Fama, Krugman, Cochrane, and all that jazz
In all seriousness I’m sick to death of this damned argument over what the accounting identity Savings=Investment means (I am sure you are too – but I get to write the blog posts and I need to vent. My apologies).
I never, for the life of me, expected such a substantial dispute to develop between economic experts on what it meant. These guys are more than smart enough to know what they are saying – which implies to me that they are hiding/twisting their value judgments to get the solution they want. Grrrrr.
Still Fama started it by stating that S=I always holds (true) and that this implies that government spending can’t influence employment (huh?). Then when the critics came raining in (they are mentioned here) he framed his view a little more – and randomly said that Brad Delong had the only potential criticism of his view. This annoyed me as Arnold Kling and Bryan Caplan had already discussed the enormous logical mis-step that he had taken – and he “answered it” by ignoring them …
Lets get back to his framing of the issue and discuss it:
Fama states:
First, however, I want to restate my argument in simple terms.
1. Bailouts and stimulus plans must be financed.
2. If the financing takes the form of additional government debt, the added debt displaces other uses of the funds.
3. Thus, stimulus plans only enhance incomes when they move resources from less productive to more productive uses.
So lets accept the first premise, then the second premise does logically “follows”, and so does the conclusion – excellent.
However, it is the interpretation of the second premise that determines how we interpret this conclusion – this is where the conclusion he made in his initial piece (that government spending can’t increase employment) failed.
Lets think about some alternative uses of the funds in a STATIC setting to start with:
- Hold it and do nothing,
- Spend it on goods and services,
- Save it at an an institution that lends it out (or uses it) at some risk.
So at a given point in time this is the business that the funds could be up to. Now, the more people that are hoarding cash, or keeping it in liquid accounts, or putting it in government treasuries, the less “funds” there are going around the economy at a point in time – and given fixed prices during a “static period” this leads to lower activity.
In a dynamic sense this is related to the idea of the velocity of money (and also the ratio of bank reserves) – the money stock is moving around the economy more slowly when people are hiding some under the pillow, reducing the money supply. Given sticky prices this leads to a reduction in activity. During this whole process S=I -> but that doesn’t stop output dropping …
Now if government can attract funds from people that are holding them doing nothing (which they definitely can at the moment!) or if it does have viable public goods it can build then government does have “more productive” alternatives for the funds than the private sector does.
This is the beauty of the John Cochrane article (ht Dani Rodrik) – he effectively states all these things working from the same prior model as Fama.
Now Paul Krugman and other commentators that were supporting him also criticised Fama’s piece – however, they also attacked John Cochrane. Sure Cochrane’s language was too strong in support of the S=I identity in his piece, and if you only read the first few pages it would seem like his is missing the point about this simply being an accounting relationship (especially given his discussion stating one dollar more government spending always leads to one dollar less private spending – something he eventually contradicts). However, after I read his piece I was under the impression that he was someone who was against fiscal stimulus as a matter of ideology – but could see scope for a stimulus under the aforementioned conditions.
By ignoring attacking the piece solely based on the a supposed misunderstanding of S=I the proponents of government spending are able to ignore his other conclusions – such as that if there is a credit constraint surely government loaning funds to private firms would at least as effective as (non-public good) government investment.
More fundamentally, the general method of Cochrane appears to be to look for “market failures” and figure out how government policy may be able to solve them. In my opinion this is definitely a preferable way to look at macro policy than arbitary stimulus packages aiming to “fill an output gap” …
Garbage in = garbage out
That’s the ruling identity in macro today.
All of these people are working with garbage — the garbage intellectual tradition of Keynes.
QED.
If the demand for money increases due to hording, prices would fall, however, you hold prices fixed. On the other hand, during your “static” period, you do not hold output fixed as well. I guess if you assume away the problem you can get the answer you desire.
Fama is absolutely correct in your indented quote. the 2 flaws with the S=I is the idea that all $ are equal and that a government can indefinitely create money to provide stimulus. Money created is confidence. Simply a promise to pay. Whilst a govt can provide stimulus by creating money it cannot do that indefinitely. Just ask yourself whether you would buy Iceland treasury bonds right now.
You are also correct that what is critical is velocity of money. If the govt allows itself to be used as a credit clearing house by borrowing from bank A and lending to bank B to lend to company C rather than Bank A lending to Bank B direct then the govt will have increased the velocity of money by taking on Bank B credit risk itself. This seems to be what is happening at present.
If the government chooses to create money it reduces the value of its promise to repay. Investors have realised that the UK government promise to repay has devalued substantially and the value of currency has fallen as people who have a choice put their funds elsewhere.
What Fama seems to ignore in making S = I is dead time in money cycling through saver, bank and government. In the long run the equation holds.
“All of these people are working with garbage — the garbage intellectual tradition of Keynes.”
Hmmm, I have no problem with the intellectual tradition of Keynes – his prescriptions were just applied too broadly.
“If the demand for money increases due to hording, prices would fall, however, you hold prices fixed. On the other hand, during your “static” period, you do not hold output fixed as well. I guess if you assume away the problem you can get the answer you desire.”
Indeed, I completely agree. However, it is an empirical fact that prices are substantially more sticky than output – which is why this result holds …
“What Fama seems to ignore in making S = I is dead time in money cycling through saver, bank and government. In the long run the equation holds.”
Even in the short-run S=I holds – however, that does not tell us about the outcome for output, contrary to what Fama stated.
Otherwise I think you hit the nail on the head in your comment Phil 😉
“Indeed, I completely agree. However, it is an empirical fact that prices are substantially more sticky than output – which is why this result holds …”
I would like to see references that show it to be an empirical fact. It is, however, a Keynesian assumption that prices are sticky. Keynes assumed one way stickiness: prices were sticky going down but not going up. Any microeconomic foundation for this assumption?
Have you bought gasoline recently?
“I would like to see references that show it to be an empirical fact. It is, however, a Keynesian assumption that prices are sticky”
Fair call, I will have a little hunt through some papers tonight 🙂
“Keynes assumed one way stickiness: prices were sticky going down but not going up. Any microeconomic foundation for this assumption?”
Well we can make “microfoundations” for any assumption we want – just pull together some game theory and make use of the folk theorem.
Why might prices be downwards rigid? Well if there are a large number of industries which are oligopolistic by nature, then a drop in demand may not initially lead to a read in price – as staying at the initial price helps to support tacit collusion between the firms. As a result, industries where prices can be painted as strategic complements provides scope for downward rigidity.
And don’t forget the fact that assumptions could come from bounded rationality and rule following. It might be a good rule to not accept a cut in your money wage when negotiating given imperfect information – however, in the rare case of a recession this rule fails …
Personally I think it is just as easy to make arguments for upward rigidity – after all if we have an increase in demand and prices are strategic complements then there is a “positive externality” from lifting prices, and firms will not do enough of it. I guess this shows how much the implied frame of reference matters.
“Have you bought gasoline recently?”
Some prices are more flexible than others, sure. As a result, we know that there will be problems with the allocation of resources – however this isn’t what we are trying to sort out. If some of the inflexible prices are important enough (eg the price of labour) then it can have substantial repercussions.
Furthermore note what “flexibility” means – it doesn’t even mean that prices physically can’t be changed. It can also be the case that strategic effects influence the price signal – as a result prices could be inflexible but still move around – it isn’t all a result of menu costs.
As an aside, lets think of the entire issue another way.
Unemployment is rising, and will be well above its natural rate by early this year. Labour is a factor of production – and for some reason avaliable labour units are not being used at the current market price. We have a massive market failure in the labour market. Obviously the wage is too high – or else we wouldn’t have the unemployment.
In such a case we are evidently not in a world were prices clear – and there is the potential for specific government actions that could improve outcomes. We have to keep in mind where the finance is coming from, yes. We have to make sure that government doesn’t remain in the middle of the labour market, and once prices clear that government puts the labour back into the private sector, yes.
However, not using a huge labour resource when the private sector is filling the coffers by purchasing treasury bonds is wasteful. Trying to say that there is a a priori case against a fiscal stimulus appears wrong-headed. Arguing that this is NOT the time for a fiscal stimulus is fine – I can more than accept that argument. But there is no argument that convincingly states that the government cannot influence employment – which is what Fama appeared to be stating.
Here is a paper on sticky price models: http://www.minneapolisfed.org/research/QR/QR2711.pdf
Bottom line: “The findings reject the joint hypothesis that the sticky-price models typically employed in policy analysis capture the U.S. economy and that commonly used monetary policy shocks represent exogenous shifts.”
“Labour is a factor of production – and for some reason avaliable labour units are not being used at the current market price. We have a massive market failure in the labour market. Obviously the wage is too high – or else we wouldn’t have the unemployment.”
May I suggest that there is structural unemployment. Low interest rate policy of the Federal Reserve (relative to say what the interest rate would have been if the Fed had followed the Taylor rule) has distorted the structure of the economy.
For example, the housing boom in large measure generated by low interest rates and the Government Sponsered Entities of Freddie Mack and Fannie Mae has coming crashing down. The sub-prime mortgage crisis has spilled in to the financial sector as well. The financial sector is by the way one of the most regulated sectors in the US economy.
The structure of the economy has been distorted by low interest rate policy, as was the case with the dot com boom in 2001 that also came crashing down, and now as the economy attempts to readjust to an equilibrium, unemployment is inevitably generated. The market reajustment is hardly to blame for boom/bust cycle created by the monetary authorities.
Also, government can influence employment, but Fama’s argument as I recall was about output. Can the government increase output of the economy? The proponent of the stimulus have said that government spending has a multiplier of 1.57. This is beyond belief. Also, Fama has argued that government spending is justified if the spending the government spending is more productive than the private spending that it displaces. In some cases this may be true, but how about having a cost benefit analysis of each of the projects proposed in the bill to see which ones are justified on this basis?
Hi Tom, thanks for all the good comments.
“Here is a paper on sticky price models: http://www.minneapolisfed.org/research/QR/QR2711.pdf”
Very interesting. At a micro level sticky prices are also rejected for many large scale areas like supermarkets.
However, as I said earlier the concept of “stickiness” in this case has to do with whether the price moves at all – not whether strategic effects cause the price to deviate from the optimal level (note the test is for observed levels of non-movement in prices, assuming monopolistic competition – which is non-strategic). We only need small strategic effects to cause a significant deviate from where prices should be.
“May I suggest that there is structural unemployment”
Most definitely – and in most cases I would completely agree with you. However, we are in a situation where employment is being lost across the board – where all industries are being hammered. Structural employment would indicate that there are some sectors that need employees but that the skills don’t exist – but that is not what the labour market figures or the survey figures are saying.
In an extreme case such as this I don’t think a structural shock by itself is enough.
“Also, government can influence employment, but Fama’s argument as I recall was about output”
Actually he says:
“Even when there are lots of idle workers, government bailouts and stimulus plans are not likely to add to employment”
And also I imagine there are circumstances where it can influence “output” quite easily. The real question (which is the hardest to answer for pro-stimulus types) is will it increase welfare – will society be happier. We can make a case for that – but even in this extreme situation the case is tight.
“Also, Fama has argued that government spending is justified if the spending the government spending is more productive than the private spending that it displaces. In some cases this may be true, but how about having a cost benefit analysis of each of the projects proposed in the bill to see which ones are justified on this basis?”
Agreed – we should do cost benefit analysis. However, everyone and their dog has suggested that government spending could be more productive – this argument is like saying the 3>2. However, the fact is that there is some debate surround WHAT it displaces.
Does it displace cash sitting under a mattress or does it displace efficient industry?
Fama’s suggestion that S=I implies that short term employment during a period of deficient demand cannot be influenced by government is wrong. However, his framework (the three points he puts forward) is right – as long as we can accept that there is a debate surrounding what we think is displaced …