Why I am a fiscalist

The FT's Cardiff Garcia has a great rundown of the two factions within the pro-stimulus camp in the macro policy debate:

Fiscalists vs market monetarists is a breakout skirmish between rivalrous, unnatural allies whose common antagonists are in retreat...To recap, this debate is about the best way to accelerate the recovery and return to pre-crisis trend growth while interest rates are at the zero lower bound. (When rates are above the ZLB, many fiscalists — in particular the neo-Keynesians — are monetarists again.)

He mentions me:

I wanted to include Tyler Cowen somewhere, but his views are complicated and would require a longer discussion. Same with Noah Smith, who I think is a fiscalist but, given his broad scepticism, I wasn’t sure.

I am broadly skeptical, especially of macroeconomic theories. However, despite that skepticism, or maybe even because of it, I am a fiscalist. Let me explain why.

Here are three propositions about the macroeconomy:

1. The Monetary Stimulus Proposition: Monetary easing will raise real output and employment in a significantly depressed economy (and doing this is now worth the costs).

2. The Fiscal Stimulus Proposition: Government spending of any type will raise real output and employment in a significantly depressed economy (and that doing this is now worth the costs).

3. The Public Capital Proposition: The U.S. is currently below the optimum level of government provision of public capital (infrastructure, etc.).

I am skeptical about all three of these propositions, and I think you should be too. Macro data is not very informative, so while empirics can  be suggestive, it won't be decisive. And for that same reason, the relevant theories have not been reliably confirmed by real-world observation. We live in a world of extreme "model uncertainty". There is evidence for and against all 3 of these propositions.

But here's the thing. If either the Fiscal Stimulus Proposition or the Public Capital Proposition is true, we need to boost government spending. In fact, even someone who is not a supporter of any kind of stimulus might support fiscalist policy recommendations. For example, John Cochrane, who in early 2012 wrote:

Let's be clear what the "fiscal stimulus" argument is and is not about.  

It is not about the proposition that governments should run deficits in recessions. They should, for simple tax-smoothing, consumption-smoothing, and social-insurance reasons, just as governments should finance wars with debt. That doesn't justify all deficits -- one can still argue that our government used the recession to radically increase permanent spending. But disliking "stimulus" is not the same thing as calling for an annually balanced budget.  

Nor is it about debt financing of "infrastructure" or other genuine investments. If the project is valuable, do it. And recessions, with low interest rates and available workers, are good times to do it. That doesn't justify all "infrastructure" roads and rails to nowhere, of course... 

So while I am not as convinced of pro-fiscal-stimulus theories as people like Paul Krugman, the fiscalist-vs.-monetarist argument is about policy. And the fiscalist policy prescription seems much more robust to model uncertainty, because of the multiple reasons to want government spending.

Let's put some numbers on this. Suppose, after looking at the data, you think it's 40% likely that pure fiscal stimulus will be cost-effective. And suppose you think it's 30% likely that the U.S. needs to spend more on infrastructure. (And suppose you think these propositions are independent). Then you think there's a 58% chance that more government spending on infrastructure would be a good thing, for one or another or both of those reasons. So while if you had only considered one of the reasons for spending, you would recommend against it, but with the combination of two reasons, you would recommend it. See how that works?

In fact, what data we have seems to suggest that the public capital impact of government spending adds to the pure stimulus effect. Alan Auerbach and Yuriy Gorodnichenko find that the fiscal multiplier for government spending on investment is much higher than the multiplier for spending on consumption.

Monetarists who argue strenuously against any fiscal stimulus are therefore forced to assert that the Public Capital Proposition is false. For example, here is David Beckworth:

Okay, but not all fiscal policy is equal. Fiscal policy geared toward large government spending programs is likely to be rife with corruption, inefficient government planning, future distortionary taxes, and a ratcheting up of government intervention in the economy. So I will pass on this type of fiscal policy.

This kind of hand-waving dismissal of the need for public capital probably carries a lot of weight in certain intellectual circles, but to those who do not share Beckworth's strong priors, the argument is unconvincing in the extreme. (I must say that I have noticed what seems to be a greater reliance on theory and assumptions among monetarists than fiscalists, but I could be wrong about that.)

So count me among the fiscalists. But like most of the other fiscalists, I am definitely in favor of doing monetary stimulus at the same time. I'm just less certain that it will work.

What is "neoclassical" economics?

New rule: the term "neo-" shall never be applied to anything other than Keanu Reeves' character in The Matrix.

OK, just kidding. But it's time to talk about one of my pet reeves...er, peeves - the use of the term "neoclassical economists".

If you read econ blogs, especially blogs by "heterodox" bloggers (Austrians, Post-Keynesians, MMTers, etc.), then you know that the term "neoclassical" gets slung around quite a lot, usually as a pejorative. See here, here, here, and here for just a few examples. The idea is that "neoclassical" econ is the dominant paradigm, and that the "heterodox" schools are competing paradigms that lost out, and were, to use Kuhn's terminology, "simply read out of the profession...and subsequently ignored."

Well and good, but I have two problems with the way the term is used. First, I don't like the sloppiness of the way it's defined, and second, I don't like its application to people as opposed to ideas.

What kind of economics counts as "neoclassical"? Wikipedia defines it thus:

Neoclassical economics is a term variously used for approaches to economics focusing on the determination of prices, outputs, and income distributions in markets through supply and demand, often mediated through a hypothesized maximization of utility by income-constrained individuals and of profits by cost-constrained firms employing available information and factors of production, in accordance with rational choice theory.

OK, makes sense. Assumption of individual rationality, utility maximization, and supply/demand. One or more of things terms probably describes most of mainstream economics theory.

But does it describe most of maintstream economics research? Theory papers have declined from over half of top-journal econ papers in 1963 to less than 28% in 2011. Empirical papers make up most of the rest, with experimental economics growing to just over 8%.

How many of those empirical papers should be described as "neoclassical"? Some of them, no doubt. Some of them explicitly include neoclassical models; others test neoclassical theories developed in other papers. But many mainstream empirical papers contain no reference whatsoever to individual rationality, utility maximization, and supply/demand.

For example, take this famous paper by Acemoglu, Johnson, and Robinson, entitled "The Colonial Origins of Comparative Development: An Empirical Investigation" (American Economic Review, 2001). This paper measures the effect of institutions on growth. It does not make use of a neoclassical model. It does not test a neoclassical model. It does not include any assumption of rationality (or indeed, any model of individual behavior at all!). It does not include utility or supply/demand.

For an example from experimental econ, take "Bubbles and Experience: An Experiment", by Dufwenberg, Lindqvist, and Moore (American Economic Review, 2005). This experiment establishes conditions under which financial markets in a laboratory will result in asset price bubbles and crashes. No assumption of rationality is made, no model is referenced or tested, and no ideas of supply/demand or utility make an appearance.

These are mainstream papers, published in the most mainstream of econ journals. And there are many others like them. Does their very mainstream-ness automatically make them "neoclassical", even though they have zero of the elements that are commonly held to define neoclassical economics? If so, then I contend that the word "neoclassical" has lost all useful meaning.

"Neoclassical" should not be synonymous with "mainstream". "Neoclassical" should be used to describe a certain set of economic methods and/or ideas. Instead, "neoclassical" seems often to be used to describe anything that does not fall within a small well-known set of "heterodox" paradigms. I think that is wrong. The net effect of that type of thinking will be to block people from thinking of new ideas, because it defines any really new approach as "neoclassical". So people who want to subvert or replace econ's dominant paradigm will be shepherded toward old alternatives such as Austrianism, Post-Keynesianism, etc.

My second objection is related to the first. People who sling around the word "neoclassical" often apply it to people rather than ideas. "Oh, he's a neoclassical economist." Etc. Does that make sense? Take Daron Acemoglu for example. He writes papers that are clearly neoclassical. But he writes others that have none of the neoclassical elements. Should he be pigeonholed as a "neoclassical"? It seems obvious to me that he should not, but he probably is.

Or take me. I've never written a paper with individual optimization or supply/demand in it (though I'm working on some now). I've just done experiments and empirical stuff that didn't rely on any neoclassical idea. But people in the blogosphere have no qualms labeling me a "neoclassical", apparently because I've discussed neoclassical ideas on my blog. Not that I am mad, but it seems silly.

Shouldn't a researcher be free to work with a number of different types of analysis, and draw on a number of intellectual traditions, without getting pigeonholed? Isn't it counterproductive to scientific progress to enforce a "one-drop rule" for paradigms, so that any researcher who ever writes "max u(x)" on a legal pad is forever labeled a "neoclassical", and every paper (s)he ever writes as a "neoclassical" paper?

To reiterate, I suspect that the net effect of all this "neoclassical"-slinging is to discourage revolutions in econ. There are probably lots of revolutionary-minded young economists out there who would love to subvert the neoclassical paradigm. But when they try to find compatriots outside of the mainstream, they are told that if they don't join one of the pre-existing revolutionary groups, then they're a "neoclassical" and should go play with their fellow "neoclassicals". Which may have the effect of pushing them back into the arms of the actual neoclassicals, who of course are happy to welcome them into the fold...while the ancient "heterodox" movements retain their claim to be the only "real" revolutionaries out there...

A stable equilibrium, if you will.


Update: Lars Syll has a response to this post that perfectly illustrates the first of my two complaints:

The basic problem with [Wikipedia's] definition of neoclassical economics – basically arguing that the [defining characteristic] of neoclassical economics is its use of demand and supply, utility maximization and rational choice – is that it doesn’t get things quite right. As we all know, there is an endless list of mainstream models that more or less distance themselves from one or the other of these characteristics. So the heart of neoclassical economic theory lies elsewhere.

This is exactly the claim that "neoclassical" = "mainstream". The clear implication of Syll's syllogism is that no matter what sort of innovations mainstream economic theory embrace, no matter what old methods it discards, no matter what revolutions it undergoes, whatever it produces will be defined as "neoclassical" simply because it is in the mainstream. To me, that is clearly a counterproductive way of thinking about the world.

Should Japan default?

There are two reasons to think about what happens in the eventuality of a Japanese sovereign default. The first is that Japan's debt might be big enough, and its bond market reluctant enough, that it is forced to either default, hyperinflate, or go into severe austerity mode. In that situation, a default might be the best option. After all, after Argentina defaulted on its debt in 2001, its economy suffered for three years but then did quite well, substantially outperforming its pre-default trend:

That looks like a decently good macroeconomic scenario. And far from being an exception, this story is the norm:

So the precedent for a default is not apocalyptic. Whether this is better than hyperinflation I will leave unanswered, but it seems likely to be better than a long grinding period of austerity-induced stagnation. Also, note that austerity would redistribute wealth from Japan's young to Japan's already-comfortable older generations; a default, in contrast, represents a big transfer of wealth from the pampered old to the struggling young.

The second reason to contemplate a default is microeconomic. Observers of Japan's economy are nearly unanimous on the need for "structural reform". But Shinzo Abe's offerings on that front were extremely anemic. And given the huge edifice of special interests in Japan, and the weak political system there, we can probably expect little progress on that front. 

Structural reform is needed because Japanese productivity is stagnant. Here's a graph, from Takeo Hoshi's much-cited paper:

Hoshi attributes the stagnant TFP to "zombie" companies - companies that continue to live only through repeated infusions of below-market-rate loans. These zombies, he claims, crowd healthy, productivity-growing firms out of the market. His research with Ricardo Caballero and Anil Kashyap supports this story.

My own suspicion is that low TFP growth is also partly due to poor corporate governance in Japan. Here is a blog post I wrote about that.

A third, and related reason for low productivity growth may be the high prevalence of family businesses in Japan. There is evidence that family businesses experience slower productivity growth than non-family businesses. In this way, Japan may be similar to Portugal; I encourage everyone to read this Matt O'Brien post on family businesses and stagnation in that country.

For structural reform, Japan would need a huge blast of "creative destruction". Zombies and family businesses would need to die en masse, and healthy, independently run companies would need to emerge. The U.S. got that kind of blast in the 1980s, but Japan is unlikely to slash regulation, open up trade, and let the corporate raiders into the henhouse. The equilibrium of entrenched political interests is too strong. 

Only a big external shock is likely to be able to cause the kind of destruction needed to clear away Japan's economic ancien regime. A default would do the trick. Banks would go bamkrupt and be nationalized, and they would be forced to cut off zombies, which would then die en masse. If Japan's history is any guide, a huge burst of entrepreneurship would probably follow this die-off; witness the emergence of Sony and Honda after the shock of WW2.

So there might be some very good reasons for Japan to choose a sovereign default. But of course there would also be large costs. What would those costs be? I see three big ones: Human cost, inequality, and political risk.

The human cost could be a jump in the already sky-high suicide rate. A large number of Japanese suicides are men who lose their jobs. The close family structure of companies means that these men essentially lose access to their entire social support network. Combine this with a culture that is not very forgiving of failure, and you begin to see why a spike in unemployment might cause a large number of self-inflicted deaths.

Then again, this cost is not certain. A recovery of dynamism in Japan's economy might ultimately save more lives than it took. And human psychology is a fickle thing; it might be that in the wake of a default, unemployment might be seen as a natural disaster rather than an individual failure, and the suicide rate might even fall.

A more definite cost would be a rise in inequality. Once famed for being a middle-class society, Japan has experienced a rise in inequality over the past two decades; it is now less equal than Europe, though still more equal than the U.S. A default might change that. Family businesses might hold back productivity, but they also anchor the Japanese middle class; if large numbers of them went under, that middle class would be set adrift. 

Finally, the biggest cost of a Japanese default would be political risk. As can be seen from Argentina's example, defaults are often followed by steep drops in GDP (and rises in unemployment) that last for two or three years. That might be bad enough to destabilize Japan's already weak political institutions, and prompt the fall of the post-WW2 regime. That in turn would likely involve violence, social disruption, and increased social repression. If the winners of the coup were the "authoritarian nationalists" - basically, Shinzo Abe and his crowd - then things might not be so bad, since those guys are generally responsible and committed to a strong, stable nation. 

But if the victors were the "fanatic nationalists" - think of Toru Hashimoto and the guys in black vans - then Japan would be in for a very bad time indeed, and would quite probably revert to an unstable, violent, socially divided, repressive middle-income country like Thailand. That would be the worst possible outcome of a default.

So basically, a default would constitute a roll of the dice - a dramatic gamble that a collapse in the old order would be followed by a repeat of the kind of explosion of positive dynamism seen in the post-WW2 economic miracle or the Meiji Restoration. If the gamble failed, however, the consequence could be the end of the beautiful, peaceful, relatively free Japan that many of us have come to know and love.

The Zero Upper Bound?

A funny thing happened the other day. As part of "Abenomics", the Bank of Japan has been buying long-term Japanese government bonds. This has seemed to have the expected positive effects - Inflation is up, inflation expectations are up, growth is up, consumption is up, exports are up, and the stock market, despite a recent drop, is way way up. But here's the funny thing - Japanese long-term government bond yields kept going up over most of the last month (meaning JGB prices went down).

That's weird, right? Econ 101 says that if you buy more of something, its price should go up, not down! In this long rant, Richard Koo attributes the rise in interest rates to increased inflation expectations. According to Koo, QE doesn't work, and Japanese private investors, realizing this, started to expect inflation without real growth, and ditched JGBs, causing rates to rise.

But Nick Rowe has another explanation. According to Rowe, the rate rise was due to greater expected growth (nominal growth, so both better real growth and more inflation). As the BOJ's easy monetary policy causes the economy to improve, Rowe says, interest rates will naturally rise; investors are simply anticipating that rise, and selling bonds now. Rowe also has these rather harsh words for Koo:

Is Richard Koo really a Keynesian? Or is he just a finance guy, who doesn't really get macro/money?

Ouch!

One person who would probably agree with Rowe is Paul Krugman, who is himself a big supporter of Abenomics. In this blog post, Krugman outlined three "stories" about falling bond prices. Although none of the three stories correspond to Japan's current situation (long-term bond prices are down, stock prices are up, and the yen is down), the "stronger recovery" scenario involves rising stock prices, which we have seen. Rising stock prices indicate positive expectations for real economic growth, not just inflation. (Koo explains away this by conjecturing that Japanese stock investors and Japanese bond investors have different expectations, which I guess is not totally unreasonable given that most JGB holders are Japanese, while most of the recent inflows into the Japanese stock market have been foreign money).

(I have to say, initially I was skeptical of the Rowe/Krugman story. Most models I know would say that the recovery would take a while to raise interest rates above the pre-Abenomics level; the general-equilibrium effect would take years to overcome the partial-equilibrium effect of increased BoJ purchases holding rates down. To get a Rowe/Krugman type of rapid interest rate rise, I think you'd need a "good equilibrium, bad equilibrium" sort of model, where Abenomics kicks the economy out of the bad equilibrium very abruptly, and the economy is shocked back to a sustainably higher rate of NGDP growth. But then again, I guess I do kind of believe in that sort of model. And Japan's deflationary stagnation has lasted much longer than is typically possible in the simpler models I learned in grad school. But I digress.)

So anyway...are Japanese interest rates rising because of an incipient real economic recovery? Let us hope to Amaterasu that they are not. Let us hope that the rising yields are a blip, a trick of expectations and volatile markets and jittery bond investors.

Why do I say such a crazy thing?

Nick Rowe hints at the answer at the end of his post, when he writes:

And we can only regret that Japan did not do this many years earlier, instead of wasting all those years and letting Japan's government debt/GDP ratio climb. Because that high debt/GDP ratio is the only reason why someone might want Japan's economic recovery but not want the higher interest rates that will accompany that recovery. Which is no reason to try to stop the recovery. Though it is one additional reason to regret not having done something like Abenomics a lot earlier.

Rowe is being far too blithe. At very very high levels of debt-to-GDP, a rate rise is disastrous.

Imagine, for the sake of exaggeration, that a country had a debt-to-GDP ratio of three thousand to one. Suppose this was all in 30-year bonds, so that every year the government would have to roll over about 1/30 of its total debt, or 10,000% of GDP. Suppose that interest rates are just barely above zero - low enough to allow the government to maintain this debt burden.

Now suppose that interest rates suddenly "normalize" to 1%. Next year, the government will abruptly owe 1% of 10,000% of GDP in interest on the portion of its debt that it had to roll over. 1% of 10,000% is equal to 100%, so the government would owe all of the country's GDP in interest costs, in the first year alone. In the second year of the recovery, it would roll over another 10,000% of GDP, and thus owe 200% of GDP in interest costs!

How could it pay up? You can't tax 100% of GDP. So the government would have to borrow the rest. It seems clear that the higher the debt/GDP ratio, the less likely it would be that the private sector would be to lend to the government at an interest rate less than the economy's growth rate (the necessary condition for "stable Ponzi finance"; see discussion in comments with Nick Rowe for why this would be the case).

The only entity that would then lend the government the necessary sum is the central bank. In other words, seigniorage would be the only option to avoid default. This would either push interest rates back down to about 0%, or cause hyperinflation. Alternatively, the government could have the central bank buy outstanding bonds to push rates back down to 0%.

Now, Japan is not close to a 300,000% debt-to-GDP ratio. Its gross debt is about 240% of GDP. But a rise in interest rates would still exact a heavy burden on Japan's public finances; according to this guy, an increase in JGB yields to 2.2% would mean that 80% of Japan's current tax revenue would be eaten up by interest costs. Whether that number is correct, it's clear that with debt at 240% of GDP, Japan's growth would have to go up by a lot more than its interest rates in order to avoid a big rise in interest costs.

Also, realize that higher interest costs could easily start to hurt an economy long before they reach 100% of GDP, or even 100% of tax revenue. Why? Because of fiscal Keynesian effects. If fiscal policy affects demand (as most Keynesians and neo-Keynesians believe it does), then raising taxes to pay higher interest costs would stall the economy, as would drastic cuts in transfers or government purchases. (Of course you could borrow to pay the increased interest costs, as mentioned earlier.) So if an incipient recovery quickly causes higher rates, higher interest costs could kill the recovery.

(Now of course, all this time, increased nominal growth would erode the debt-to-GDP ratio. But that takes a long time to work. And Japan, which runs big primary deficits, is probably going to see its debt-to-GDP ratio continue to climb even if a recovery comes.)

So if monetary expansion can only cause the kind of recovery where interest rates rise, Japan is in deep shiitake. Japan's only hope is to cause the kind of recovery where interest rates stay very low for a very long time. If Japan is living in a Nick Rowe type world, that will prove impossible, and Japan's only options will be stagnation, default or hyperinflation. We should pray with all our might that we are living in a Richard Koo world instead, and that there is an economic policy that will allow Japan to boost growth and inflation while keeping interest rates low.

Anyway, this whole exercise raises the possibility that very high debt-to-GDP ratios could act as a long-term growth trap. We often talk about the "zero lower bound" on nominal rates, but very high debt-to-GDP ratios mean that there is also a zero upper bound. If recoveries always cause rates to rise (as Rowe contends), then high-debt-to-GDP ratios force governments to allow their economies to stagnate forever (or default/hyperinflate). In that case, the much-maligned Reinhart and Rogoff would be right.

If you get into a high government debt situation, a long periods of very low interest rates and robust real growth is really your only hope for a clean escape.


Update: Paul Krugman weighs in, saying my concerns are unwarranted. I've been thinking in terms of nominal rates this whole time, but Paul says that what really matters are real quantities; if the real interest rate stays low, it's all good.

Update 2: The more I think about it, the more certain I am that this post (of mine) confused and obfuscated more than it clarified. I think there was a good point somewhere in here, but I failed to make it. Oh well. That happens. For a simpler and (in my opinion) better take on the matter, see this brief post by Brad DeLong.

Update 3: Nick Rowe has a follow-up post in this series that is also much better than mine. This is basically the point I was trying to make, but stated much more cleanly. Excerpt:

Let's assume the worst-case scenario. Let's assume that I am right and Paul [Krugman] is wrong, so r increases [when the economy recovers]. And let's also assume that r increases more than g, so (r-g) increases. (Or (i-n) increases, if you prefer.) So economic recovery, by assumption, makes it harder for Japan to service the debt. [<-- at="" blockquote="" get="" i="" in="" my="" nbsp="" post.="" scenario="" the="" this="" to="" trying="" was="">

Let us also assume that Japan is like an OverLapping Genererations model, where Ricardian Equivalence is false, and where the equilibrium level of (r-g) is an increasing function of the debt/NGDP ratio. 

These worst case assumptions mean that there must exist some maximum Debt/NGDP ratio, call it Rmax, such that if the actual debt/NGDP ratio exceeds Rmax, then it would be impossible for Japan to service its debt if economic recovery causes interest rates to rise. Japan would either have to default, or create a big enough unanticipated rise in the price level to inflate away the old debt and bring the debt/NGDP ratio back down below Rmax. 

Let us further assume that Japan's current debt/NGDP ratio exceeds Rmax. 

In other words, I have deliberately set up a case in which Richard Koo would be right (maybe for the wrong reasons, but let that pass). I have deliberately made worst-case assumptions so that the higher interest rates caused by loosening monetary policy creating economic recovery would cause Japan to default on its debt, either literally or via very high inflation. 

Does this mean that "Japan cannot afford recovery"?

No. It means that Japan is already dead. It just doesn't know it yet... 

If Japan is already past the point of no return, then recovery will mean default. But delaying recovery will simply mean an even bigger default.

Now I feel even more ashamed for writing a sucky post. But at least I can link to similar posts that do not suck.

What is "derp"? The answer is technical.

There has been much discussion lately concerning the word "derp" and its appropriate usage. For example, Josh Barro used the word to describe conservative bigmouth Erick Erickson, and Paul Krugman used it as well. This prompted a primer on the history of the term, followed elsewhere by the usual hand-wringing by self-appointed cultural policemen annoyed by the word.

Now, I myself have used the word "derp" quite a lot. Possibly more than any other pundit I know, with the exception of Dave Weigel. But in any case, not only do I consider myself an expert in the use of "derp", I also have a very precise idea of what "derp" means, and how it should be used. I think "derp" is incredibly useful as a term for an important concept for which the English language has no other word.

It has to do with Bayesian probability.

Bayesian probability basically says that "probability" is, to some degree, subjective. It's your best guess for how likely something is. But to be Bayesian, your "best guess" must take the observable evidence into account. Updating your beliefs by looking at the outside world is called "Bayesian inference". Your initial guess about the probability is called your "prior belief", or just your "prior" for short. Your final guess, after you look at the evidence, is called your "posterior." The observable evidence is what changes your prior into your posterior.

How much does the evidence change your belief? That depends on three things. It depends on A) how different the evidence is from your prior, B) how strong the evidence is, and C) how strong your prior is.

What does it mean for a prior to be "strong"? It means you really, really believe something to be true. If your start off with a very strong prior, even solid evidence to the contrary won't change your mind. In other words, your posterior will come directly from your prior. (And where do priors come from? On this, Bayesian theory is silent. Let's assume they come directly from your...um...posterior.)

There are many people who have very strong priors about things. For example, there are people who believe, very strongly, that solar power will never be cost-efficient. If you confront them with evidence of solar's rapid price declines, they will continue to insist that, despite this evidence, solar will simply never be cost-competitive with fossil fuels. That they continue to insist this does not necessarily make them irrational in the Bayesian sense; they simply have very strong priors. Someday they may be convinced - for example, if and when unsubsidized solar power starts being adopted on a mass scale. It'll just take a LOT to convince them. (A more entertaining example can be seen in this classic comedy video.)

But here's the thing: When those people keep broadcasting their priors to the world again and again after every new piece of evidence comes out, it gets very annoying. After every article comes out about a new solar technology breakthrough, or a new cost drop, they'll just repeat "Solar will never be cost-competitive." That is unhelpful and uninformative, since they're just restating their priors over and over. Thus, it is annoying. Guys, we know what you think already.

English has no word for "the constant, repetitive reiteration of strong priors". Yet it is a well-known phenomenon in the world of punditry, debate, and public affairs. On Twitter, we call it "derp".

So "derp" is a unique and useful English word. Let's keep using it.

(Also, the verb associated with "derp" is "herp". It describes the action of coughing a large sticky mass of derp onto the internet in front of you. For example, to use it in a sentence: "That twerp just herped a flerp of derp!" A "flerp" is a unit I made up. It is the amount of derp that can be herped by one twerp. See?)

Atlantic column: "Should we trust economists?"

Excerpts from my latest Atlantic column:

Imagine you are the Royal Physician in England some time during the 14th century. The prince is sick, and you've been summoned to help. You call in two experts for advice. The first says: "Use leeches to suck out the evil humors." The second says "No, you must bleed him to get the evil humors out." Just to make sure, you summon a third expert all the way from Austria, who says “No, the disease is God’s punishment for the prince’s sins, you should let it run its course.” After the Austrian expert is duly led off to the dungeons and beheaded, you’re still left with the question of whether to treat the prince with leeches or bleeding. They start to argue, insulting each other in nasty epistles. "Leech guy is secretly working for the French!" alleges Bleeding Guy. "Bleeding Guy just wants the prince to die because the prince wanted higher taxes on the nobles!" Leech Guy fires back. 

What's the right move? Well, in an ideal world, you would go and get 999 patients who have illnesses similar to the prince's and give them all a variety of household substances, such as bread mold. Then you would take careful note of who died and use statistical analysis to figure out which household substances cured disease. Thus, you would discover penicillin and invent modern medicine. 

Sadly, this is not what you do, because a) if you proposed it, you would be led off to the dungeons and beheaded right next to the Austrian guy, b) it's the 14th century and you have no concept of the scientific method, and c) you don't really have the right tools for that experiment, anyway. Instead, it's bleeding or leeches. So you take your best guess and you pray you're right. 

The economic situation we find ourselves in today is a little bit like the example above...
If economists ever do succeed in developing formal models that work better, then we'll be able to go to them with questions (like "Should the Fed print more money?") and simply trust their expert advice. But until that day, all economists can really give us is intuition, suggestions, and ideas... 

No matter how much we might wish they were, economists are not go-to experts who know just how the world works or how to fine tune it...But they do have a lot of interesting things to say. They might help you clarify or re-evaluate your own beliefs about how the economy functions. They can also help you spot the flaws in each other's arguments. 

And in the end, you're the Royal Physician. You may not know everything, but the prince is dying, and you pick from among the "experts" you've got.

You can read the whole thing here. Unfortunately, the part about the Austrian guy was edited for length. :(

Regular blog readers will recognize material from some of my past blog posts, such as:

"What can you do with a DSGE model?"

"The swamps of DSGE despair"

"A world without macroeconomists?"

"A satisfactory philosophy of ignorance"

What does it mean to have "predicted the crisis"?

Since 2008, quite a lot of people have boldly claimed that they "predicted the crisis". Usually, the claimants use this "fact" to argue for the superiority of their economic school of thought, modeling approach, investing approach, or personal intuition. But what does it mean to have "predicted the crisis"?

First of all, there are different things that get labeled "the crisis". These include:

1. The big drop in U.S. housing prices that started in 2006-7.

2. The systemic collapse of the U.S. financial industry that began in 2008.

3. The deep recession and the long stagnation that began in late 2008.

Predicting one of these is not the same as predicting the others. It is possible, for example, to have missed the housing bubble and the finance industry collapse, but to have successfully predicted, after seeing these events happen, that a deep recession and long stagnation would be the result; this is what Marco Del Negro et al. claim to have done, and a number of pundits and commentators made informal recession predictions after housing peaked in 2006. Alternatively, it is possible to have predicted the bursting of the housing bubble without foreseeing the systemic damage that this would cause to the financial system; some economists, such as Dean Baker and Nouriel Roubini (and of course, Robert Shiller), seem to have called the bubble far in advance, as well as some writers like Bill McBride. It is also possible to have predicted the collapse of the big banks and their mortgage-backed bonds - and made money off of this - while staying agnostic about the macroeconomic consequences; this seems to have made a lot of money for investors like Steve Eisman and John Paulson. Of course, in theory it might have been possible to predict all three events.

Then there's the question of what it means to "predict" something. Here are some alternative definitions:

1. You could predict the timing of an event, e.g. when the housing bubble would burst.

2. You could predict the size or severity of an event, e.g. how much house prices would decline or how much the economy would contract in 2009.

3. You could predict the duration of an event, e.g. how long our economy would stagnate after the recession, or how long it would be before housing prices reached their pre-crash peak.

4. You could describe the particular characteristics of an event, e.g. what would cause banks to fail, or whether they would be bailed out, or whether inflation would remain subdued after the recession.

Next, there is the question of with what degree of confidence you make a prediction. Saying "this event is a conceivable possibility" is different than saying "the risk of this event is high," which is different from saying "the risk of this event has increased," which is different from saying "this event will happen."

Also, there is the question of how far in advance a prediction was made. That could be important.

Finally, there is the question of whether the prediction was made by a model or by a human. If it's a model, then there's the hope that humanity has a tool with which to predict future crisis events.

Anyway, how should we evaluate these claims? There are so many different combination of "predictions" and "crises" here that it's very difficult to lay out an explicit taxonomy of who got it "more right," and who got it "less right." As a more humble goal, we can examine a specific individual or model, and identify which events he/she/it predicted, with what degree of confidence, and when.

As an example, let's take Steve Keen.



Steve Keen, formerly a professor at the University of Western Sydney, is known for claiming more loudly and confidently than just about anyone else on the planet that he "predicted the global financial crisis". According to Keen, this should be a reason to believe his extensive critiques of neoclassical (i.e. mainstream) economics, and his suggested alternative paradigm, known as "Post-Keynesianism".

So in what way did Keen "predict the crisis"?

Searching the internet, I can find no record of an ex-ante prediction by Keen of a large-scale U.S. housing bubble. He did, however, predict an Australian housing bubble, in 2007 after the U.S. housing bubble had already begun to pop. That prediction has so far yet to materialize; Australian housing prices have not collapsed yet. As a result of this incorrect prediction, Keen lost a high-profile bet.

Did Keen predict the collapse of the U.S. finance industry (the Lehman shock and subsequent bailouts)? Not that I can find. Nor did he warn of the risk of such an industry collapse, as far as I can find.

How about the recession and stagnation? Here, Keen makes his strongest claim to have made an ex ante prediction. His argument is laid out in this paper. (Warning: as others have noted, Keen's papers are nearly unreadable.)

Much of the paper covers the history of macroeconomics as Keen sees it. Later, on page 10, we get to the part where he explains how he "predicted the crisis". Keen presents a macroeconomic model; actually, a class of macroeconomic models. Each of the models is a system of deterministic Ordinary Differential Equations describing the behavior of macroeconomic aggregates. He claims that this sort of model would allow one to realize that a crisis of the type we observed could potentially occur.

Notice, therefore, that this is not a prediction of timing. It is a prediction of the particular characteristics of a recession. And as to whether or not it is intended to be a prediction of the severity or duration of the recession...that's not clear. Keen isn't saying when a recession would happen, he's saying that his model shows what it would look like.

And what would it look like? Well, one of the models Keen presents (a "Goodwin" model, apparently from the 1960s) produces cycles of employment and output that look like this:

As you can see, these cycles are periodic, and of constant amplitude. But we know that this is not what business cycles really look like. (More complicated versions of this type of model might veer from periodicity into extreme nonlinearity and chaos, but chaotic models by definition have little to no predictive power.)

The next model he references is one of his own, produced in 1995. That model contains the possibility of something like a complete economic collapse:

My own simulations in Keen (1995) illustrated this possibility of a debt-induced collapse if the rate of interest was too high. For a low rate, a convergence to equilibrium occurred (Figure 4): 

At a higher rate, the system approached the infinite debt to output ratio equilibrium...

However, we have not observed an approach toward the infinite-debt-to-output ratio and near-total unemployment equilibrium that . Also, interest rates were still historically low when the financial crisis began. So this 1995 Keen model does not appear to describe the crisis we really had. Keen also adds:

[T]he 1995 model lacked price dynamics.

It's also noteworthy that Keen's 1995 model, like the "Goodwin model", contains plenty of periodicity, which as I mentioned is not observed in real life.

Keen then goes on to present a model that does include price dynamics. The figures he presents from that model is labeled "Schandl (2011)", indicating that it was made after the crisis and cannot therefore cannot be regarded as a prediction. Note that in that model, as presented by Keen, the economic collapse takes 40+ years to happen, and involves unemployment going to 100%:

In any case, it is clearly apparent that nowhere in this paper - or in any other paper that I can find - does Keen present a model whose output bears even a passing resemblance to the crisis we experienced in the late 2000s. (As an aside, note that many models, including a simple neoclassical Ramsey model, have equilibria in which the economy collapses completely. Building such a model is very very easy. But complete economic collapses - total and permanent cessations of economic activity - haven't yet been seen in the real world...ever.)

Therefore, we can conclude that there is no Steve Keen model that predicted the recession and long stagnation that we've experienced. And in fact, there does not seem to be any "Post-Keynesian model" whose features closely resemble the financial crises and recessions that we see in the real world.

So did Steve Keen himself warn in the early or mid 2000s of the impending possibility of an economic collapse? He claims that he did warn of an "impending global recession" in 2005 (see also here). I cannot find any actual writings by Keen from 2005, but I will take him at his word, since if he had made this up, I'm sure that his fellow Aussies would quickly tar and feather him for it. (If you have links to the 2005 prediction, please post them in the comments section.)

So Steve Keen presumably did warn in 2005 that a global recession was coming. This means that, counting his prediction of an imminent Australian crash, he has a 50% success rate. Remember that, according to Bayes' Theorem, the predictions of someone with an unconditional 50% success rate (i.e., coin flips) convey no information.

But is that his true success rate? After all, how many earlier predictions of imminent global recession has Keen made, that did not materialize? According to this website, Keen was predicting an imminent global recession as early as 1995. It was 12 or 13 years before his prediction came true; this long time lag makes the prediction a bit less impressive, since someone who in 1933 predicted a global recession - which did come, 80 years later - would nevertheless now be seen as having been "wrong". Now, 12 years is better than 80 years, of course.

Anyway, so we see that Steve Keen's prediction of the global financial crisis was considerably less impressive than his bold claims would have us believe. He does not have a model that can predict bubbles, financial collapses, or recessions. His personal warnings of doom often don't seem to materialize for over a decade...if they materialize at all. If you trust Steve Keen as an economist or as a personal prognosticator based on his 2005 warnings of imminent global recession, you may be falling victim to the common behavioral phenomenon of overconfidence. (Not that I expect this fact to give pause to many of his...um...ardent followers. Remember that pundits get more fans by displaying self-confidence than by being right!)

Of course, all this is not to say that Keen should receive zero plaudits, respect, or commendation for his 2005 warning - or, for that matter, for his 1995 and 2007 warnings. There are plenty of people out there who said that finance has nothing to do with recessions. There are plenty of people out there - including some very prominent mainstream economists - who said that big recessions couldn't happen anymore. However right Keen did or didn't get it - and even if he made his predictions just by reading old Minsky books and nodding his head in vague agreement - those mainstream people got things far less right.

Anyway, a similar exercise can be applied to any other economist, model, or pundit whom you think may have "predicted the crisis". You will obtain varying results, though my bet is that few will be as spectacular as you might hope.

In conclusion: Predictions are hard, especially about the future. Sometimes people get things right because they understand how the world works, and sometimes they get things right by luck. The idea of a brilliant Cassandra-like sage, shouting in the wilderness while everyone ignores his or her trenchant warnings, is occasionally true, but not as much as we would like to think.


(Update: Naturally, a bunch of people have been asking me: "So, Noah, blah blah blah, but who do you think predicted the crisis the best?" Well, I don't know. Back in 2002 and 2003 I was reading Dean Baker talking about a housing bubble and bank failures. And I remember believing that, and as a result not being too surprised when the crisis came. I'm fairly sure Baker also predicted that the macroeconomic knock-on effects would be severe. Nor do I recall him predicting a bunch of other crises that never happened. So from my very limited set of knowledge, I'd guess that Baker did very well as a prognosticator. But to really know, I'd have to go back and check systematically. Note also that Dean is a quite humble guy and doesn't go around thumping his chest about having "called the crisis"...)

DSGE + financial frictions = macro that works?

In my last post, I wrote:

So far, we don't seem to have gotten a heck of a lot of a return from the massive amount of intellectual capital that we have invested in making, exploring, and applying [DSGE] models. In principle, though, there's no reason why they can't be useful.

One of the areas I cited was forecasting. In addition to the studies I cited by Refet Gurkaynak, many people have criticized macro models for missing the big recession of 2008Q4-2009. For example, in this blog post, Volker Wieland and Maik Wolters demonstrate how DSGE models failed to forecast the big recession, even after the financial crisis itself had happened:

This would seem to be a problem. 

But it's worth it to note that, since the 2008 crisis, the macro profession does not seem to have dropped DSGE like a dirty dishrag. Instead, what most business cycle theorists seem to have done is simply to add financial frictions to the models. Which, after all, kind of makes sense; a financial crisis seems to have caused the big recession, and financial crises were the big obvious thing that was missing from the most popular New Keynesian DSGE models.

So, there are a lot of smart macroeconomists out there. Why are they not abandoning DSGE? Many "sociological" explanations are possible, of course - herd behavior, sunk cost fallacy, hysteresis and heterogeneous human capital (i.e. DSGE may be all they know how to do), and so on. But there's also another possibility, which is that maybe DSGE models, augmented by financial frictions, really do have promise as a technology.

This is the position taken by Marco Del Negro, Marc P. Giannoni, and Frank Schorfheide of the New York Fed. In a 2013 working paper, they demonstrate that a certain DSGE model was able to forecast the big post-crisis recession.

The model they use is a combination of two existing models: 1) the famous and popular Smets-Wouters (2007) New Keynesian model that I discussed in my last post, and 2) the "financial accelerator" model of Bernanke, Gertler, and Gilchrist (1999). They find that this hybrid financial New Keynesian model is able to predict the recession pretty well as of 2008Q3! Check out these graphs (red lines are 2008Q3 forecasts, dotted black lines are real events):

I don't know about you, but to me that looks pretty darn good!

I don't want to downplay or pooh-pooh this result. I want to see this checked carefully, of course, with some tables that quantify the model's forecasting performance, including its long-term forecasting performance. I will need more convincing, as will the macroeconomics profession and the world at large. And forecasting is, of course, not the only purpose of macro models. But this does look really good, and I think it supports my statement that "in principle, there is no reason why [DSGEs] can't be useful."

Remember, sometimes technologies take a long time to mature. People thought machine guns were a joke after they failed to help the French in the War of 1870. But after World War 1, nobody was laughing anymore.

However, I do have an observation to make. The Bernanke et al. (1999) financial-accelerator model has been around for quite a while. It was certainly around well before the 2008 crisis. And we had certainly had financial crises before, as had many other countries. Why was the Bernanke model not widely used to warn of the economic dangers of a financial crisis? Why was it not universally used for forecasting? Why are we only looking carefully at financial frictions after they blew a giant gaping hole in the world economy?

It seems to me that it must have to do with the scientific culture of macroeconomics. If macro as a whole had demanded good quantitative results from its models, then people would not have been satisfied with the pre-crisis finance-less New Keynesian models, or with the RBC models before them. They would have said "This approach might work, but it's not working yet, let's keep changing things to see what does work." Of course, some people said this, but apparently not enough. 

Instead, my guess is that many people in the macro field were probably content to use DSGE models for storytelling purposes, and had little hope that the models could ever really forecast the actual economy. With low expectations, people didn't push to improve the existing models as hard as they might have. But that is just my guess; I wasn't really around.

So to people who want to throw DSGE in the dustbin of history, I say: You might want to rethink that. But to people who view the del Negro paper as a vindication of modern macro theory, I say: Why didn't we do this back in 2007? And are we condemned to "always fight the last war"?

Update: Mark Thoma has some very good thoughts on why we didn't use this sort of model pre-2008, even though we had the chance.

Update 2: Some commenters and Twitter people have been suggesting that the authors tweaked ("calibrated") the parameters of the model in order to produce the impressive results seen above. The authors say in the paper (p. 13, section 3.1) that they did not do this; rather, they estimated the model using only data before 2008Q3. 

Which is good, because calibrating parameters to produce better forecasts is definitely something you are not supposed to do!! There is a difference between "fitting" and "pseudo-out-of-sample forecasting". The red lines seen in the picture above are labeled "forecasts". To do a "pseudo-out-of-sample forecast", you train (fit) the model using only data before 2008Q3, and then you produce a forecast and compare it with the post-2008Q3 data to see how good your forecast was. You should never fiddle with the model parameters to make the "forecast" come out better! 

From Section 3.1 of the paper it seems fairly clear that del Negro et al. did not make this mistake. But I think the authors should explain the forecasting procedure itself in greater detail in the next iteration of the working paper...just in case readers worry about this.

What can you do with a DSGE model?

When the Bank of England invited me to give a talk at their workshop on macroeconomics, I wasn't sure if they wanted me to provoke (i.e. troll) them with the kind of skeptical stuff I usually write on this blog, or to talk about my own research on artificial markets and expectations. So I did both. Now, this is a central bank event, which means secrecy prevails - so I can't tell you what the reaction was to my talk, or what other people said in theirs. But I thought I'd reproduce part of my talk in a blog post - the part where I talked about DSGE models. (In other words, the provocative part.)

"DSGE" is a loose term. It usually implies much more than dynamics, stochastics, and general equilibrium; colloquially, to be "DSGE" your model probably has to have things like infinitely far-sighted rational expectations, rapid clearing of goods markets, certain simple types of agent aggregation, etc. So when I talk about "DSGE models", I'm loosely referring to ones whose form is based on the 1982 Kydland & Prescott "RBC" model.

In recent times, of course, RBC models themselves have fallen out of favor somewhat in the mainstream business-cycle-modeling community, and have gone on to colonize other fields like asset pricing, international finance, and labor econ. As of 2013, the most "mainstream" DSGE models of the business cycle are "New Keynesian" models. The most important of these is the Smets-Wouters model, which has gained a huge amount of attention, especially from central banks, for seeming to be able to forecast the macroeconomy better than certain popular alternative approaches. If you know only one DSGE model, Smets-Wouters is the one you should know.

Anyway, my talk asked the question: "What can you do with a DSGE model?" Most people who evaluate the DSGE paradigm don't focus on this question; they either trace the historical reasons for the adoption of DSGE (the Lucas Critique, etc.), or they discuss the ways DSGE models might be improved. Instead, in my talk, I wanted to take the perspective of an alien econ prof who showed up on Earth in 2013 and tried to evaluate what human macroeconomic theorists were doing.

A DSGE model is just a tool. It's a gizmo, like a fork lift or a lithium-ion battery. The U.S. and Europe have invested an enormous amount of intellectual capital - thousands of person-years of our best and brightest minds - in creating, testing, and using these tools.

So what can you do with these tools?


1. Forecast the economy?

One thing you might want to do with a business cycle model is to forecast the business cycle. DSGE models have improved enormously in this regard. Though early RBC models were notoriously bad at forecasting, more recent, complex DSGE models have proven much better, and are now considered slightly better than vector autoregressions, and about as good as the Fed's own forecasts.

But as Rochelle Edge and Refet Gurkaynak show in their seminal 2010 paper, even the best DSGE models have very low forecasting power. Check out these tables from that paper:

These tables show the forecasting performance for the Smets-Wouters model (which, remember, is the "best in class") from 1992 through 2006. The first table is for inflation forecasts, the second is for growth forecasts. Look at the R-squared values. These numbers loosely describe the amount of the actual macroeconomic aggregate (inflation or growth) that the model was able to predict. An R-squared of 1 would mean that the forecasts were perfect. You'll notice that most of the numbers are very, very low. The Smets-Wouters model was able to predict a bit of inflation one quarter out (though the Fed's internal forecasts were much better at that horizon), and not at all after one quarter. As for growth, the DSGE model had very low forecasting power even one quarter ahead.

Now, this doesn't necessarily mean that DSGE models are sub-optimal forecasters. These things might just be very very hard to predict! Humanity may simply not have any good tools (yet) for predicting macroeconomies, just like we aren't yet able to predict earthquakes.

But there's also some evidence that we could be doing better than we are. In this 2013 paper, Gurkaynak et al. test the "forecast efficiency" of DSGE models, and find that their forecasts are not optimal forecasts. Also, they find that simple univariate AR models are often significantly better at forecasting things like inflation and GDP growth than the best available DSGE models! This is not an encouraging finding for the DSGE paradigm, since AR models are just about the simplest thing you can use.

Also, in this discussion of forecasting, remember that the deck has already been stacked in favor of DSGE models. Why? Because of publicity bias and overfitting. If DSGE models don't do well at forecasting, researchers will add features until they do better. As soon as they do well enough to look good, researchers will publicize the success. This is a perfectly appropriate thing to do, of course - it's like improving any machine until it's good enough to sell. But it means that the publicized models will have a tendency to overfit the data, meaning that their out-of-sample performance will usually be worse than their in-sample and pseudo-out-of-sample performance.

(Update: Via a commenter, here's a good survey of DSGE models' forecasting ability, including how they did in the Great Recession. See my new post for more...)

In other words, DSGE models are probably not very good as forecasting tools...yet. But they're about as good as anything else we have. And they have improved considerably compared to their early incarnations.


2. Give policy advice?

This is what DSGE models are "supposed to do" - in other words, most academics will tell you that this is the purpose of the models. Actually, a model can be perfectly good for policy advice even if it's bad at forecasting. This is because forecasts have to deal with lots of different effects and noise and stuff that's all happening simultaneously, while policy advice only requires you to understand one phenomenon in isolation.

But here's the problem: To get good policy advice, you need to know which model to use, and when. So how do you choose between the various DSGE models? After all, there's a million and one of them out there. And they're usually mutually contradictory; since they're fitted using many of the same macroeconomic time-series (e.g. U.S. post-WW2 GDP, employment, and inflation), one of them being a good model (even just in one specific situation) means the others must then not be good models.

So how do you choose which model to use to give you advice? Old methods like "moment matching", which were used to "validate" the original RBC models, are, simply put, not very helpful at all.

What about hypothesis testing? Again, not very helpful. If you make the model itself the null, then of course you'll reject it, because any model will be too simplified to explain everything that's going on in the economy. If you make the null the hypothesis that the DSGE model parameters equal zero, you'll almost always reject that null, even if the model is grossly misspecified.

In principle, I think you should use some kind of goodness-of-fit criterion, like an R-squared, using out-of-sample data and adjusted to favor parsimonious models. At the macro conferences and seminars I've attended, I haven't see people saying "Look at the out-of-sample adjusted R-squared of this model! We should use this one for policy!" Maybe they do say this, though, and I just haven't seen it. (Update: Here, some people, including Smets and Wouters, do evaluate the fit! Definitely check out this paper if you're into macro modeling.)

But anyway, there's a few more problems here. One is the lack of clearly defined scope conditions; macro theorists rarely work on the difficult problem of when to stop using one model and start using another (see next section). Another is the nonlinearity problem; most DSGE models are linearized, which makes them easier (i.e. possible) to work with, but means that their policy recommendations often don't even match the model.

(As an aside, many people say "OK, we don't know which DSGE model is right, so just combine a bunch of models, with some weights." Fine...but the weights aren't structural parameters, so by doing this you give up the supposed "structural-ness" of DSGE models, which is the main reason people use DSGE models instead of a spreadsheet in the first place.)

So to sum up, DSGE models could offer policy advice if you used an appropriate model selection criterion, and dealt carefully with a bunch of other thorny issues, AND happened to find a model that seemed to fit the data decently well under some clearly defined set of observable conditions. But I don't think we seem to be there yet.


3. Map from DSGE models to policy advice?

OK, so it's really hard to give definitive policy advice with DSGE models. Maybe you could instead use DSGE models as maps from policymakers' assumptions to policy advice? I.e., you could say "Hey, policymaker, if you believe A and B and C, then here are the implications for policies X and Y and Z." In other words, since DSGE models are internally consistent, maybe they can help tell policymakers what they themselves think can be done with regards to the macroeconomy. (Another way of saying this is that maybe we can leave model selection up to the priors of the policymaker.)

There's just one problem with this. DSGE models are highly stylized, meaning that it's often not possible even to figure out whether you buy an assumption or not.

Let me demonstrate this. Let's take a look at a DSGE model - say, Christiano, Eichenbaum, and Evans (2005). This New Keynesian model is very similar to the Smets-Wouters model mentioned above. Here is a VERY truncated list of the assumptions necessary for this model to work:

• Production consists of many intermediate goods, produced by monopolists, and one single consumption good" that is a CES combination of all the intermediate goods.
• Firms who produce the consumption good make no profits.
• Firms rent their capital in a perfectly competitive market.
• Firms hire labor in a perfectly competitive market.
• New firms cannot enter into, or exit from, markets.
• All capital is owned by households, and firms act to maximize profits (no agency problems).
• Firms can only change their prices at random times. These times are all independent of each other, and independent of anything about the firm, and independent of anything in the wider economy. (This is "Calvo pricing". The magic entity that allows some firms to change their prices is called the "Calvo Fairy").
• The wage demanded by households is also subject to Calvo pricing (i.e. it can only be changed at random times).
• Households purchase financial securities whose payoffs depend on whether the household is able to reoptimize its wage decision or not. Because they purchase these odd financial assets, all households have the same amount of of consumption and asset holdings.
• Households derive utility from the change in their consumption, not from its level ("habit formation"). Households also don't like to work.
• Households are rational, forward-looking, and utility-maximizing.

OK, I'll stop. Like I said, this is a VERY truncated list; the full list is maybe two or three times this long.

How many of these assumptions do you believe? I'm not sure that's even possible to answer. Formally, most of these are false. Some are very obviously false. The question is how good an approximation of reality they are. But how do we know that either?? Is it a good approximation of reality to say that households purchase financial securities whose payoffs depend on whether the household is able to reoptimize its wage decision or not? How would I even know? 

In principle, you could look at the micro evidence and see which of these assumptions looks kinda-sorta like real micro behavior. Some people have tried to do that with a few of the assumptions of the Smets-Wouters model; their results are not exactly encouraging. But if you tried to go ask a policymaker "Which of these things do you believe?", you'd get a blank stare.

So DSGE models don't make a clear map from assumptions to conclusions. But how about using them just to explore the robustness of models to variations in assumptions? A central bank (or the academic macro community) could make a bunch of DSGE models and compare their results, just to see how different modeling assumptions affect conclusions. In fact, that's probably what the academic macro community has been doing for the past 30 years. This seems somewhat useful to me, but there's a problem. DSGE models are not very tractable, so it's probably the case that nearly all of the modeling assumptions usable in DSGE models are poor approximations of reality. In that case, we'll be stuck searching next to the lamppost.

4. Communicate ideas?

DSGE models can definitely be used as a language in which to communicate ideas about how the economy works. But they are probably not the best such language. Simpler econ models, like OLG models, or even partial-equilibrium models, are much more flexible, and can be understood much more quickly by an interlocutor. DSGE models have a ton of moving parts, and it's generally very hard to see which assumptions end up causing which results. The better a model matches data or forecasts future data, the more moving parts it will generally have. This is called the "realism-tractability tradeoff". 

So if you only work with DSGE models, and if you try to understand everything in terms of DSGE models, you'll have a hard time communicating with other economists. I can see this being a problem in a central bank, where people need to communicate ideas very quickly in times of crisis.

So, what else would you have us do?

There are a number of alternatives that have been proposed to DSGE models. Different alternatives are generally proposed for the different purposes listed above.

For communicating ideas, the most popular alternatives are simpler, OLG-type models (which are, technically, DSGE, though not what we typically call "DSGE"!), and partial-equilibrium models (suggested by Robert Solow). I've seen some people use these at seminars, especially the OLG type, so I think this alternative may be catching on.

For forecasting, the common alternatives are "spreadsheet" type models (Chris Sims' dismissive term) that don't assume structural-ness. This is the kind of model used by the Fed (the FRB/US) and by some private forecasting firms like Macroadvisers.

Policy advice is the thorniest question, since you need your model to be structural. For this, the main alternative that has been put forth is called "agent-based modeling". I don't know too much about this, and the name is weird, because DSGE models are also agent-based. But basically what it seems to mean is to specify a set of microfoundations (behavioral rules for agents), and then do a big simulation. The big difference between this and DSGE is that with DSGE you can write down a set of equations that supposedly govern the macroeconomy, and with ABM you can't.

So are we wasting our time making all these DSGE models, or not?

My answer is: I'm not sure. So far, we don't seem to have gotten a heck of a lot of a return from the massive amount of intellectual capital that we have invested in making, exploring, and applying these models. In principle, though, there's no reason why they can't be useful. They have flaws, but not any clear "fatal flaw". They're not the only game in town, and realization of that fact seems to be slowly spreading, though cultural momentum may mean that the more recently invented alternatives (ABM) will take decades to catch up in popularity, if they ever do.

Bets do not (necessarily) reveal beliefs

Bryan Caplan is well-known for demanding that people bet on their macroeconomic beliefs and theories. The idea (which I endorse, btw) is that people don't really know much about macroeconomics, and tend to project an unwarranted sense of certitude in their ideas. Of course, Bryan is hardly alone in this belief; Alex Tabarrok famously declared that "a bet is a tax on bullshit".

But this idea, attractive as it is, is not quite true. The reason is something that I've decided to call the Fundamental Error of Risk. It's a mistake that most people make (myself often included!), and that an intro finance class spends months correcting. The mistake is looking at the risk and return of single assets instead of total portfolios. Basically, the risk of an asset - which includes a bet! - is based mainly on how that asset relates to other assets in your portfolio.

This means when people make bets, you don't necessarily know anything about what they really believe. Here is an example. A while ago I made a bet with Brad DeLong that U.S. inflation would go over 5% by 7/28/2015. Brad, who bet against inflation, gave me 50-to-1 odds. Now, if this were my only inflation-related bet, you could infer that I believe that there is a greater than 2% chance of 5% inflation between now and 7/28/2015. But you cannot infer that. In fact, my bet with Brad reveals nothing whatsoever about my inflation beliefs.

Why? Because I also made the exact opposite bet (i.e. that inflation would stay under 5%) with Patrick Chovanec, and gave him only 25-to-1 odds! In other words, I can't possibly lose money, no matter what inflation does (if pizza bets could scale perfectly, I could have executed an arbitrage, but I didn't bother; as things stand, I either break even or win 25 pizza dinner equivalents).

So we see that a bet does not reveal beliefs, because a bet is often used as a hedge. To use another example, I might bet on Sarah Palin winning the presidency, in order to partially hedge my personal sadness in that unfortunate state of the world.

Actually, if you take modern portfolio theory seriously - if you don't believe in any sort of mental accounting at all - then you'd have to look at my entire financial portfolio in order to determine what I really believe about inflation. Even had I not made the countervailing bet with Patrick, I might have been net long in nominal debt (i.e., I might have some cash in a bank account), meaning that my bet with Brad might have just been a hedge against my overall inflation risk.

This is definitely a problem that crops up in finance experiments. Experimentalists try to measure subjects' beliefs about asset price changes by asking them to make side bets about those changes. But we have to be careful to make sure that subjects can't use those side bets as a hedge against their choices in the other parts of their experiment. (You can do this by designing the payoffs such that it's optimal for people to bet on their true beliefs in the side bet, or you can randomly assign people to "prediction" and "investing" groups).

So we see that bets are not necessarily taxes on bullshit. This only becomes more apparent when we bring non-monetary payoffs into the picture. In reality, people make public bets based on all kinds of considerations other than financial gains - ego, fun, the need for posturing, or an excuse to go out for pizza. It's definitely not clear how these other payoffs interact with the monetary payoff of the bet, or with the payoffs of a person's other asset holdings (including other opportunities for ego, posturing, fun, and pizza dinner!). For example, in my case, I made the bet with Brad largely to help publicize the fact of low inflation, and to have an excuse to go to the excellent Zachary's Pizza. And I made the countervailing bet with Patrick not to hedge the risk of a $20 loss, but so I could write blog posts like this. (Also because I am highly mercurial and whimsical.)

Tyler Cowen summed it up best when he tweeted: "I say portfolios reveal beliefs, bets reveal personality traits and public posturing." Exactly.

Short blogging hiatus

In Europe for a week, won't be able to blog...