Sunday, March 29, 2015

Every TV News Report On the Economy In One

Charlie Brooker’s Weekly Wipe is a British TV show, more or less like The Daily Show. Here they’ve put together a faux news report containing all the features that usually go into one of these things … with just about as much actual economic content as usually go into these things:

You shouldn’t always take these things you see on TV too seriously. There often isn’t very much to them.

Via Alex Tabarrok from Marginal Revolution.

Wednesday, March 25, 2015

This Post Is Not a Joke

Would you love to have a sidekick? Ideally a useful one? You sort of do already if you use your smartphone for more than texting links to kitten videos. This isn't just claptrap ... or maybe it is.

Wait for it ...


It's difficult to get across to people the scale of economic growth produced by compounding of technological improvements (e.g., a 4,000 to 1 improvement in the cost of lighting over 175 years).

However, it's critically important to do so because, even in a developed country like the U.S. that is not posting ridiculously high growth rates (like, say, the 10-15% per year that Mongolia has been averaging) it's still reasonable to forecast a doubling of well-being during an adult's working lifetime, and perhaps a quadrupling from birth to death. What exactly will that entail, when there's lots of stuff (like food, clothing, and transportation) that can't increase much? Well, it means a lot of other stuff that will go from nothing to something.†

But, it's worse that that. All this growth is caused by improvements in the productivity of workers. This comes from increased use of capital. But we've now figured out how to further improve our ability to use capital by inventing technology that allows individuals to control more pieces of productive capital. That's what technology does: it turns workers who are productive because they use capital into more effective workers who are more productive because they use technology to increase the amount of capital they control.

Now, put on your tinfoil hat: how far are we from a future in which you can control a piece of technology mixed with capital that is capable of controlling the other capital you already control with technology? What is the world going to look like when you have ... not just a drone ... but an avatar?

An example may help: I'm not talking about transitioning from human-driven cars to self-driving cars, but rather a transition to an avatar that hangs around you and that you can tell to go to the car and drive to school and sign your sick kid out of the nurse's office and bring them home?

Think about the numbers for this. We float along year after year benefiting from improvements in technology of say 2% per year, and we float through our lives while that compounds over 10 years to a smooth and steady 22% improvement that we take for granted. And then you buy your first avatar, and it isn't perfect, but it immediately gives you a 20% jump in productivity. The thing that makes this thought experiment different from other ones we might conceive is that I want to assume that the avatar also benefits from future technological improvements: because it does what you do, when your productivity improves its productivity improves too, so you'll be compounding your compounding of productivity growth.

You end up with something that charts out like this:
What's critical about this is not that the blue "curve" has jumps (each time you get a new and improved avatar), or that it's higher than the orange curve (because the avatars improve your productivity), but that the slope of each long segment of the blue curve is steeper than the orange curve. As a result the slope of the jumps gets steeper too. This is what it happens if a sequence of improved avatars gets the same productivity improvements that its person gets.

BTW: What are you going to do when you're old if you're on the blue curve instead of the orange one, and you live long enough? What will you buy? Trinidad? Or will you be outbid by someone with 2 avatars?

I'm writing this because yesterday, a retrospective video game box set went on sale: Borderlands: The Handsome Collection — Claptrap-In-A-Box Edition.

When you play Borderlands, you're assisted by an annoying but useful robot sidekick named Claptrap that can act somewhat independently of you.

The thing is, the new box set comes with your own physical version Claptrap.

This Claptrap doesn't do too much. But it moves, it repeats phrases, and so on. Very much like other toy robots that have been around for 15 years or so. (Did you know that in Japan people are starting to have funerals for their robot dogs that are starting to die from something akin to old age?). But this Claptrap runs the android operating system, and can be programmed from your phone (even remotely). How many generations of marketing are we away from a physical avatar that comes with the game being more useful than the character within the game? So, you might need Claptrap to find water in the game, and the Claptrap in your home shows up a couple of minutes later with a mug of hot chocolate that it made from scratch because it knows its still winter outside and you'd like that better, especially if the milk is free of antibiotics?

Even better — some of you know that I'm taking Josh Price's Stata class with about half the students from this class — what will happen to you when a Dave Tufte can have his avatar learn Stata to compound Dave Tufte's productivity? Your absolute productivity will go up from learning Stata, but your relative productivity may fall further behind a Dave Tufte who leverages his avatar first.

Folks, I'm not sure how advanced such avatars will be in your lifetime. But I'm pretty sure that they're going to be independently productive enough that they're going to influence how we measure and understand productivity at the macroeconomic level.

† One of my favorite nothing to something improvements in well-being that's impossible to quantify is an app called SoundHound. I'm old enough to remember when few people knew the names of the songs they liked. That's weird, right? This was the mid to late 1970's. And there were two new technologies that were disrupting the lives of teenagers: FM radio and cassette tapes. FM allowed broadcasts in stereo with greater dynamic range (there's a reason that AM still sounds worse than FM). Cassette tapes allowed people to record cheaply and easily. The problem for the music industry was that people could tape songs off the radio. The way the music industry got around this was by encouraging a bunch of antisocial behaviors on the part of radio DJ's: fading in and fading out songs, talking over the fade in and fade out, and mentioning the artist's name but not the song's title (or discussing the lyrics too much). DJ's got kickbacks to behave this way! And listeners had to go pay real money to get all that stuff that we might now call metadata. The reason was so you'd hear this, say, really trippy song on the radio and all you'd know was that it was by the Eagles, and you'd go to the record store and there'd be a bin filled with copies of an LP called Hotel California and you'd buy the whole thing to find the one song (at $5.99 when it came out in 1976, or about $25 in today's dollars). I kid you not: to avoid this people would have tapes that they'd love, and their friends would love too, and the labels were blank because no one knew the names of all the songs. No one knew the lyrics either, which is why in 2015 sites like KissThisGuy feature misheard lyrics of mostly older songs: that site is a response to what used to be an actual problem for music listeners. And now we have SoundHound. Do you know this app? If not, you turn it on when you hear a song you like that you don't know ... and it listens to the song for a few moments, identifies the song, locates its metadata, and sends it back to your phone ... with streaming lyrics if you're into that ... and links to buy that song with a click or two.

All of this footnote may seem like so much navel-gazing. It isn't. Back then we spent money on records that got counted in GDP that no longer gets counted ... so the numbers tell us we're poorer now. But we aren't. Further, we spent ridiculous amounts of time back then looking through friend's record collections trying to glean little bits of metadata. We wanted that, so it must have been valuable, and should have been counted in GDP, but it wasn't so the numbers back then were telling us we were richer than we really were. Flash forward, and SoundHound provides us all of that in a form that's impossible to measure with GDP, so we're richer on this count too, but the numbers don't show it.

Monday, March 23, 2015


Here are a couple of bits about Singapore that are relevant for this semester.

1) We have a student, JB, who did an exchange program in Singapore. So we have inside information.

2) Lee Kuan Yew passed away this weekend. He was prime minister of Singapore from 1959 to 1990. Singapore has a one party political system, so this position was akin to CEO. He then took another position, more advisory, less day-to-day, that might be regarded as something like a corporate president, which he held until 2004. After that he took an emeritus position for 7 more years. Anyway, he's such a big deal that his obituary is on the front page of today's issue of The New York Times (they do that once or twice each month). In the annals of developing countries, Lee Kuan Yew is the ne plus ultra of effective leaders: Singapore was a developing country when he took over, and was a developed country when he stepped down. No one else has done that. If you're ever in a discussion of the success, or lack thereof, of developing countries, you'd better have him in the mix.

3) This is a good time to bring up an observation, based on a descriptive growth model (that we'll be covering over the next month), of how we explain high growth rates in some East Asian countries. This is based on an article entitled "The Myth of Asia's Miracle", written by Paul Krugman, and published in Foreign Affairs in 1994. Krugman's position is that the common understanding that growth in East Asian countries is a symptom that they are doing something special that should be emulated elsewhere ... is an urban myth.
[This conclusion is] called into question by the simple observation that the remarkable record of East Asian growth has been matched by input growth so rapid that Asian economic growth, incredibly, ceases to be a mystery. [quote corrected after class]
Essentially, Krugman is saying that other countries don't match the growth in East Asia because they do not have societies or governments that are capable of marshalling capital as readily.

Krugman's position has been widely influential. But there's also a cautionary tale. Simple simulations of descriptive growth models show that output growth derived from capital growth is fairly limited, and that developed countries are as rich as they are because they transitioned from capital growth to technological growth. He cautioned that Singapore has not matched the growth numbers of Hong Kong because it is primarily a technology importer rather than a technology innovator. This argument carries over to contemporary China: a lot of discussion in developed countries that China's growth rates are not sustainable relates back to the continued absence of technology driven increases in growth rates.

Thursday, March 19, 2015

Do Minimum Wage Increases Cause Unemployment?

Theoretically, economists usually assert that increasing a binding price floor will decrease quantity transacted.

The minimum wage is a price floor, and it is binding in many locations. So we expect to see increases in the minimum wage being associated with reduced employment or increased unemployment.

Except that this has been remarkably hard to show in the data. The seminal paper here is Card and Krueger's 1994 piece entitled "Minimum Wages and Employment: A Case Study of the Fast-Food Industry In New Jersey and Pennsylvania". Their work found little effect of minimum wage increases on unemployment.

There have been many criticisms of this paper over the last 20 years, and it has taken some hits. But don't criticize it casually: Card and Krueger are major researchers from Princeton who published this in  American Economic Review. Their results should not be taken lightly.

The latest paper in this fight is from Jeffrey Clemens and Michael Wither of UC San Diego. They looked at the three national increases in the minimum wage that occurred each July from 2007 to 2009. And they did find some negative effects.

We talk a lot in here about the decline in labor force participation over the last 15 years or so. But that includes both the employed and the unemployed. No one is saying that changes in the minimum wage have significant effects on the opportunities of the already unemployed. So they focus on a slightly different statistic: the employment/population ratio rather than the (employment+unemployment)/population ratio. This too has fallen, although it peaked a little later (in April 2000). Here's what it looks like:

So, obviously, there's a big drop off with the Great Recession. And the government was increasing the minimum wage just before the peak, just after the peak, and in the middle of that big downward slide. Roughly, we're talking about a drop from 62.9% to 58.7%.

What they found was that about 1/6 of that big drop was due to the minimum wage increase. So that's far smaller than the changes produced by the demographics of aging baby boomers, which accounts for perhaps half of that big decline. But it's not nothing: it's about 1.4 million people who lost their jobs as a result of the three increases.

That's a big deal because far fewer people work for the minimum wage than is commonly recognized. Only about half the population actually works for wages (about 76 million). The rest work for salary, commission, piece-work, and so on. Of that 76 million, only 3.3 million worked at the minimum wage in 2013. So this is saying that would have been roughly 4.7 million without those minimum wage increases.

So what we have here is that 3.3 million people got a 40% increase, and 1.4 million got a 100% decrease in their wage.

It's not an easy ethical question to address whether that was worthwhile or not.

But, here's a thought experiment. The product of the 3.3 and the 40% is 1.32, which is roughly the same as the product of 1.4 and 100%. Thus, to an approximation, an increase in the minimum wage for workers who retain their jobs is paid for by completely cutting enough people to cover the difference. So, if you worked with 5 other people, would you vote to approve firing one of them at random, and splitting their wage to give everyone left a 25% increase? I wonder.

Extended Unemployment Benefits Are Probably a Bad Idea

One of the things that was done differently in the U.S. in the Great Recession was the extension of extended unemployment benefits nationwide. Depending on your location, the Feds guaranteed you up to 47 extra weeks of unemployment. This policy was unprecedented at the national level.

The motivation for this was good-hearted: the unemployed have a tough row to hoe, and recessions aren't the fault of individuals. Of course, there was also all that hyperbole about this being The Great Recession, when really it was just ... as we showed in Case 5 of the time series portion of the class ... worse than the last two. You know just maybe using the worst thing in a generation as a justification for doing something you've never, ever, done before wasn't the best idea.

The "Depending on your location" is the key part. Because when Congress rescinded those benefits in December 2013, it provided a natural experiment for looking at the before and after effects.

Do note that in the past in this class that we have discussed the effects of extended benefits, and at the time I remarked that I didn't see that they'd made much difference because there just wasn't much evidence either way. I was wrong.

The results are now in and  ... are not pretty. Hagedorn, Manovskii and Mitman have an NBER working paper attributing 1.8 million of the 3.1 million jobs created in 2014 to rescinding those benefits. Further, they estimate that 1.0 of those 3.1 million were people who entered the work force: that is they saw people who'd been collecting benefits get them cut and go get jobs, and figured they could do the same thing too. The latter is critical because frequently critics of cutting benefits make the casual argument that if you cut benefits people just leave the labor force: the data says it's the other way around.

You can't readily disentangle the raw numbers to get at how much the unemployment rate would have changed has Congress not rescinded extended benefits. But, the unemployment rate dropped from 6.7% to 5.6% over the year, and it seems plausible that half of that was due to the benefits change.

N.B. Do note that now that we're in Spring 2015, people call the behavior of the labor market full employment, and remark that we somehow turned the corner on a weak recovery over the last year. Perhaps our members of Congress just got smarter one day in late 2013.

America: Is There a Problem with Your Rich People?

Here's two interesting bits of research.

Let me lay out some stylized facts:
  • There's a broad feeling across Democrats and Republicans that there's something wrong with America.
  • Republicans tend to view this as a combination of Democratic policies holding back the rich and encouraging the poor.
  • Democrats tend to view this as "the game is rigged in favor of the rich".
  • America has a problem with the appropriate valuation of housing.
  • America has a problem with not enough people being interested in working.
I think that's a set that most would think form a reasonable description. So what do we make of these two bits of research?

1) Adelino, Schoar and Severino have a new NBER working paper entitled "Changes in Buyer Composition and the Expansion of Credit During the Boom". The story we've been telling the last 5 years is that the financial crisis was brought on by making mortgage loans to poor people who couldn't afford the payments. This is known as decoupling (of income from ability to make payments). Adelino et al. provide evidence that this is false: decoupling is an artifact of analyzing the data by zip codes. Instead, if you look at individuals, you see no evidence of decoupling.
... There were no severe distortions towards poor or low-income people in the way banks allocated capital at loan origination.
Instead, what we see is rich people borrowing too much because the bubble looked too good to pass up:
... Lenders and borrowers bought into high house price expectations and ignored potential equilibrium effects from a large fraction of borrowers being levered to their maximum ...
The paper includes a chart showing that the portion of mortgage delinquencies among the rich increased from 7% to 32% from 2003 to 2006.

2) Robert Hall, one of the senior generation of macroeconomists on everyone's medium list for an eventual Nobel Prize reported to a Republican Congress that, now that the Great Recession is a memory, the problem in the labor market is the participation of ... the rich. Check it out:
Of course, this is showing that lack of labor force participation has been a consistent reason why the poor are poor. But what is new is that this problem is distinctly not getting worse. Instead what we are seeing is roughly a 5% decline in labor force participation across the top half of the income spectrum over the last generation. That roughly explains all of the decline in U.S. labor force participation over this period. The raw numbers are ridiculous: labor force participation by teenagers from the top quartile (the blue and green at the top) is down by 16%.
In particular, the data do not seem to support the view that the social safety net is discouraging participation ...
Note that these numbers are over the course of a generation, so we're not seeing business cycle effects (in fact it's pretty hard to see the two recessions in this chart). Further, this is participation. That includes people who are unemployed: so this isn't about a lack of jobs but rather a lack of interest in working at all.

Let me summarize: rich adults made stupid real estate purchases, and are raising kids who aren't interested in work. So, the problem with America is ... the OC.

Greece Is Not Just a Teenager, But a Difficult One At That

Remember that post from a few weeks back that everybody pays back the IMF eventually?

Catch this:
International Monetary Fund officials told their euro-area colleagues that Greece is the most unhelpful country the organization has dealt with in its 70-year history, according to two people familiar with the talks.
In a short and bad-tempered conference call on Tuesday, officials from the IMF, the European Central Bank and the European Commission complained that Greek officials aren’t adhering to a bailout extension deal reached in February or cooperating with creditors ...
This kind of points towards a conclusion that Syriza's worldview is that because they've chosen to wear white hats that this makes the other guys bad. There's a lot of window dressing on it, but at the core that may be what's going on here.

Via Marginal Revolution.