial offering, it leaves a backdoor to a public option.

From the Department of Who’daThunkit

From here, a ranking of occupations by whether they donate to liberal or conservative politicians.  As usual, the precise centering of the graph is not terribly important, both because it depends on the vagaries of the survey and because this graph only counts folks who donate, not the median voter.  But the rank-order is pretty interesting.  On the left we have Hollywood workers, professors, and lawyers.  On the right we have engineers, dentists, and accountants.

Two trends of note:

1. On both sides, you have folks who are ideological because that’s the side their bread is buttered on.  Civil servants are Democrats and oil and gas workers are Republican, which makes sense given the respective platforms.
2. Overall, the professions on the left are more reputation-based, with few quantifiable signals of success – you don’t have a “Certified Hollywood Producer” the same way you’d have a CPA, for example. Mostly, they rely on word of mouth and personal reputation.  On the right, you have professions whose output is more quantifiable – accountants, real estate agents, and the like.  I’m not sure what to make of this, except maybe a measure of whether you prefer ostensible egalitarianism vs. blatant meritocratic hierarchy.

A note on the “Healthcare Professionals” data point – when I was interviewing for med schools, I was often asked my views of healthcare reform.  More often than not, that was a platform for the interviewer to explain his own views on the matter.  I found that there was a sharp difference between academics and practitioners – doctors who spent most of their time teaching liked universal healthcare; those who practiced hated the idea.  So I suspect that the middling position of healthcare professionals is a composite of a bimodal distribution, with academics on the left (like professors in general?) and practitioners on the right.

Thoughts prompted by HBO’s Rome

Despite being a pinnacle of ancient civilization, Rome was poor. The very few aristocratic families scraped out an existence comparable to a middle-class American. (Bonus: slaves.  Downside: no medicine.) The vast majority of Romans, from the mercantile class downward, were arguably materially poorer than inhabitants of third-world countries today. If ancient Rome were transported to the present, it would be one of the worst-off countries in the world. Slave labor is nice for non-slaves but it cannot get you past primitive technology and (at the time of the series) poor governance.

More broadly, what this implies is that if you’re really concerned about income inequality, you should be just as concerned about temporal inequality as with inequalities in a static time frame, such as between rich and poor in the present-day USA. (This is forgetting about concerns about redistributing from the rich US to the global poor, which no serious politician places as a high priority). Transferring goods those poor blokes in the past is, of course, impossible. But this line of thought also implies that we should really be worried about is improving our own lot at the expense of those lucky bastards who live in the future, where technological progress makes material goods increasingly cheap, and who will probably look at our living conditions as backwards savagery.

Now this we can do.  How?  Higher consumption, and lower investment. Spending more money today will make us better-off today. The downside is that we don’t get the payoff of investment in the future – but they don’t need it, what with their Playstation 10,000s and flying cars. Yes, at some level, underinvestment will decrease the total welfare of humanity, but this is a tradeoff that plain-vanilla redistributionists are already willing to make to some extent. (Again, the people who lose from this policy are those well-off far-future people.)  And given the current rate of technological growth, you could reduce investment by a significant amount without creating a futurity that’s worse off than us, which would defeat the whole point.

You could argue that we already do this tradeoff to some extent, by saving below the welfare-optimal amount. But I think that we naturally care about our children and descendants, more so than we care about our poor neighbors, and more so than would be welfare optimizing. Given this powerful emotion I think it’s safe to say that we aren’t consuming anywhere near what intertemporal redistribution would imply. So, next time you hear criticism of Americans spending too much and saving too little, just remember that we’re not being selfish; we’re just ensuring equality!

(If you buy that, by the way, another consequence of this belief is that we shouldn’t do that much about global warming. People in the future will be better-off and will be better able to deal with the problems; inaction on the global-warming front is a form of disinvestment in the future.)

Crossposted from Facebook.  Contents may be flippant.

Still haven’t written that facebook post, but…

Here’s an article on Facebook as MMORPG.

A lot of the praxis around virtual worlds — and indeed, games in general — has been co-opted by social media….

• Formal group identity is taken to a level well beyond that of the typical virtual world on social networks. (We’ve been saying for years that we should support multiple guild membership in MMORPGs… check out the typical number of groups a Facebook user belongs to…)
• Points and quantified reputation are rampant. Arguably, excessive.
• Similarly, publicly visible profile data has become the defining characteristic of much of the social web. Facebook is a collection of “avatar” pages where you can browse only one’s clothing, achievements, guild memberships, and skills — in a manner of speaking.
• Formalized user roles are also the norm, on the admin-vs-user level.
• And perhaps most pleasantly, malleability of environment is also a key characteristic. Even the most simplistic of farming games on Facebook ranks higher on the “affect your world” scale than World of Warcraft does… and this sort of personalization of the environment is standard not only in social games but across the social web today.