Monday, December 28, 2015

Production and the Black Box

In economics (and other fields) we use the metaphor “black box” to describe a process where we see the inputs and the outputs, but we don’t really understand (or maybe really need to know) what goes on inside.

In the coconut plantation story (or lemonade stand, or apple orchard) we tell to motivate students to think about how production works, we invoke that black box metaphor: labor and capital go in, and output (or real GDP) comes out. When we add a production function, what we’re trying to do is put mathematics on what goes on inside the black box that’s both tractable and economically sensible.

With any luck, this cartoon will make the black box idea stickier for you:

WuMo - Black Box Production

I do want to emphasize that the block box doesn’t necessarily mean better inputs produce better outputs (although most of us believe that to be true most of the time in the real world). In this case, we kept one input the same (labor), changed the other input, and with (what looks like the the exact same) black box produced the same amounts of different outputs. But it could just have easily been different combinations of the same inputs. What’s important is that we don’t need to know more about this to get the point.

Friday, December 18, 2015

Why Is Macro So Hard? The Bridge's Weight Limit Problem

We've all seen signs like this:
Bridge Weight Limit
There's an old joke about how engineers figure out these limits: they build a bridge, drive successively larger vehicles over it until it collapses, and then they rebuild the bridge and put up the sign.

Of course, in reality, do engineers try to do better than that? Well, of course they do. I'm not implying that they don't, just that we should expect some failures.

Unfortunately, most policy applied at the macroeconomic level is non-experimental. You can't run a controlled experiment to see if it will work or not. Instead, you just do it, and reflect later on whether it worked the way you thought it would. If we're lucky, the decision-makers may even update their thinking. So most of macroeconomic policy is analogous to the bridge's weight limit problem.

This came up when New Orleans was hit by Hurricane Katrina: the levees held ... until they didn't hold any more. Then we cleaned up and pointed fingers. (For readers who have never seen the levees, this was one of the biggest federal infrastructure projects in history).

And yet, Charles Marohn writing at Strong Towns notes that almost the exact same excuse was used by former Federal Reserve Chair Ben Bernanke in describing the policy response to the Great Recession:
My favorite part of that interview was Bernanke reacting to himself from a 2005 MSNB interview where he said housing problems were localized and would not impact the national economy. Here's what he said:
I absolutely – first things you said, by the way, when saying in 2005 and 2006 the economy was going to continue to do well, it did do well.  2007 was not a bad year until the end.  So, the economy was doing OK in a broad sense.  What we missed, what we didn’t anticipate, was that the decline in house prices and the problems in mortgages would generate this huge panic. So that — you know, I can’t, I can’t deny that.  I think that I wouldn’t give us a particularly good grade before the fall of 2007.  After that, when we began to see what was going on, there, we were much more aggressive in responding.
In other words, when he said things were doing well, they were doing well, until everything fell apart, and then they reacted.
Now, let's change direction a bit. Returning to the bridge joke, what do you think get's budgeted for? Building one bridge, or building two? The Katrina case should make it clear to everyone that ultimately we pay twice (or more). But that's not the way we start out. Instead, we budget one version, and then we add to that later on.

So, here's another quote from former San Francisco mayor Willie Brown:
Referring to huge cost overruns during the construction of San Francisco’s four-and-a-half-billion-dollar Transbay Transit Center, Brown wrote, “We always knew the initial estimate was way under the real cost…. If people knew the real cost from the start, nothing would ever be approved. The idea is to get going. Start digging a hole and make it so big, there’s no alternative to coming up with the money to fill it in.” [italics are from the source article]
In terms of macroeconomics, this means that you don't let on how big your policy is going to ultimately be. Instead, you start small, and add to it. Then if there's a failure, you use that to double-down and expand the program.

Oxford business professor Bent Flyvbjerg† dubs this "survival of the un-fittest":
... the least deserving projects get built precisely because their cost-benefit estimates are so misleadingly optimistic. [italics are original]
So now we've got three related issues:
  • We don't really know whether some things will work until they fail.
  • We don't like to pay enough for something that won't fail, so we start with something cheap and add on to it.
  • We fib to ourselves about the cost of doing things right, so we end up choosing what seems cheapest (because we didn't include the opportunity costs of doing it right or doing it twice).
Is there an example of how this works in macroeconomics? Well, how about Social Security, or even better the program that was added to it in 1965: Medicare.

Of course, on the first point, Social Security and/or Medicare haven't failed. But, there have been ominous rumblings for a few decades now.

On the second point, we started with social security in 1935. This offered a government check to former workers, aged 65 and older. Life expectancy for a 65 year old at the time was 2 years. Many things have been added to our government's social security offerings: coverage for spouses and dependent children in 1939, coverage for disabled workers in 1954, early retirement in 1961 (even though people were living longer), coverage for the disabled who weren't workers (SSI) in 1972, and automatic cost-of-living adjustments in 1977. But by far the biggest one is Medicare. Initially, in 1965, this was just two parts (A and B, for hospital stays, and more general insurance). But we got Part C in 1997 (with more broader and more flexible coverage), and Part D in 2006 (with prescription coverage). Now, here's a little secret. "Obamacare" was pitched to the public as being about patient protection and affordable care. But everyone who follows policy knows that a huge part of Obamacare was new fixes to Medicare to rein in costs.

That covers the start small and make additions part of the argument. What about the idea that costs are understated because no one would agree to them?

Well, Medicare is a liability of the federal government (a liability being something that generates cash outflows rather than inflows). The money for those checks needs to come from somewhere. If not, it's called an unfunded liability. Current estimates are that the unfunded liabilities of the Medicare system are about $36T (that's the number that Obamacare reduced, by perhaps $10T). That's twice the size of the "official" national debt. That unfunded liability is not included in the national debt because it hasn't been borrowed ... yet. This sort of like carrying a $1,000 balance on a credit card, and knowing that you're going to have to put a $2,000 car repair on the card, yet still thinking that your debt is only $1,000. Well, technically it is, but this is clearly magical thinking.

Then there's the third point: did we choose to structure the laws this way because we were in denial about some of the costs? I don't know of any direct evidence to support this, but the indirect evidence is how we treat increasing life expectancy. It's no secret that life expectancy has been increasing in the United States for, well, not really decades but in fact a couple of centuries now. And yet 80 years into our era with Social Security, we still do not have any concrete plan for matching up when people can start to collect benefits with how long they can be expected generate costs to the system.

Maybe macro isn't that hard. Maybe what's hard is admitting to this pattern of behavior.

† I swear I did not make that name up.

BTW: Thanks to Cold Spring Shops to pointing me to this source article, from which I'm getting the previous entry in this series too.

Why Is Macro So Hard? Orderly but Dumb or Chaotic but Smart? Choose one.

Here's a quote from Charles Marohn, writing at Strong Towns:
Projects coming from the top down tend to be orderly but dumb while projects coming from the bottom up tend to be chaotic but smart. We all prefer smart to dumb, but we Americans have a really strong preference – and have established systems that enable bureaucrats and elected officials to ensure – that we get orderly over chaotic, even when it means accepting dumb as a result. [emphasis is original]
This goes some way towards explaining things like the "Obama stimulus packages" of 2009. A new government responded in a timely fashion for a representative democracy (over a period of several months), with a large series of nice sounding programs (that roughly match up with the textbook Keynesian prescription for a bad recession), wrapped up in a few large multi-purpose bills.

That's the orderly part.

But mixed in there were some clunkers like ... well ... the "cash for clunkers" program. You know, let's destroy some productive capital because it has relatively larger negative externalities that are still minute when compared to the stream of beneficial services delivered by the capital.

BTW: Thanks to Cold Spring Shops to pointing me to this source article, from which I'm getting the next entry in this series too.