Arnold Kling thinks he disagrees with me . . .

. . . but he’s wrong.  In a recent post I made this argument:

I don’t like using any kind of interest rate as an indicator of Fed policy.  But if you insist (as most economists seem to) why not at least use ex ante real interest rates?  Then we can very easily explain the crisis this way:

1.  Between July and November 2008 the Fed adopted an ultra-tight monetary policy.

2.  Real interest rates on indexed 5 year T-bonds rose from 0.5% to 4.2%

3.  The tight money caused NGDP growth to plunge from its usual 5% to negative 3%.

4.  The AS/AD model predicts you’d get about 1% inflation and negative 4% real growth.

5.  That’s what we got, therefore modern macro nicely explains the recent recession

Kling responded by quoting the 5 points (but not the short preceding paragraph.)  He then responded:

According to Sumner, we know that the Fed tightened because we know that real interest rates on TIPS rose. The real rate is the nominal rate minus the expected rate of inflation. Because the nominal rate did not rise, the increase in the real rate came from a big drop in expected inflation. So we know that the Fed tightened because expectation inflation dropped, and we know that expected inflation dropped because the Fed tightened. That’s too circular for my taste.

I’m also not convinced that we can read expected inflation in the TIPS market. Take right now, for instance. The TIPS market is saying that inflation is going to be low. But is that what the people who are buying gold believe? I don’t think so.

But of course that’s not what I said; I think real interest rates are less bad than nominal rates, but still a poor indicator of monetary policy.  It’s the rest of the profession that seems to like them.  If we must use some form of real rates, I’d prefer Beckworth’s suggestion that we take the difference between nominal rates and NGDP growth, but I don’t much like that one either.  I’d rather not discuss interest rates at all.

Since Kling raised this issue, however, I’d like to clear up one misconception.  Kling argues that in order to figure out the ex ante real interest rate, we first must estimate the expected rate of inflation.  That is often done by looking at TIPS spreads.  But in fact that is not necessary at all.  The ex ante real yield on (some off the run) TIPS is known with (near) certainty, and is in no way contingent on dicey estimates of expected inflation.  Sure, if we wanted to know the ex ante real rate on non-indexed T-bonds, we’d need to estimate the expected rate of inflation.  But why would we be interested in that real rate?  Suppose it was estimated at 1%, while TIPS had a real yield of 4.5%.  All that would prove is that a mad rush for liquidity had driven the real yields on nominal bonds to absurdly low levels.  It would make nominal bonds meaningless as a monetary policy indicator.  TIPS are less liquid, just as almost every other bond in the universe is less liquid that nominal Treasury securities.

If TIPS yielded 4.5%, no banker in his right mind would use estimated real yields on conventional T-bonds as an opportunity cost of funds.  If you could get a risk free real rate of 4.5% on a TIPS, you’d need to get an extremely high rate on (much riskier) loans to the private sector to make lending worthwhile.  Especially loans occurring against the backdrop of a collapsing economy.  No one has yet explained to me why the TIPS yield, which is an ex ante real yield, is not the minimum opportunity cost of any funds lent by banks in late November 2008.

Of course after November real rates fell sharply, just as you’d expect when corporate investment was tanking.  This is another reason why real rates are not a useful monetary policy indicator.

Tags: , ,


11 Responses to “Arnold Kling thinks he disagrees with me . . .”

  1. Gravatar of Nick Rowe Nick Rowe
    2. December 2009 at 18:17

    Wandering off-topic a little: why doesn’t the Fed buy TIPS?

  2. Gravatar of David Stinson David Stinson
    2. December 2009 at 18:54

    Nick said:

    “Wandering off-topic a little: why doesn’t the Fed buy TIPS?”

    Perhaps TIPS are not sufficiently liquid? That, at least, is the implication of previous posts and comments here. Which raises the question, I suppose, as to why TIPS are not sufficiently liquid. The answer, I guess, may be that the US gov’t doesn’t create a big enough market. One speculates that the reason for that is that the US government might prefer it if someone else bears the inflation risk.

  3. Gravatar of Josh Josh
    2. December 2009 at 21:32


    The phrase that I latched onto was Kling’s assertion that “we know that the Fed tightened…” I know that we differ somewhat on the minutiae, but the Fed doesn’t have to change their policy instrument for their policy to become tight. In fact, I made the case in a post today:

    The Fed’s policy became tight in September (I know that you believe it was even earlier than this) without the Fed tightening in the traditional sense of adjusting the federal funds rate. (Not to be outdone, they did of course start paying interest on reserves.) Nonetheless, I think that this is the point that is separating others from accepting this view. They see that the monetary base doubled and they cannot imagine that this not inflationary.

    What’s more, this point is entirely consistent with Friedman and Schwartz in The Great Contraction — the Fed allowed broader measures of the money supply to fall and failed to increase the base to prevent such declines. I do not maintain the view that these broad measures fell in the recent circumstance, but given the decline in nominal income the increases in the base were clearly not sufficient to maintain a “monetary equilibrium” in which the Fed offsets increases in the demand for money (reductions in the money multiplier and velocity) by increasing the supply of base money.

    (Note: I apologize to those who are not familiar with my views as this comment might seem rather opaque. I would refer those of you to the link above.)

  4. Gravatar of Scott Sumner Scott Sumner
    3. December 2009 at 06:08

    Nick and David, I don’t know. But aren’t the bid-asked spreads in TIPS tiny? I thought I had seen data to that effect. Yes TIPS are relatively small, but in absolute terms are still pretty big. (at least well up in the $100s of billions.)

    Josh, That’s a great point that I overlooked. By suggesting the the Fed hadn’t tightened because nominal rates hadn’t increased, Kling was making the same mistake Joan Robinson made when she said easy money couldn’t have caused the German hyperinflation because nominal rates weren’t low.

  5. Gravatar of Joe Joe
    3. December 2009 at 10:42

    How do you square TIPS with Gold? Does gold have better marketing, or are the TIPS measure of CPI skewed, and large number of market players think Gold is a better predictor of inflation? Maybe gold is a better prediction of worldwide inflation, people worried that the solution of cooperative money printing will come about (I think Eichengreen had thsi idea in a Vox EU paper).

  6. Gravatar of ssumner ssumner
    3. December 2009 at 12:38

    Joe, There was a recent article that gold discoveries are tailing off, and demand from India and China is rising fast. Gold is traded in a world market, it is not a good indicator of US inflation, or Japanese inflation, or Chinese inflation.

  7. Gravatar of Doc Merlin Doc Merlin
    3. December 2009 at 14:23

    TIPS and T-bills are also traded on a world market. But anyway….

    I don’t think the skyrocketing gold price is supply driven.
    Supply of *new* gold has dropped off a little bit, but not that much. It went from 2,518 metric tons in 2000 to 2,469 metric tons in 2007. The output in 2008 fell by around 3% to about 2,400 metric tons. but producer hedging also fell, considerably, so the supply of new gold is still quite high relative to total gold production.
    Also, only about 161,000 tons have been mined in the history of the world and almost all of that is still around because gold is almost perfectly durable. So, roughly 1.5% of all gold ever mined was mined last year.

    This leads me to think that the gold price rise isn’t supply driven. As production of new gold has been *almost* flat since 2001, and the old gold is still around.
    We have other three other possibilities:

    1. If its driven by demand for gold then we have to ask why is gold demand so high, particularly when jewelry sales to individuals are way down, but people are selling lots of old jewelry which is being melted into bullion. I think that, possibly, instead of thinking of gold as a commodity its better to think of it as a reserve currency.
    And since it expands so steadily and regularly, gold can be a good proxy for the difference between a Selginesk natural rate of inflation/deflation and the actual rate for the wold taken as a whole.

    2. The transaction costs of buying gold are down due to Gold ETFs.
    This may be part of the effect, but I doubt it can be all of it. Its not THAT much easier to use an ETF instead of a future, although the transaction costs for an ETF are a lot lower than for small amounts of physical gold, its not hard to buy physical gold at a discount. Anyway, ETFs like GLD are really neat they act almost like a 100% reserve currency in gold.

    3. The only other possibility is that gold is an asset bubble caused by easy money.

    My attempt to reconcile 3 with your idea that money is ‘tight.’:

    If we have tight money then then precious metals are not a bubble.

    One possibility that looks like ‘tight money’ to your measurements, but would still create asset bubbles would be ‘loose money’ with all the extra looseness sucked off into assets. This would make the ‘looseness’ irrelevant and make it act just like tight money only with more asset bubbles.
    If this is true, that means that the Fed has lost all ability to control the value of the USD, and is now completely impotent.

  8. Gravatar of Scott Sumner Scott Sumner
    3. December 2009 at 17:41

    Doc Merlin, Even if supply was the cause of the higher gold price, the numbers you provide would not be expected to impact gold prices. This is because efficient markets priced in the data you provided a long time ago. Prices move on new information. So it must be the recent articles I have been reading predicting a sharp drop-off in future gold production. Again, past data from 2008 would have no impact on 2009 gold prices.

    My hunch is that it is a combination of fast rising Asian demand, some Westerners who have been reading Bob Murphy’s blog, and sharply falling expected supply.

  9. Gravatar of Josh Josh
    3. December 2009 at 20:35


    I don’t know how much of the increase in the price of gold is actually reflecting inflation expectations. First, there is a flow supply (gold mining) and an existing (stock) supply of gold. The flow supply of gold has slowed since around 2001 when gold mining output peaked. The stock supply has also stopped increasing as central banks have by and large stopped selling gold and, given the financial crisis, actually started buying gold in some cases. Meanwhile, the demand for gold as an investment and as jewelry has increased in China and India over the past few years. In addition, newly created gold ETFs have increased the demand for gold because they are required to purchase and store gold.

    Given that other indicators of inflation expectations are quite low and the significant changes in the gold market, I am more inclined to believe that movements in the gold price have been caused by changes in the gold market than from macroeconomic factors.

    I just wrote a post on this today:

  10. Gravatar of Doc Merlin Doc Merlin
    3. December 2009 at 21:26

    Scott if you were right about it being supply issues that only recently came to light, it wouldn’t be almost every metal commodity that is doing this sort of behavior over the last 18 months.

    over the past 18 months almost every single metal has done this (platinum being an exception).
    JCC (copper etf) is up well over 100% since the copper bottom.
    GLD (gold etf) is up about 50% since the gold bottom in 2008.
    SLV (silver etf) is up around 100% since the silver bottom.
    JJU (Aluminum etf) is up about 40% since the aluminum bottom in early 2009.

    Other indexes have also climbed:
    NYSE has climbed 45%
    OIL (crude oil etf, yes I know its not a metal) has climbed about 40% since its bottom, even though crude stores are massive right now, and we have major over-supply that is being stored in warehouses.

    The main difference is that gold has been in a secular bull market for about as long as the dollar has been in a secular bear market.

    Not every commodity has done so well: Food commodities are way down, this is probably due to the biofuels bubble finally bursting. Its gotten so bad that gasoline production in the US will face legal awkwardness this year. The law requires use of a certain amount of ethanol, but the law also limits the amount of ethanol that is legally addable to gasoline to 10%. This year there isn’t enough gasoline production to fulfill the mandate.
    Platinum also hasn’t faired so well, but this is probably due to two factors. New technologies replacing it with cheaper materials in industrial uses and the platinum mining companies actively discouraging speculation that involves actual buying metal, for example refusing to sell to ETFs.

    So, to summarize: Gold’s rising over the last year is commensurate with other metals and quite a few indexes, therefore it is unlikely that new forward looking supply information in 2009 is what is making it rise so much. This is another reason, I think you are wrong and that my above analysis is correct.

  11. Gravatar of ssumner ssumner
    4. December 2009 at 07:30

    Josh, I agree. And again, recent news stories point to even lower gold production in future years.

    Doc Merlin, Actually, the fact that all metals are soaring supports my explanation, not the gold bugs. I think the boom in Asia is dramtically increasing the demand for all metals, not just gold. If the gold bugs were right you’d see gold and silver prices rising, but not other (base) metals.

    I will say this, I think the fears of deflation last March have receded, and that has helped both gold and equities. But I still think the expected rate of inflation is quite low. I expect 1-2% CPI inflation next year. I think the markets also expect that sort of inflation.

    It is possible that markets expect higher long term inflation because of fiscal deficits (though I doubt it.) But if this is true, the best way to lower those long term inflationary expectations is with an easy money policy today, which will reduce the debt/GDP ratio.

Leave a Reply