Blissful Ignorance, George Carlin, and the Damned Multiplier
Scott Sumner is the guy that really gets me to rethink my support of expansionary fiscal policy. He’s allergic to the intellectual laziness that pollutes so many conservative publications (the Wall Street Journal seems to think it can draw trendlines from sawdust). He also genuinely cares about unemployment and the weak in a way that many on the right do not.
But I think his argument today is wrong:
Lucas argued that discretionary monetary policy was undesirable. He argued that monetary stimulus was ineffective if anticipated, as it would already be factored into wages and prices. […] Suppose you wrote down “Every time unemployment rises above 7% we’ll do an unexpected monetary stimulus.” Then it would no longer be unexpected! In other words, the monetary authority could only surprise the public by doing “wild and crazy” things which people had no reason to expect. How likely is that to work?
I think wages and prices are stickier than Lucas believes, so I think even systematic monetary policies will have some real effects. But the logic of his argument applies pretty well to the monetary offset issue. The claim today seems to be “fiscal stimulus can work if it catches the Fed off guard, as with the April 2013 sequester.” Yes, but now think about this game from a “timeless perspective,” (not a 2013:2 perspective) where you don’t know the starting point. […] But it’s absurd to think we can have long run success from fiscal stimulus via a policy of Congress continually fooling the Fed. It’s hard to imagine a less nimble organization than Congress.
The excellent George Carlin once said, “Never argue with an idiot. They will only bring you down to their level and beat you with experience”. The idiocy of Congress is our best argument against Sumner. As Paul Krugman would say, Congress is run by a bunch of knaves, fools, and lunatics. Some are willfully ignorant and others willfully feign ignorance to represent the willfully ignorant. Congress is incentivized not by the success of our real economy, but by reelection, bringing to fore the importance of vested interests, primaries, and local advocacy groups.
Representatives are also highly “fragile” creatures. One bad remark can stain a campaign forever. Elections and policies are lost on sexual escapades and decided in cryptic dinners at maybe the Bombay Club. Big policies like tax reform are difficult to anticipate and citizens have barely a clue about a Congressman’s actual priors. Fair to say that the FOMC is endowed with more information than I, but also fair to say that not even a super-rational being can understand idiocy.
Of course, not all policies, P, are unpredictable. If a country bombed San Francisco, there’s a fantastic chance we’d retaliate (or at least ramp up our defense systems). Investors of Lockheed Martin would benefit not after P = “retaliate” is instituted, but after the unexpected attack itself. Because we all know Congress will fight back.
But some policies, especially economic ones where there exists a considerable amount of political debate and ignorance, are far harder to predict. Moreover, even if the optimal policy is certain (disclaimer: it’s not), we have no idea how lobbyists, political discounting, election j-curve factors, and a slew of other complex phenomenon will affect voting and hence Congressional behavior. And things are getting more unpredictable, still. There was significant controversy and volatility behind providing relief for Hurricane Sandy. This would be unheard not too long ago.
Imagine if a relatively big, but not TBTF, bank failed. There would probably be contagion effects and a minor financial crisis, something we can’t afford right now. (By the way, if we didn’t subsidize debt in the ridiculous ways we do, I would be staunchly against bank bailouts). But even if government relief is the optimal policy – and there is sufficient expert consensus on the topic – I would be very uncertain about the resultant action because I know there are very significant voices on both sides of the aisle averse to bailouts of any kind. Perhaps they have honorable intentions, like cautiousness against the socialization of risk or the virtues of a smaller government. Or maybe they need to pander to a hyper-conservative electorate. Whatever the case, I know that I’d have for support only Mr. Wall Street Chuck Schumer. I, and others, will be eager to hear the final decision.
So then we can posit that the effect of a policy P is in some capacity proportional to its informational value. Think about how we’d bet on a coin toss. If we have a certain prior (say “Bet Heads”) it would be updated to a far greater extent – over an infinite trial set – on a fair coin than one biased either way. We’re just far more surprised at the answer when probabilities are equal. Note, “betting” isn’t necessarily the right term here, because in such a situation only my prior and not the update itself matters. The concept remains.
Ludwig Boltzmann (above) formalized this notion as entropy. This idea, though designed in the context of thermodynamics, is pretty apt in abstracting the economy. It’s intimately connected with the efficient market hypothesis, which Sumner strongly defends. As Pele and Tepus put it:
This study aims to demonstrate that the extreme values of returns distribution are mostly associated with particular periods of stock markets inefficiency, when their level of uncertainty reaches a local minimum. We propose an estimator of uncertainty, through the entropy of probability density function of returns. The relationship between the level of uncertainty of a stock market and extreme values of returns distribution is illustrated trough a binary logistic regression model estimated for main indexes of four stock markets from Central and Eastern Europe.
In a rational and competitive arena like the stock market, periods of low uncertainty and hence inefficiency and rents are rare. In politics, they are not. (Beautifully noted by Thomas Jefferson in his feud with Alexander Hamilton on federal assumption of state debt. Indeed, in an age of slow information, the Northerners who learned first earned huge profits on arbitrage of bonds in the South). This is also why it’s disconcerting that Congress plans on repealing the STOCK Act. No matter, the point is informational asymmetries between Congress and even the smart members of FOMC imply that there is no “zero multiplier” on fiscal policy at the zero lower bound. This contrasts Sumner’s belief that any fiscal multiplier is only a measure of central bank incompetence, to the extent inability to predict clowns is not “incompetence”.
Sumner comments that this is irrelevant:
Ashok, I think you are missing the point of the Lucas argument. If institutions have an “effect” because they are crazy and irrational, then you don’t want to use that institution to try to stabilize the economy. It’s not about whether fiscal policy can or cannot have an effect right now, (it can) it’s about whether fiscal policy should be used at all. Lucas is making a very subtle point, and even today I think most people miss the point.
An irrational fiscal policy REGIME will make things worse, even if right this minute they might make things temporarily better, by dumb luck.
This is a respectable view. If fiscal policy sets a precedent it might wreck all the good it can do today with all kinds of disastrous policies in the future. Actually, my argument above wasn’t totally fair, Sumner does note:
Over time, there will be equal number of cases where fiscal stimulus is greater than expected, and less than expected. Thus on average fiscal stimulus will have no effect.
This depends on the scope of our argument. Are we arguing what Congress should do. Or are we arguing what rational experts should advice Congress to do. I think in both cases fiscal policy wins out – given the circumstances – but definitely so in the latter group. If Sumner accepts fiscal policy can have benefits, even “minute”, today he should advocate to that effect. (Note, this is assuming a rational, monetary offset of fiscal policy is the only argument to the contrary, i.e. the zero multiplier. There are other points made like the efficiency of a small government, preponderance of rent-seeking behavior, and general fear of higher debt. So don’t remind me of those, and in return I’ll keep my argument for fiscal policy strictly in the context of monetary offset.) Before I go on, there’s another really important point here. Congress isn’t going to get smart and rational in the near future. Actually, all chances are it won’t. Period. So by Sumner’s own logic, they’ll screw up the economy with bas fiscal policy eventually. But if he admits that there’s some “minute” benefit today, as he did, ignoring it would just yield a negative rather than zero multiplier. So even if you take his path. Fiscal policy today is a good deal.
In the first circumstance, let’s say Congress implicitly adopts a so-called Krugman Rule (KR) “Fiscal policy only at the zero-lower bound”. Note, this is a reverse implication not a biconditional for that would remove all uncertainty and has no chance of capturing reality for reasons mentioned above. How would this be statistically represented? We could imagine Congress passing fiscal policy is log-normally distributed with mean m and standard deviation s. This is probably too simplistic, but better than a normal because there’s a definite right-skew. Adoption of KR, today at the ZLB, would increase m and decrease s. That is, there is a greater chance that Congress passes a higher stimulus with a tighter bound.
But there’s still uncertainty. Because the conditionality of KR runs only one way, the market and FOMC cannot anticipate and hence render useless fiscal policy. But KR also means that bad precedents wouldn’t be set because the Fed can anticipate no fiscal policy unless interest rates hit zero.
Of course, there’s an uncertainty to KR itself. But that’s a meta-level argument. (Just like arguing the validity of the Lucas Critique, is – I just have a different initial belief than Sumner). Further, the general culture of debt aversion (except during unexpected events like invading Iraq) will probably render fiscal policy in times of expansion, beyond expected automatic stabilizers, unlikely. So an “irrational policy regime” couldn’t make things worse.
Remember, ignorance is bliss. Big bankers have far more political information then me, and know that Congress will save them if they fail. That kind of certainty corrodes systems and breeds corruption. Imagine if the dumb and clunky Congressmen were so random that even bankers didn’t know what would happen. Well, they wouldn’t have made stupidly risky bets or leveraged themselves to the heavens.
But we don’t want Congress to be a coin. That would be disastrous. The real policy debate should be on the parts of Congressional activity that remain uncertain, and the parts that don’t. Taking this to the logical conclusion, we know certain rights – enshrined the Constitution – will be guaranteed. While there are certain threats to the Bill of Rights, these are generally checked by a independent judiciary.
Here’s a parting thought experiment. Sumner might think my assumption of a Krugman Rule is too strong (because maybe Congress will be pressured to act in a non-ZLB recession, or rents, or whatever). What if this was enshrined in the Constitution with the Chicago IGM as Supreme Court. Would he, then, support fiscal policy today? More later. This was one of my favorite posts!
This is Ashok. 