Friday, August 01, 2014

Paul Ryan's poverty plan signals an ideological shift

(This post originally appeared at Bloomberg View.)
A few days ago, Republican Congressman Paul Ryan released a plan for helping people out of poverty. He unveiled the outlines in a talk at the American Enterprise Institute, a Washington-based think tank that seems to have emerged as the intellectual center of the so-called reform-conservatism movement. The plan involves making large block grants, called Opportunity Grants, to states, and instructing them to implement a raft of antipoverty programs. The most innovative of the programs involves having social workers directly help poor people take concrete steps to improve their lives in a number of dimensions.
As soon as the plan was released, a lot of liberal-leaning writers began to criticize it. Ezra Klein complained that the poverty plan would conflict with Ryan’s own deficit-cutting proposals:
Ryan's budgets and his poverty plan aren't merely different. They're flatly contradictory. They cannot be implemented in the same universe at the same time. His budget, for instance, cuts deep into the funding stream that powers the Earned Income Tax Credit. His poverty plan sharply increases spending on the Earned Income Tax Credit. His budget cuts deep into food stamps and other income-support programs. His poverty plan holds their spending constant...Ryan's poverty plan can be seen either as an effort to move the Republican Party forward on poverty or as a Trojan Horse-like effort to achieve his budget's goals by other means[.]
Paul Krugman dismissed the plan outright, saying anything that comes from Ryan can’t be trusted. Emily Badger wrote that Ryan’s punitive, deadline-based approach to personal assistance is inconsistent with what we know about how poor people make decisions. Annie Lowrey called the plan too paternalistic, labeling it “condescending and wrongheaded.” And Jamelle Bouie wrote that what the poor really need is not a “life coach,” but more money.
This is just a small sampling of the negative responses from writers on the left. Only a very few struck a positive note, such as Matt O’Brien, who pointed out some elements of Ryan’s plan that should please liberals.
These reactions are understandable -- Ryan has made a name for himself as an idea man, but this usually entails releasing plans that look bold but won’t work. And many of the criticisms liberals make represent real shortcomings of the plan -- for example, it’s clear that overall funding for poverty reduction would have to be increased substantially if it was to work.
But liberals are being much too quick to bash Ryan here. Ryan’s plan, which is being hailed by conservatives and Republicans, potentially represents a huge tectonic shift in the conservative movement’s -- and the Republican Party’s -- approach to the problem of poverty.
First, there’s the recognition that material poverty is important. Until recently, when confronted with the issue of poverty, conservatives often tended to sniff that the American poor were much richer than people in other countries, that relative poverty was just a product of envy, or that what the poor really needed was spiritual, not material, improvement.
Ryan’s plan reflects a different attitude. It recognizes that chronic unemployment and underemployment are personally destructive and a drag on the economy. It expands the EITC, which is a way of writing poor people a check. It focuses on getting poor people better jobs, higher incomes, even some additional education. The idea of personal responsibility and good behavior is still there, but now it is treated as a means to an end, and the end is better material well-being.
Second, and even more importantly, Ryan’s plan represents a sea change in the way Republicans see the role of government. In his first inaugural address, Ronald Reagan famously declared: “In this present crisis, government is not the solution to our problem; government is the problem.” Over the next 2 1/2 decades, Republicans and conservatives tended to drop the “in the present crisis” part. They’ve treated government as an obstacle to human welfare always and everywhere, instead of a tool that can sometimes be used to improve things.
Ryan’s plan is the first glimmer of a big awakening on the right -- the realization that the crisis we now face isn't the same as the one we faced in 1981. Perhaps a decade-and-a-half of falling real incomes and falling mobility has finally cracked the hard shell of triumphal post-Reaganism. If so, the fear that the conservative movement would degenerate forever into obstructionist self-parody -- that the Tea Party is the future -- has proven unfounded.
Think about it: In 2014, the Republican Party’s main idea man -- who just two years ago ran for vice president on the same ticket as a man who called the poorer half of America “takers” -- is now proposing to use a government bureaucracy to send social workers to help poor people make more money, while simultaneously mailing them government checks. That is a big, big deal. Compared with that epochal shift, the particulars of Ryan’s plan hardly matter.

Tuesday, July 29, 2014

Holey poverty, Batman!



One interesting ramification of the Paul Ryan poverty plan is that you see liberals attacking the notion that social workers can help people out of poverty. In days gone by, this was the sort of thing conservatives did, but it's a strange world we live in.

Matt Bruenig, writing for the website Demos, has an interesting blog post attacking the idea that paternalism can reduce poverty. His argument is that because poverty is "structural," changing people's behavior won't reduce poverty rates. Bruenig explains the "structural" theory of poverty thus:
The right-wing view is that poverty is an individual phenomenon. On this view, people are in poverty because they are lazy, uneducated, ignorant, or otherwise inferior in some manner. If this theory were true, it would follow that impoverished people are basically the same people every year. And if that were true, we could whip poverty by helping that particular 15% of the population to figure things out and climb out of poverty... 
The left-wing view is that poverty is a structural phenomenon. On this view, people are in poverty because they find themselves in holes in the economic system that deliver them inadequate income. Because individual lives are dynamic, people don't sit in those holes forever. One year they are in a low-income hole, but the next year they've found a job or gotten a promotion, and aren't anymore. But that hole that they were in last year doesn't go away. Others inevitably find themselves in that hole because it is a persistent defect in the economic structure. It follows from this that impoverished people are not the same people every year. It follows further that the only way to reduce poverty is to alter the economic structure so as to reduce the number of low-income holes in it.
I myself have often wondered about topics like this. Personally, I heavily doubt that paternalistic behavioral interventions could eliminate most poverty. But still, I have some problems with Bruenig's analysis. Here are my issues, in no particular order:

1. Note that Bruenig's two explanations are not mutually exclusive. It could be part one and part the other. For example, take cancer risk. Cancer risk has a well-known behavioral component - if you smoke, you're much more likely to get cancer. Cancer also has a well-known "structural" component - if you live near a toxic waste dump, you're more likely to get cancer. Cancer also has a random component, which is a theory of poverty that Bruenig did not mention. All of these stories about cancer risk are simultaneously true.

Bruenig, in his blog post, claims to demonstrate some structural reasons for poverty. Assuming that he succeeds in demonstrating this, why does that mean that individual behavior does not factor in poverty at all? That would be like saying that because toxic waste dumps cause cancer, that smoking does not. Which would make no sense. Bruenig does not show that "structural" factors quantitatively explain all of poverty. The particular "structural" factor he discusses - namely, life-cycle effects - explains at most a fraction of poverty. What about the rest? 

2. Next, there's the question of how to differentiate "structural" factors from "individual" behavior on a mass scale. Suppose everyone, acting independently and on their own, just stopped doing any sort of work. The entire world would instantly fall into poverty. But would that poverty be "structural" or "individual"? Note that if you're the one person who wants to work in a society of people who do no work at all, you're not going to be very productive, because you can't specialize. So in the case where no one works, no individual's behavior can hoist him or her out of poverty. So that's "structural", right? But if you simultaneously persuaded a large fraction of people to work, then large numbers of people would climb out of poverty. So is the poverty "individual"? To be honest, I'm not sure, since I don't think Bruenig has defined the terms as generally as he might have.

This is a technical/philosophical issue, but it's an illustration of how the distinction between "structural" and "behavioral" poverty can get very slippery in these discussions.

3. It occurs to me that if poverty is entirely "structural," then depending on who gets assigned to the poverty "holes", we might not care about poverty at all! For example, suppose that income depends only on age - for some reason, you make less when you're young, more when you're middle-aged, and less when you're old, but everyone who is the same age makes the same amount. In other words, suppose that the structural factor identified by Bruenig were 100% of the poverty story. In that case, we wouldn't really care about poverty, since people could just borrow when they're young, save when they're middle-aged, and then consume their savings when they're old, thus smoothing their consumption over their lifetime. Income inequality might be high, but consumption inequality would be low, and everyone would have plenty to eat at all times. So no biggie.

Or alternatively, suppose that income is completely random, but changes at a high frequency? Suppose that everyone is assigned to a random spot on the (fixed, structural) income distribution every year, but that your income one year doesn't affect your chances next year. In this case, people could just buy insurance from each other, and this would solve almost all of the entire problem, since people could then smooth their consumption over the years. Consumption inequality across the broad economy would be very low, and everyone would have plenty to eat. Again, no biggie.

So just showing that poverty is "structural" doesn't imply that we need government to plug the poverty "holes". In fact, it might imply the opposite. Bruenig's post sort of seems to suggest that poverty just a temporary phase of life, in which we all do our time and then escape. I think there's clear evidence that that's not the case, but if it were, it would seriously undermine liberals' case that poverty in America is a big problem.

In other words, I worry that Bruenig, in his zeal to strike down Paul Ryan's paternalistic ideas, might throw out the entire case for poverty reduction in the first place. 

4. Bruenig draws a general inference about the "structural-ness" of poverty that I believe is invalid. He asserts that since the set of poor people is always changing, and since individual behavior is fixed, poverty cannot be due to individual behavior. But the poverty line is just an arbitrary line. A family of 4 making $18,000 $24,000 in 2014 is not "poor" by the U.S. government's definition, but that sure doesn't sound like a lot of money. If random shocks (unemployment, health issues, family issues, etc.) push a bunch of people back and forth over this line, that will show up as a change in the set of poor people each year, but it won't necessarily make a difference in who is materially deprived.

When you ignore the line and look at the more general picture, you see that persistent, long-term inequality has increased in America. That is worrisome. The trend might be due to "structural" changes, and it might be due to large-scale changes in individual behavior. So although personally I predict that "structural" changes - especially globalization - have played a much bigger role than behavioral changes (i.e. family breakdown among uneducated Americans) in the recent increase in poverty, this cannot be proven just by observing that many people cross over the poverty line each year.

5. Bruenig also draws an inference about age and poverty that I believe is invalid. He notes that younger workers consistently get paid less than older ones, and assumes that this is a structural reason, unrelated to individual behavior. But what if workers simply learn as they work, increasing their productivity with experience? And suppose there were some interventions the government could do, early on, to give workers a skills boost before they even started? That paternalistic modification of individual behavior (on a mass scale) would compress the age-based income distribution, and allow young workers to get a head start.

We might even call such an intervention "public school."


In conclusion: I think Bruenig's heart is in the right place, and he raises some important issues. I agree partially with his conclusion (that behavioral interventions won't eliminate poverty), but not entirely (because I suspect they could do some real good, based on the example of public schools). And I think the evidence about "structural" vs. "individual" explanations is not so easily assessed as Bruenig asserts. It's not an open-and-shut case.


Updates

Bruenig responds at his blog. A couple points:
[Quantifying the structural causes of poverty] isn’t difficult for someone who is good with numbers, like Smith is. You can track certain categories like age (which nobody has control over) and see the same patterns year after year and country after country. Young adulthood, childhood, and old age show persistent market poverty spikes in every year and every country. Disability is the same way. Unemployment is the same as well, but people want to quibble over whether this is voluntary sometimes for whatever reason.
Actually, being good with numbers (actually I have software for that, but I appreciate the compliment) doesn't help one identify all the possible "structural" causes of poverty. Age won't account for all of poverty, as Bruenig's own graphs demonstrate. Observe:



As you can see from the y-axis, age doesn't explain the whole thing. (This is one of those graphs where the x-intercept is not set at zero.) Actually, this graph represents an upper bound on the amount of poverty that structural age-related factors can explain, since age may also be correlated with individual behavioral choices; when I was 23, I recall being distinctly interested in partying and trying new...experiences, and distinctly uninterested in doing any more work than I had to do in order to get by...

Unemployment is a biggie, and Bruenig is right that it's not clear how much is structural and how much is behavior-related. As in all these cases, the two can be intertwined (consider the case of youthful choices that screw you over later in life). But even without unemployment, there are a lot of people who just don't make much money, ever - whether or not they're above the official "poverty line". The paper I cited above on "persistent inequality" shows that pretty clearly.

Later, Bruenig catches a mistake of mine:
[$18,000] is actually [below the poverty line]. Well below the poverty line for a family of four, in fact.
Quite true; I was looking at the wrong number. The poverty line is $23,850 for a family of four in 2014. I have updated this in my post above. But the general point I was making is just as valid - you don't have to be below the poverty line to experience material deprivation.

More substantively, here's Bruenig talking about income fluctuations:
We care about poverty because having very little income is hard on you in the episodes where that is the case. It literally scrambles young children’s brains, for instance, fucking them up for the rest of their lives. One in four children in this country are born into it, and this is attributable to the structural problems families have in market systems (including that adding a child increases income needs but the market does not respond and that people make the least amount of money right at the same time they have kids). The existence of structural child poverty is not excused because, when children get to be middle-aged adults, they generally are not in poverty. You don’t eat food across the life-course. You eat it every day.
I think there's a problem here. It's not very little income that's hard on you, it's very little consumption. To understand the difference with a couple of examples, observe that A) many comfortably retired people have low income, but consume the principal of their retirement savings, and B) many children of rich families have zero income, but consume what their parents give them.

The more general point is that, theoretically, you can borrow to smooth consumption. In lean years, borrow money. In good years, save money. Then income fluctuations will not bother you much, and you will eat plenty of food every day! Of course in the real world, there are limits to how much you can borrow, how much you can save, and how much you can predict your future income - but economists find that "consumption smoothing", though not universal, is still very widespread across the economy.

So my main problem with Bruenig's argument is still that he seems to equate structural poverty with transitory poverty. Transitory poverty is not a big deal, because people can smooth their consumption. Intuitively, most liberals realize this - we worry about the poor people for whom poverty is not just a short-lived phase. We worry about the people who are mired in poverty. We don't worry about the entry-level worker who is just starting to climb the ladder. We don't worry about the well-off elderly couple who are spending their nest egg. If government poverty statistics call these people "poor", that's a problem with the government statistics, not with the structure of society. The people we want to help are the people who never (or almost never) manage to get ahead. I think Bruenig, in his zeal to demonstrate his idea, is in danger of forgetting that.

Saturday, July 26, 2014

How to beat Russia



Geostrategic analysis is so easy, anyone can do it. Just make some historical analogies, lean a little too heavily (but subtly!) on national stereotypes, and try to sound smart. Anyone can get into this game! So why not Yours Truly?

OK, so what is Russia up to in Ukraine, and what can we do about it? Let's think about this.

Thought 1: Historically, Russia is a cautions nation. After their initial burst of expansion, they have made relatively few territorial acquisitions, even after their victories in WW2 and the Napoleonic Wars. Caution makes sense for a big country - you already have a lot, so you have more to lose. Smaller countries, like Germany and Japan and France, have been more tempted to throw the dice.

Thought 2: If Russia is going to expand, it's going to expand to the west. To the east, there are A) some sparsely populated Central Asian countries that Russia already dominates, and B) China, which is now Russia's ally. Russia already has so many resources that further resource-grabbing is much more trouble than it's worth - the only kind of expansion that would really change Russia's status is to absorb some of the populated lands of East Europe. (Why Russia would want to take over these places is not clear, but back to that later).

Thought 3: Russia has a pretty simple strategy for dominating places in Eastern Europe. Basically, take advantage of ethnic divisions (if necessary, stirring them up), and just be by far the biggest, most unified, most powerful ethnic bloc in the area. The weaker ethnic group in any local conflict will naturally look to Russia to be their patron. You can see this strategy at work in South Ossetia, Abkhazia, Transdniestria, and now in Ukraine. It's a slower, more cautious variant of a very old, very effective imperial strategy used by the Mongols and British, among others. It probably explains why Russia is so keen to keep Russian-speakers in the region from abandoning the Russian language.

Thought 4: Russian institutions are just not that effective. Russia's economy has never been that good. They have never had a smooth system for transferring power. Their property rights are weak, their infrastructure is poor, their industries are uncompetitive, they have poor health, etc. Most importantly, Russia suffered massive political collapses twice during the 20th century, and no one really knows if Putin has built anything more durable than the Romanov and Soviet systems.

So what can we do to stymie Russia's expansion into Ukraine and (in the future) elsewhere? First, the current strategy of creating high and uncertain costs for Russian intervention seems to have worked OK - Russia has so far refrained from sending troops into East Ukraine, despite Ukrainian successes against the rebels. Increasing our military forces in Poland and the Baltics may also be necessary.

Second, we should try to boost the economies of countries surrounding Russia, in the hope of encouraging greater popular solidarity with the central governments of those countries. The way for us to boost their economies is to implement free trade agreements between the U.S. and those countries. Ukraine, unfortunately, is already too chaotic to do this, but Romania, Poland, and the Baltic countries are perfect candidates. Also, countries traditionally allied with Russia, like Serbia, Belarus, and Armenia, are good targets for FTAs. (Update: As a commenter pointed out, we're actually not allowed to negotiate bilateral FTAs with EU countries, so we really just have to conclude an FTA with the EU itself as fast as possible.)

But third, and most importantly, what we should do is just wait. Russia's system is not robust to shocks. Putin will grow older and die, leaving no robust, stable system in his wake. Energy prices will fluctuate, wreaking havoc on Russia's economy. Low fertility will put a massive strain on the government's finances. And as Russia absorbs the costs of newly acquired satellite states and territories, without reaping any economic benefit, additional strain will be put on the Russian economy. Even if Russia takes half of Ukraine, reabsorbs Belarus, and slices off a couple pieces of Georgia, it will just collapse again before it ever becomes a real threat to the core of Europe.

And in the meantime, we should of course try to destabilize Russia by creating disruptive cheap energy technologies. Another option is to welcome mass immigration from Russia, thus decreasing the size and power of the Russian ethnic group in the long term.

I think Obama has realized the wisdom of the basic "just wait" approach, and is carrying it out. The key is not to worry - unless you're part of one of the tiny squabbling impoverished ethnic groups on Russia's intentionally chaotic border, you're safe from the bear.

Friday, July 25, 2014

Why doesn't the finance industry use DSGE models? (cont.)



(This post originally appeared at Bloomberg View. It's sort of a mass-market version of this earlier post.)


If you care at all about what academic macroeconomists are cooking up (or if you do any macro investing), you might want to check out the latest economics blog discussion about the big change that happened in the late '70s and early '80s. Here’s a post by the University of Chicago economist John Cochrane, and here’s one by Oxford’s Simon Wren-Lewis that includes links to most of the other contributions.
In case you don’t know the background, here’s the short version: Around 1980, macroeconomists abandoned the models they had been using and switched to something very different. The old kind of model was called structural econometric modeling (SEM), based on equations for economic aggregates -- investment in office buildings, consumption of cars, etc. These models were also called “Keynesian,” because they usually included some assumptions that were loosely based on the writings of economist John Maynard Keynes. The new type of model was called dynamic stochastic general equilibrium (DSGE), and it tried to account for the individual decisions of consumers and producers. Everyone, and I mean everyone in academia, abandoned SEMs in a very short period of time, and many switched over to DSGEs.
Why did DSGE models take over? Two reasons. The first was the stagflation of the 1970s. The Keynesian SEMs predicted that when the Federal Reserve lowered interest rates, it should have given the economy a boost; instead, all it did was create useless, harmful inflation. This made a big impression on economists. About a year ago I asked a group of economists whether the Fed should temporarily adopt a higher inflation target. Robert Lucas, who probably has more claim than anyone to being the father of modern macroeconomics, thundered: “We tried (stupid) inflation! It didn’t (dang) work!”
Now, it was possible to tweak the old Keynesian SEM models to explain why inflation didn’t boost the economy. But at the same time, the aforementioned Lucas and some other heavyweights such as Tom Sargent were revealing that there was a very deep reason those SEM’sshouldn’t work. It boils down to the famous saying that “correlation doesn’t equal causation.” Suppose economists noticed that businesses where people wear Star Trek T-shirts are more productive than others. Simple -- just have everyone wear a Star Trek T-shirt, and you’ll boost national productivity, right? Wrong.
In the same way, Lucas showed that trying to boost gross domestic product by raising inflation might be like the tail trying to wag the dog. To avoid that kind of mistake, he and his compatriots declared, macroeconomists needed to base their models on things that wouldn’t change when government policy changed -- things like technology, or consumer preferences. And so DSGE was born. (DSGE also gave macroeconomists a chance to use a lot of cool new math tricks, which probably increased its appeal.)
OK, history lesson over. So why is this important now?
Well, for one thing, the finance industry has ignored DSGE models. That could be a big mistake!
Suppose you’re a macro investor. If all you want to do is makeunconditional forecasts -- say, GDP next quarter – then you can go ahead and use an old-style SEM model, because you only care about correlation, not causation. But suppose you want to make a forecast of the effect of a government policy change -- for example, suppose you want to know how the Fed’s taper will affect growth. In that case, you need to understand causation -- you need to know whether quantitative easing is actually changing people’s behavior in a predictable way, and how.
This is what DSGE models are supposed to do. This is why academic macroeconomists use these models. So why doesn’t anyone in the finance industry use them? Maybe industry is just slow to catch on. But with so many billions upon billions of dollars on the line, and so many DSGE models to choose from, you would think someone at some big bank or macro hedge fund somewhere would be running a DSGE model. And yet after asking around pretty extensively, I can’t find anybody who is.
One unsettling possibility is that the academic macroeconomists of the '70s and '80s simply bit off more than they could chew. Modeling a big thing (like the economy) as the outcome of a bunch of little things (like the decisions of consumers and companies) is a difficult task. Maybe no DSGE is going to do the job. And maybe finance industry people simply realize this.
There are signs that some academic macroeconomists are starting to come to a similar conclusion. In another post, Cochrane talks about attending a recent macroeconomics conference, and how researchers are abandoning big, all-encompassing theories in favor of simpler, more targeted models designed to explain specific ideas rather than predict the whole economy. And at the Fed, academically trained economists have flat-out refused to abandon their old-style SEMs. That seems like a strong sign that the Fed hasn’t found any DSGE model that convincingly explains the business cycle.
So it seems to me that industry people and academics need to have more of a conversation than they’re having. If industry simply missed out on the big intellectual revolution of the '80s, academics need to help them get on board. On the other hand, if academics have set themselves an impossible task, they need to think hard about what to do instead.


Updates

This James Heckman quote from 2010 is sort of the converse of my post. He asks: If old-style "Keynesian" SEM's are so bankrupt, why is Wall Street still using them exclusively?
What struck me was that we knew Keynesian theory was still alive in the banks and on Wall Street. Economists in those areas relied on Keynesian models to make short-run forecasts. It seemed strange to me that they would continue to do this if it had been theoretically proven that these models didn’t work.
The standard answer to this is that the "Keynesian" models are OK at unconditional forecasting, but not at policy-conditional forecasting. But since DSGE models should, in theory, be the best models for policy-conditional forecasting, Heckman's question leads naturally to mine...

Thursday, July 24, 2014

Why I like Frequentism



My post about "Bayesian Superman" wasn't actually intended to be a knock on Bayesianism - it was just about a quirk of rationality. I certainly don't think Bayesianism is a "dangerous religion that harms science"!

But reading this Andrew Gelman essay made me think about Bayesian inference in science, which then got me to thinking about Frequentist inference, and why I think Frequentism is a bit underrated these days.

Frequentist hypothesis testing has come under sustained and vigorous attack in recent years. It's arbitrary, it doesn't obey the likelihood principle, it throws away information, it can lead to silly results. All this is true. But there are a couple of good things about Frequentist hypothesis testing that I haven't seen many people discuss. Both of these have to do not with the formal method itself, but with social conventions associated with the practice. These are:

1. The unwritten rule that you "don't protect the null hypothesis" (or that you "penalize type I errors relative to type II errors"), and

2. The implicit three-valued logic of "hypothesis rejection".

The first of these is about what kind of prior (to use Bayesian language) the scientists should start with. It basically says that you should bias your conclusions against your own hypothesis. This is in contrast to, say, using flat or "uninformative" priors.

The second social convention is more amorphous and difficult to define. It's about what conclusions you draw from the hypothesis test. If you don't protect the null, and you reject the null in favor of your own hypothesis, Frequentism says you've found something interesting, and you should follow up on it. If you fail to reject the null, though, it doesn't mean you believe in the null any more than you did before - it means you shrug and move on. Frequentism implicitly assigns results to one of three categories: 1. "interesting evidence against a hypothesis", "interesting evidence for a hypothesis", and "nothing interesting". It's sort of like three-valued logic, in a way. Compare this to the implicit logic of the likelihood principle, in which you compare alternative hypotheses directly.

Why do I like these social conventions? Two reasons. First, I think they cut down a lot on scientific noise. "Statistical significance" is sort of a first-pass filter that tells you which results are interesting and which ones aren't. Without that automated filter, the entire job of distinguishing interesting results from uninteresting ones falls to the reviewers of a paper, who have to read through the paper much more carefully than if they can just scan for those little asterisks of "significance". Naturally, this filter also has a downside - it creates publication bias against "negative results" that may in fact be interesting. But that may be a small price to pay to avoid the flood of paper submissions that would result if everyone just wrote up and sent in the results of any estimation exercise.

Second, the discipline of the Frequentist social conventions acts against scientists' natural tendency to favor and promote and believe in their own theories. It tries to enforce the idea of scientific honesty. Feynman talks about this in his famous speech, "Cargo Cult Science":
It's a kind of scientific integrity, a principle of scientific thought that corresponds to a kind of utter honesty--a kind of leaning over backwards. For example, if you're doing an experiment, you should report everything that you think might make it invalid--not only what you think is right about it: other causes that could possibly explain your results; and things you thought of that you've eliminated by some other experiment, and how they worked--to make sure the other fellow can tell they have been eliminated. 
Details that could throw doubt on your interpretation must be given, if you know them. You must do the best you can--if you know anything at all wrong, or possibly wrong--to explain it.
The kind of integrity Feynman is talking about concerns systematic error. The Frequentist social conventions are an attempt to do something similar for random error. This provides a natural defense against "scientific trolling", which is a term I just invented to mean "the tendency of unscrupulous researchers to report weak results to advance some ulterior agenda." Scientific trolling means that ulterior-motivated researchers will submit a flood of weak results, while scrupulous, scientifically-motivated researchers will voluntarily restrain themselves from reporting equally weak results that go in the opposite direction. That sort of reporting bias will tend to contaminate the beliefs of a neutral observer. (I can think of at least one very good real-world example of this, but I will be polite and not discuss it on a blog.)

Now, of course, the Frequentist social conventions are weak, inadequate defenses against subjectivism and noise. They have drawbacks, like discouraging the reporting of negative results. And they are subject to being gamed by unscrupulous researchers. But at least they are something.

Bayesian inference seems to me like a perfectly fine and good method of inference. It's more appealing in many ways than classical Frequentism. But I think that Bayesianism might want to get some standardized social conventions similar to (and hopefully superior to) the Frequentist ideas of "not protecting the null" and "statistical significance" (note: it may already have these, and I'm just not aware of them). These conventions would unavoidably be arbitrary, and would throw away some information in many cases. But they would help lean against the natural incentives of the scientific reporting and publishing process. Maybe there could be more than one set of conventions, for use in recognizably different situations.

Classical Frequentist hypothesis testing is probably on its way out in the long term. But the fact that it has survived and dominated scientific publishing as long as it has, in spite of all its well-known problems, might be a testament to the usefulness of the unspoken social conventions associated with it.

National unity sure is helpful



(This post originally appeared at Bloomberg View.)

These days, many people tend to dismiss anything Pat Buchanan says. But in a recent column in WorldNetDaily, Buchanan raises an important point: America could use a bit of nationalism. By “nationalism,” I don’t mean jingoistic sentiment and aggressive militarism. I mean a sense of national cohesion, of being one unified people instead of just a collection of unrelated individuals who happen to live in the same geographical space. Here’s Buchanan:
Will America remain one nation, or are we are on the road to Balkanization and the breakup of America into ethnic enclaves?...
We were not a nation of immigrants in 1789. They came later. From 1845-1849, the Irish fleeing the famine. From 1890-1920, the Germans. Then the Italians, Poles, Jews and other Eastern Europeans...From 1925 to 1965, the children and grandchildren of those immigrants were assimilated, Americanized. In strong public schools, they were taught our language, literature and history, and celebrated our holidays and heroes. We endured together through the Depression and sacrificed together in World War II and the Cold War.
By 1960, we had become truly one nation and one people...But we are no longer that “band of brethren.” We are no longer one unique people “descended from the same ancestors, speaking the same language, professing the same religion.”
We are from every continent and country. Nearly 4 in 10 Americans trace their ancestry to Asia, Africa and Latin America. We are a multiracial, multilingual, multicultural society in a world where countless countries are being torn apart over race, religion and roots.
We no longer speak the same language, worship the same God, honor the same heroes or share the same holidays. Christmas and Easter have been privatized. Columbus is reviled. Stonewall Jackson and Robert E. Lee are out of the pantheon. Cesar Chavez is in...
If a country is a land of defined and defended borders, within which resides a people of a common ancestry, history, language, faith, culture and traditions, in what sense are we Americans one nation and one people today?
Buchanan is right to be worried about American people’s sense of national unity. A famous 1999 paper by William Easterly, Reza Baqir, and Alberto Alesina documented that in places with more ethnic divisions, the government was worse at providing public goods -- things like infrastructure, public health and other goods and services that the government is usually the best at providing. A substantial body of research supports this conclusion. Although people still argue about the reason, it seems clear that when the people of a country don’t consider themselves to be “one people,” the government becomes less effective.
America may be suffering somewhat from this problem. Since the '70s, elements of the Republican Party have exploited ethnic divisions to reduce support for government spending, especially in the South where racial tensions run high. Basically, a large number of poor and middle-class white people seem to have become (wrongly) convinced that government spending only benefits black or Hispanic people. In practice, cutting government spending usually doesn’t mean cutting redistributive transfers like Medicare -- it means cutting things that offer small benefits to large numbers of people, such as infrastructure and scientific research.
Has immigration exacerbated this problem? Maybe, but I think Buchanan is seriously overstating the danger. If Hispanic immigration led to the Southwest turning into a Spanish-speaking region -- America’s version of Quebec -- then there might be cause for alarm. But that is just not happening. Hispanic immigrants are becoming English speakers even faster than the earlier waves of European immigrants that Buchanan cites. Hispanics are rapidly increasing their educational attainment, and mixing socially with other groups at high rates. Intermarriage between Hispanics and other races is extremely high. And all of these trends are even more pronounced for Asian immigrants, who have recently taken over from Hispanics as America’s biggest group of new immigrants.
So on the language, education and marriage front, there is nothing to worry about. But what about the cultural and historical front? The continued existence of anti-immigrant sentiment on the political right is reminiscent of the “know-nothings” who resisted German and Irish immigration in the 1800s. Are lower-class conservative Americans going to continue to refuse to see the new immigrants as their fellow countrymen?
Buchanan suggests that the Depression and the World Wars, along with a unified history curriculum in public schools, were responsible for the assimilation of the last immigrant wave. Well, we are just now emerging from the closest analogue to the Depression that we are (hopefully) likely to see. And our history curriculum, while it has changed in substance, doesn't seem to vary from state to state more than it used to – in fact, it may vary less, given the influence of national tests like the advanced-placement history test. Stonewall Jackson and Robert E. Lee, for their part, probably don’t mind being written out of the “pantheon” of the nation from which they fought so hard to secede.
As for world wars -- well, let’s just hope that those aren't necessary for national unity.
Overall, things look good for immigrant assimilation. The main threat to Buchanan’s cherished ideal of national unity comes from nativist elements of the political right, and their refusal to accept Hispanic immigrants as legitimate Americans. Let's hope the pro-immigration wing of the conservative movement, typified by Ronald Reagan, George W. Bush, and Marco Rubio, will win out.

Tuesday, July 22, 2014

Bayesian Superman



Consider Proposition H: "God is watching out for me, and has a special purpose for me and me alone. Therefore, God will not let me die. No matter how dangerous a threat seems, it cannot possibly kill me, because God is looking out for me - and only me - at all times."

Suppose that you believe that there is a nonzero probability that H is true. And suppose you are a Bayesian - you update your beliefs according to Bayes' Rule. As you survive longer and longer - as more and more threats fail to kill you - your belief about the probability that H is true must increase and increase. It's just mechanical application of Bayes' Rule:

P(H\mid E) = \frac{P(E\mid H) \cdot P(H)}{P(E)}

Here, E is "not being killed," P(E|H)=1, and P(H) is assumed not to be zero. P(E) is less than 1, since under a number of alternative hypotheses you might get killed (if you have a philosophical problem with this due to the fact that anyone who observes any evidence must not be dead, just slightly tweak H so that it's possible to receive a "mortal wound").

So P(H|E) is greater than P(H) - every moment that you fail to die increases your subjective probability that you are an invincible superman, the chosen of God. This is totally and completely rational, at least by the Bayesian definition of rationality.

If you think this is weird, consider a closely related question: Suppose H was true, and you were an invincible God-protected superman. How would you know? Simple: you'd go get in a bunch of dangerous situations, and when you failed to die, you'd gradually realize the truth. The more dangerous the situations you chose, the faster the truth would be revealed - if you jumped in front of a train and failed to die, that would be stronger evidence than if you just drove to the store without a seat belt and made it back safely.

But nevertheless, every moment contains some probability of death for a non-superman. So every moment that passes, evidence piles up in support of the proposition that you are a Bayesian superman. The pile will probably never amount to very much, but it will always grow, until you die.

So does this mean that any true Bayesian will eventually accept invincibility as his working hypothesis? No. There are other hypotheses that explain your failure to die just as well as H does, and which are mutually exclusive with H. The ratio of the likelihood of H to the likelihood of these alternatives will not change as you continue to live longer and longer. But this gets into a philosophical thing that I've never quite understood about statistical inference, Bayesian or otherwise, which is the question of how to choose the set of hypotheses, when the set of possible hypotheses seems infinite.

Anyway, it seems to me that "Bayesian superman" is not just a prescriptive definition of how a rational person reacts to his own failure to die, but also a description of the observed behavior of teenagers. Teenagers - especially boys - are commonly described as "thinking they're invincible." Often this belief noticeably decreases when a young man sustains a major injury, witnesses a major injury or death, or has a close call with death. This, I suspect, is because most people's belief in their own invincibility is a little more complex than the "Proposition H" I've written above - it includes things like "God is looking out for my friends," and "God will stop me from getting the s*** kicked out of me," etc. A lot of people seem to start life with a non-negligible belief that this kind of special protection is watching over them. Sadly, they all eventually learn otherwise.

Monday, July 21, 2014

You'll get no edge with Zero Hedge



(This post originally appeared at Bloomberg View.)


I can’t tell you that gold is a bad investment. Even after the recent plunge, if you bought gold in 2004, your investment would have earned you an annualized rate of about 10.4 percent, after accounting for inflation. That is darned impressive. If you bought in 1994, it would have earned about 3.9 percent per year -- not too shabby. Even if you bought all the way back in 1984, you would have earned 1.8 percent in real terms. (Of course, this assumes that shadowstats.com is wrong, and that inflation hasn’t been massively understated.)
In addition to delivering decent long-term returns, gold has been a way to spread or offset investment risk. As my co-blogger Yichuan Wang showed last year, gold’s return is somewhat negatively correlated with interest rates, so that a bet on gold is to some degree a bet on lower rates. This is actually the prediction of some old economic models, which also indicate that gold should have a positive rate of return over the long term. But a lot of the variance in the price of gold isn't explained by such factors, meaning that some small amount of gold is a valuable addition to any well-diversified portfolio.
But there are two big words of caution with respect to gold. The first is that you shouldn’t believe the standard story for why gold will go up. The second is that you should be very wary of websites and media outlets that constantly push you to buy gold.
The standard story for why you should buy gold is that it’s a hedge against the inherent weakness of the fiat money system. Unfortunately, it isn’t. For example, gold is a poor hedge against inflation. The correlation is very weak. Remember that gold had its huge bull run in the 2000s and a long slide in the '80s…but inflation was higher in the '80s than in the 2000s!
A more speculative and extreme version of the story is that the whole fiat money system is destined for collapse, and that after this happens we’ll go back to using gold as money. If that did happen, you’d want to own a lot of gold at that moment. Unfortunately, there are some big problems with this story, too. Technology has advanced to the point where we can use things like Bitcoin instead of heavy, easy-to-steal physical commodities like gold. And if civilization collapsed to the point where we couldn’t even use computers anymore, I’d advise you to invest in guns, ammunition, seeds and antibiotics instead of gold.
A bigger problem with the gold story is the question of why you should expect it to earn a good return. For gold’s price to keep rising steadily due to the failure of the fiat money system, it has to be the case that more and more people will steadily realize that the story is true. So a bet on this gold story is a bet that your macro perspective is way, way ahead of the macro perspectives of most other investors. That’s a highly speculative, risky bet.
So you should beware of media outlets that constantly push this story on you. The most important such website is probably Zero Hedge. If you read Zero Hedge, you’ll see this story about gold and fiat money being promoted again and again and again, often mixed with a healthy dose of political ideology and references to “Austrian economics.”
If you actually take Zero Hedge’s constant gold-flogging to heart, you could lose a lot of money. Since gold hit a peak in 2011, it has lost about 33 percent of its value in real terms. Zero Hedge kept touting gold all the way down. For example, in November 2011, Zero Hedge ran an article saying that gold could be “on its way to infinity.” In March 2012, Zero Hedge urged its readers to “stay long gold.” An October 2012 article made the same recommendation. As the price fell, Zero Hedge assured us that the collapse was only in “paper” gold, not the physical commodity. Needless to say, if you took Zero Hedge’s advice at any of these points, you would have lost a lot of money.
Now, if the gold crash is only temporary, and someday gold heads toward infinity, then losing money on paper is no problem…unless, of course, you have to sell to cover retirement expenses or pay some medical bills.
A lot of finance people seem to treat macro stories, like the one Zero Hedge pushes, as entertainment rather than actionable information. That’s a healthy attitude. And it’s true that Zero Hedge occasionally does some excellent reporting, or publishes other good information. So maybe you can read the site just for those tidbits, and either ignore or just smile indulgently at the huge volume of gold-flogging politics-tinged macro-propaganda that the site hurls at you day after day. But then you’re like one of those rare people who really does read Playboy magazine for the articles.
Zero Hedge is still pushing gold. It’s still pushing Austrian economics. But should you be listening? Only if you’re very good at separating bedtime stories from reality.

Sunday, July 20, 2014

Are liberals rescuing marriage?

(This post originally appeared at Bloomberg View.)
When we think about the issues of family, sex, and morality there’s a standard story that frames our thinking. Ross Douthat, writing back in January, did a good job of telling that story:
[Liberals might want to] acknowledge the ways in which liberalism itself has undercut the two-parent family — through the liberal-dominated culture industry’s permissive, reductive attitudes toward sex, and through the 1970s-era revolution in divorce and abortion law.
It seems obvious that sexual permissiveness discourages marriage. In the old days, marital sex was the only socially acceptable kind of sex, so if you wanted to have sex, you had to get married; allow people to have sex outside of marriage with no social sanction, and you take away a big part of the impetus to get married. Liberal values, therefore, are corrosive to families.
But is this true? Last year, I read a book that changed my outlook on American society: Charles Murray’s ``Coming Apart: The State of White America, 1960-2010.” In that book, Murray convincingly argues that a class divide has emerged in America, between educated people (“Belmont”) and uneducated people (“Fishtown”). The surprising thing is that Murray finds that the educated Belmonters, despite being much more socially liberal than their Fishtown counterparts, actually have more traditional family values -- they get married more, get divorced less and pay more attention to their children. After a period in the '70s and early '80s when educated people dabbled in single parenthood and high divorce rates, they went back to a traditional-family structure. Uneducated Americans, on the other hand, are abandoning marriage and two-parent child-rearing in droves.
It isn't just Murray finding this. Richard Reeves of the Brookings Institution has identified something very similar. In a landmark article in Atlantic magazine, he gave a detailed, data-driven case that educated Americas are rebuilding the institution of marriage:
[C]ollege graduates in the United States are reinventing marriage as a child-rearing machine for a post-feminist society and a knowledge economy. It’s working, too: Their marriages offer more satisfaction, last longer, and produce more successful children.
The glue for these marriages is not sex, nor religion, nor money. It is a joint commitment to high-investment parenting…Right now, these marriages are concentrated at the top of the social ladder, but they offer the best—perhaps the only—hope for saving the institution.
How can this be? How can the people who preach sexual license be the same ones who have rebuilt their families, while lower-class people who profess more conservative values are seeing their families crumble?
Social conservatives have a few hypotheses. One is that high-investment parenting is only for the wealthy -- that in our stratified society, lower-income people know that investing a lot of effort in the kids won’t pay off, so they just don’t bother. A variant of this is the idea that lower-class men are too risky for lower-class women to marry.
In his book, Murray implies that educated liberals are a bit hypocritical. He calls on them to “preach what they practice,” in the hope that they can spread their newly recovered family values to the Fishtown masses. But a thought occurs to me: What if that’s exactly what they’ve been doing all along? What if sexual permissiveness and feminism, instead of being toxic to the institution of marriage, are the key to saving it?
That might sound crazy, but here’s the case in a nutshell. If you wait until marriage to have sex, you’re taking an enormous risk. What if you’re not compatible? Or what if you regret not having shopped around?
Sexual permissiveness means that sex isn’t about marriage. But that means that marriage isn’t about sex. Most of the upper-class liberal educated Americans I know who are in stable, happy marriages had their share of premarital sex. Knowing what that lifestyle is like -- and realizing that they wanted more -- allowed them to be more content in their marriages, and more realistic about what marriage is all about (i.e., lifetime companionship and raising kids).
Feminism may be even more important for families. With traditional gender roles, only a man who can be a sole breadwinner is a worthwhile mate. That rules out a lot of men, and it might be a reason why less-educated Americans’ conservative values are holding them back from getting married. Feminism, on the other hand, rewards fathers for sharing child care and housework, and frees them from the heavy burden of antiquated expectations.
In other words, maybe liberal morality is simply better adapted for creating stable two-parent families in a post-industrialized world. Maybe conservative family values are hard but brittle, like diamond, while liberal family values are strong like titanium -- able to bend without breaking.
If this is true, then I feel hopeful about American families. In many cases, educated and upper-class Americans are social trend-setters, while less-educated Americans catch up eventually. Perhaps America’s “Fishtown” class is simply doing the same thing the “Belmont” class did 30 years ago -- experimenting with single parenthood -- and will eventually learn how to do gender-equal marriage and high-investment parenting.