09 Nov 2013

O’Driscoll vs. Sumner

Federal Reserve, Scott Sumner 9 Comments

There is just too much good stuff to excerpt from Gerald O’Driscoll’s lead essay at Cato Unbound on central banks. OK fine, here’s just a taste:

Does the literature make an intellectual case for ending the Fed? The 19th century economic journalist and Economist editor Walter Bagehot thought it would have been better if the Bank of England had never been created. In Lombard Street, Bagehot argued that a decentralized system of many banks of approximately equal size would have been preferable. Instead of reserves being concentrated at the Bank of England, reserves in the competitive system of banking would have been dispersed among all banks.

Concentrating reserves at a central bank was the cause, not the cure, for panics. The concentration of reserves exposed the banking system to periodic panics and scrambles for liquidity. Bagehot’s famous dictum that Bank of England must lend freely at penalty interest rates in times of panic was a second-best solution to a problem caused by centralizing reserves in that institution.

Now Scott Sumner responded (in a very polite and scholarly way). You know how I often complain that Sumner is a bit slippery?

Watch this. I’m going to excerpt from Sumner’s article, then make a point about it:

I believe O’Driscoll is overly optimistic about the effectiveness of gold standard regimes. In addition, I would argue that we should avoid policy regimes that are “tamper-proof.” We don’t know what sort of policy regime is best. Therefore we should have policy regimes that are easy to alter in a situation where they appear to be causing grievous economic harm.

O’Driscoll briefly discusses the possibility of combining fiat money and free banking. He ends up concluding that this sort of regime would be too susceptible to political tampering. In my view that’s a powerful advantage of a fiat money regime. Throughout history there are many examples of rigid monetary regimes that went seriously awry and caused great damage before they collapsed. In the early 1930s prices fell sharply under an international gold standard regime. Countries did not begin recovering from the Depression until they abandoned the regime. Argentina suffered from falling prices and nominal GDP in the late 1990s and early 2000s under a rigid “tamper-proof” currency board.

In many respects the eurozone of today is an even more “tamper-proof” monetary regime than a gold standard or currency board regime. It is extremely difficult for an individual country to exit the euro without triggering a collapse of their banking system. The euro is much more than a fixed exchange rate regime. The peripheral economies of the eurozone certainly would have sharply devalued their currencies if they had been able to do so.

So then it becomes a judgment call. How bad are the future mistakes under fiat money likely to be? And how bad might things end up under a rigid monetary regime such as a gold standard? In my view the downside risks from a return to a gold standard, however constituted, are far greater than the risks of persevering with fiat money and trying to make incremental improvements.
[Bold added.]

As I hope the bolded statements above make clear, Sumner was not making offhand remarks that his proposal was less “rigid” than O’Driscoll’s. Nope, Sumner was stating it as the reason O’Driscoll was wrong (or at least one of the reasons).

In light of comments such as the ones I quoted above, the editor(s) at Cato understandably titled Sumner’s reaction essay, “In Defense of a Flexible Monetary Policy.” How could anyone possibly object to such a title, right?

Oh wait, Sumner did. On his blog linking to the exchange, Sumner writes: “My only quibble is the title they gave my essay. I favor maximum policy rigidity—the pegging of the price of a NGDP futures contract.”

My point in bringing this up, is I want you Sumner fans out there to realize I’m not merely whining when I say the guy is slippery. I’m not saying he’s a liar, just saying he is slippery. You spend months reading him, to understand his world view and why, for example, you should “never reason from a price change,” and then you’re told that a jump in stock prices proves QE is working. It’s truly why I am not following him as closely as I used to, because I don’t feel that he’s a stationary target.

9 Responses to “O’Driscoll vs. Sumner”

  1. Transformer says:

    I think Sumner sees a difference between policy flexibility and monetary flexibility. NGDPT in Sumner’s view would be a rigid policy for flexible money.

    All of the bad examples he gives have in common the fact that they prevent the money supply from adjusting to changes in demand.

  2. D. F. Linton says:

    Walking away because he’s a constantly moving target is what led Hayek not to write a critique of the General Theory. Much badness followed….

  3. Major_Freedom says:

    Re: Sumner…

    He is often inconsistent, but I believe in this case he is not. Here he is distinguishing political rigidity from monetary rigidity. He is not against monetary policy flexibility, but is in favor of political rigidity. He believes that the setting of rules politically is rigid part, and allowing the central bank to print as many dollars as is required to obey the rigid political rule is the flexible part.

    Having said that, there are quite a number of problems in the essay. He writes:

    “Throughout history there are many examples of rigid monetary regimes that went seriously awry and caused great damage before they collapsed. In the early 1930s prices fell sharply under an international gold standard regime. Countries did not begin recovering from the Depression until they abandoned the regime.”

    Murphy, this is an example of Sumner Slipperiness. He reasons from a price change. And I believe I know why. He knows that during the 1920s the aggregate money supply expanded significantly, and that prices during that time remained more or less stable as well. So by claiming that monetary policy defined in terms of prices suddenly collapsed in the early 1930s, he can end up arguing that monetary policy in general was modest until suddenly there was a sudden and significant change to monetary policy defined in terms of prices in the early 1930s. Sudden significant changes are problematic, and we are to believe that if the central bank were not “tied to gold”, then it allegedly would have been able to stop the Great Depression.

    Now, what would happen if we instead looked to aggregate money supply? Sumner knows that the money supply grew substantially during the 1920s. He would then be compelled to write that there was a significant change to monetary policy in the inflationary direction during the 1920s. He won’t be able to ignore the 1920s. In terms of money supply, it grew significantly during the 1920s and fell significantly during the 1930s. So overall to make a fair argument he would have to accompany any analyses of significant deflation during the 1930s with significant inflation during the 1920s. In other words, he won’t be able to claim that in general central banking was operating modestly “until suddenly they deflated”. He would have to ponder the possibility that maybe the reason for the significant deflation had something to do with the prior significant inflation, the effects of which would linger without a deflation. Looking at prices on the other hand enables him to tell the “All of a sudden there was a significant deflation in the 1930s!” story, which rescues central banking in his mind, and hence his intellectual investment.

  4. Major_Freedom says:

    If there was no central banking during the 1920s, then any increase in productivity would have resulted in falling prices of consumer goods, and falling prices of capital goods. It might have been the case that consumer prices, instead of rising only modestly with great monetary inflation and great productivity (as what occurred during the 1920s), they would have fallen greatly with modest inflation and great productivity. This is productivity based price deflation. Thus it makes sense that the great monetary inflation during the 1920s was accompanied by only modest price inflation. But, and this is the key point that “reasoning from a price change” market monetarists seem to be confused about: even if prices rise only modestly, it does not follow that inflation is “normal” and not causative of any major dislocutions in the capital structure of the economy as a whole, the effects of which include the potentiality and actuality of great monetary deflation.

    Market monetarism cannot seriously address why the Fed found itself having to “act” in order to reverse the effects of a sudden decline in the money supply during the 1930s. Or, another way of saying the same thing, they cannot seriously address why there was a rise in the length of time money is held before it is spent, why there was debt deflation, and why there was a decline in the money supply and hence in volume of spending and prices.

    They cannot seriously address this because they would have to find the answer in (individual) subjective value theory, including thymology. Subjective value theory and thymology are prerequisites in ascertaining the cause for why individuals suddenly did what they did during the early 1930s, the effects of which the Fed was incapable of reversing due to being “tied to gold.” This gap is of course superficially dealt with by the flawed excuse of “The Fed caused the money supply to collapse.” This is like saying an absence of cures for cancer is the cause of cancer. Well, what about the cancer itself? Why did that person get cancer, such that the absence of a cure posed an additional problem for them besides the cancer itself?

    The inability of Keynesians to seriously address this leads them conclude “animal spirits.” Minskyists to “Too much debt”. Monetarists to “monetary inflexibility”. None have the wherewithal to understand the business cycle because their economic epistemologies are holistic.

  5. Bob Roddis says:

    Mike Norman’s take on Sumner:

    Michael NormanNovember 9, 2013 at 5:21 PM

    Here’s the response: Total government spending in 2013 has been higher than 2012. The Keynesian model has not failed anyone. Please see Daily Treasury Statement for FY 2013 ending Sep 30. Here. So despite the sequester and tax increases, total spending has still been higher than 2012. I’ll just give you the numbers: FY 2012 $4.18T vs FY 2013: $4.2T.


    Just think how much better things would have been with more spending, but for those evil Austerians.

    • Andrew_FL says:

      Are those nominal numbers or after applying an inflation adjustment?

      Anyway, I wouldn’t go by any numbers for 2013 until the year is good and over. The spending estimates are pretty consistently unreliable. I’m not sure what point they couldn’t possibly think they are making, though. Very confused, very muddled…

  6. Major_Freedom says:

    “[…] Because the supply of gold grows at a relatively slow and stable rate, fluctuations in the demand for gold largely explain price level fluctuations in the period up until 1940.”

    Why the fluctuations in the demand for gold? I don’t see that question being addressed.

    “The basic problem is this; neither individuals nor banks have an incentive to adjust their demand for gold in a way that would stabilize prices or nominal GDP. Assume global real interest rates fell close to zero due to an investment bust. That would sharply increase the demand for gold. Suppose that at the same time rapid economic growth in a continent with 60% of the world’s population (and a strong cultural affinity for gold) led to soaring non-monetary demand for gold. If both of those things happened at the same time the real value of gold would soar dramatically higher. Any country with a currency fixed to gold would suffer severe deflation, as the price level is inversely related to the value of money.”

    Sumner fails to take into account the fact that humans learn and adapt to new monetary orders. Individuals and banks do indeed have an incentive to guard against recurring problems. But fluctuating NGDP is not one of them. What causes NGDP fluctuations is the problem.

    For central banking, we have adapted to the central banking monetary order by occasionally purging the economy of widespread malinvestment, i.e. “recession”. Why not adapt to it by preventing malinvestment once and for all? We need unhampered economic calculation to do this. But that is what central banking prevents. Our hampered economic calculation compels us to solve the problems in stages, culminating in a final rejection of the currency.

    For a free market money, perhaps gold, we would also adapt. However, Sumner would have us believe that in a world where gold is money, people will not seek to hold gold for longer, and thus not spend less out of existing gold balances, as a habit or as a matter of regular activity, to avoid precisely the effects he describes. He has to believe that people are stupid, and will constantly find themselves occasionally “scrambling for gold” when there are changes to real interest rates, and thus bring about recurring large scale changes to aggregate spending. Every advocate of a central plan must fall back on holding people as stupid in the precise way that the central plan is needed. Holding people as “stupid” is often communicated as “they have no incentive to do otherwise”.

    But hold on a second. If people are not stupid, then why couldn’t they hold vastly higher gold balances as a percentage of their total assets, and in so doing guard against their own potential spending (income) fluctuations caused by changes to real interest rates. This would have the effect of no sudden significant change to actual spending after all. Gold production does not need to keep up exactly with every change to the times of gold holding. Holding a large gold to asset ratio guards against the problem of the individual suddenly losing their income. Of course individuals would have an incentive to hold high cash balances in a fiat money regime, but there can’t be government guarantees of welfare. Scrap welfare, and there would be an incentive to holding far greater cash balances than we do now.

    We have speculators in the commodities markets who hold larger and smaller inventories of commodities depending on demand. We have warehouses for goods that store a supply of goods much greater than actual demand for those goods at that moment. If there is a change in demand, there is a buffer in inventory. Individuals DO have an incentive to guard against changes in demand. There is no reason they can’t guard against changes to money holding times.

    If NGDP fluctuates significantly, then this in and of itself is not a problem. The problem is behind an explanation of what caused this sudden significant change to NGDP. Any correlation of output/employment and NGDP is not caused by preventing sudden changes to NGDP via central banking. Any correlation is due to prior factors, which have as a consequence the correlation. These factors might be those that central banking negatively effects, for example economic calculation and capital structure sustainability.

    • Bob Roddis says:

      What is the necessity for having anything “nominal” in the first place? I thought that was supposed to be the “cure”, but everyone spends all of their time trying to outsmart and out run it at the expense of everyone else.

      • Major_Freedom says:

        “What is the necessity for having anything “nominal” in the first place?”

        Are you asking what is the necessity for having money? Or are you asking what is the necessity for having nominal factors in one’s economic analyses? If the latter, then from my perspective it is to answer unwarranted criticisms and confusions concerning free market money systems.

Leave a Reply