Alternatives to interest rate targeting

[Over at Econlog, I have a post explaining 3 MMT fallacies.]

In my recent Mercatus paper I criticized interest rate targeting and also suggested an alternative. One problem with interest rate targeting is that it doesn’t work well when nominal rates fall to zero, or slightly below zero. Fortunately, there are many possible alternatives.

One famous alternative is money supply targeting, as proposed by Milton Friedman. But even Friedman moved away from this idea late in his career. Here I’d like to focus on alternative asset price targets. After all, if you target short-term interest rate then you are also implicitly targeting the price of assets such as one-month T-bills. Here are some better options:

1. Exchange rates: This is the system that Singapore uses:

Fourth, the choice of the exchange rate as the intermediate target of monetary policy implies that MAS gives up control over domestic interest rates (and money supply). In the context of free capital movements, interest rates in Singapore are largely determined by foreign interest rates and investor expectations of the future movements in the Singapore dollar.

The downside of the policy is that it probably doesn’t work for big economies—the trading partners would object.

2. A stock price index: This is the system that Roger Farmer advocates. The downside is that stocks can move for reasons unrelated to changes in NGDP, such as a corporate tax cut that increases the share of after-tax GDP going to corporations.

3. A gold price target: This was the method used by FDR during November 1933. The problem is that it worked because markets understood that movements in gold prices were an indicator of the future path of monetary policy, once convertibility was restored (in February 1934). That’s not the case today.

4. A commodity price basket: Supply siders used to occasionally advocate targeting the price of a basket of actively traded commodities. The problem is that even a diversified basket of commodities can have a very unstable relative price.

5. The TIPS spread: First proposed in 1989 by Robert Hetzel, this idea was recently revived by John Cochrane. One problem is that there might be a time varying risk premium on TIPS spreads. Another is that it doesn’t address the Fed’s dual mandate.

6. NGDP futures targeting, guardrails version: This is my preferred policy, with no IOER and OMOs used as the policy instrument. It doesn’t really have any downsides (as you’d expect of a policy that I favor) except that central banks are reluctant to adopt such a radical policy.

7. Hybrid model/market-based policy; The central bank targets the NGDP forecast, with 50% weight given to a prediction of NGDP based on basket of asset prices and 50% weight given to a prediction of NGDP from Fed models and private forecasters. Once the non-asset price part of the forecast is established, policy targets the basket of asset prices at a level expected to produce on-target NGDP, when averaged in with the non-asset price NGDP forecast.

All these approaches should use level targeting.

For some reason, MMTers often seem to think that the Fed can’t target anything but interest rates. Singapore shows that’s not the case. In Singapore, both money and interest rates can be viewed as endogenous, as it should be.



30 Responses to “Alternatives to interest rate targeting”

  1. Gravatar of Benjamin Cole Benjamin Cole
    29. November 2020 at 16:44

    Good post.

    But central-bank policymakers need the tools to hit NGDPLT, and in a timely manner.

    Stanley Fischer suggests that central banks be given a fiscal facility.

    Maybe that is a bad idea, but then we used to have central bankers who suggested the Fed should target 1% annual deflation (Charles Plosser). Others suggested a recession would be necessary and even salubrious, as it would hold down inflation (Richard Fisher).

    What should John Q. Public believe?

  2. Gravatar of Spencer B Hall Spencer B Hall
    29. November 2020 at 16:57

    Mechanically, you target R-gDp, the nominal anchor.

    The potential output gap for R-gDp is determined by the trajectory of long-term money flows. There are different gaps relative to various long-term money targets and how fast you want to close the gap.

  3. Gravatar of Thomas Hutcheson Thomas Hutcheson
    29. November 2020 at 19:41

    What about just targeting what Congress TOLD the Fed to target: prices and employment? If the Fed thinks it can follow its mandate using interest rates, or exchange rates, or M0 as a INSTRUMENT, well they have a lot of economists to help them figure out how moving interest rates, or exchange rates or M0 will get us to 2% (or some other %) annual increase in the price level and maximum employment.

  4. Gravatar of Michael Sandifer Michael Sandifer
    30. November 2020 at 00:24

    The objection to using stock indexes doesn’t apply if you target the earnings yield of the S&P 500, for example. Trailing earnings will increase in response to corporate tax cuts to match the increased expected earnings that increase stock prices. And if you accept that the earnings yield is a nominal variable, you maintain any target you want anyway.

  5. Gravatar of William Peden William Peden
    30. November 2020 at 02:52

    Thomas Hutcheson,

    The Fed has one instrument on which it can intervene: base money, both via supply (OMOs) and demand (reserve requirements, positive/negative IOER etc.). As Jan Tinbergen clarified, hitting two targets requires two instruments:

    This is especially important in the case of the Fed’s mandate, where hitting 2% inflation will often require deviating from full employment and vice versa. Instead, the Fed needs some indicator that incorporates concern for both parts of the mandate, like NGDP levels relative to trend.

  6. Gravatar of Postkey Postkey
    30. November 2020 at 03:21

    “What about just targeting what Congress TOLD the Fed to target: prices and employment? ”

    ‘If “full employment” is anything under 5% unemployment and “price stability” is core inflation below the Fed’s 2% target rate then the Fed has achieved its dual mandate a whopping 3.5% of the time since 1957 when core inflation was first tracked.  Yes, you read that right.  THREE POINT FIVE PERCENT OF THE TIME.*  That means the Fed has failed to simultaneously achieve both its mandates 96.5% of the time.  I wouldn’t call that failure.  I’d say they’re not even trying. And maybe they’re not?’

    A ‘successful’ central bank?
    “ Werner (1992, 1997, 2005, 2011b), using Japanese data, shows that credit for GDP transactions explains nominal GDP well over several decades, while alternative explanatory variables (including interest rates and money supply) are eliminated in a reduction from a general to the parsimonious specific model.” P23.

    ‘Princes of the Yen: Central Bank Truth Documentary’

  7. Gravatar of rayward rayward
    30. November 2020 at 03:56

    I haven’t been following Roger Farmer lately, but as I understood his version of targeting, it is very similar to Sumner’s preferred version (i.e., NGDP targeting), the difference being that Farmer would include stocks (in addition to government bonds) among the assets the Fed would purchase when NGDP falls below the target. I suppose one might call that stock price targeting, but I don’t believe Farmer would. I’m a skeptic. We are already dependent on rising asset prices for prosperity, and Farmer would make it even more obvious (and make us even more dependent).

  8. Gravatar of Spencer B Hall Spencer B Hall
    30. November 2020 at 05:59

    Asset valuation prices are driven by Gresham’s law: “a statement of the least cost “principle of substitution” as applied to money: that a commodity (or service) will be devoted to those uses which are the most profitable (most widely viewed as promising), that a statement of the principle of substitution: “the bad money drives out good”.

    Credit is extended by both the banks and the nonbanks. Velocity explains the difference. Velocity is simply captured by “isolating money intended for spending, from the money held as savings”.

    Economists are stupid. The U.S. Golden Era in Capitalism was financed almost entirely by velocity, putting savings back to work.

    Today, in contrast, the economy is being financed almost entirely by money.

    I mean, why do you think stocks are falling today? The rate-of-change in money flows, volume times transactions’ velocity, is now in a downswing.

  9. Gravatar of rayward rayward
    30. November 2020 at 08:34

    One theory about Trump’s over-performance in the election is that voters were responding to the growing economy and rising stock prices under Trump’s policies up to the collapse caused by the pandemic. One trillion dollar a year budget deficits (fiscal stimulus) have that effect. Just consider the effect on economic growth and prosperity if the fiscal stimulus were targeted at investment in productive capital. I know, that’s not what interests owners of assets, rising asset prices being their main concern.

  10. Gravatar of ssumner ssumner
    30. November 2020 at 09:11

    Thomas, You asked:

    “What about just targeting what Congress TOLD the Fed to target: prices and employment?”

    I believe that NGDP targeting is the best way to do this.

    Michael, Aren’t earnings reported with a lag?

    Rayward, Did you read Farmer’s proposal?

  11. Gravatar of Ted Durant Ted Durant
    30. November 2020 at 09:25

    How is your monetary theory affected by:
    1) consumers whose purchases are made entirely by debt (loans, leases, credit card) and therefore whose only need for cash is to service debt; and
    2) the status of the U.S. dollar as the global reserve currency and U.S. Treasury bonds as the global safe asset?

  12. Gravatar of Michael Sandifer Michael Sandifer
    30. November 2020 at 09:25


    Yes, earnings are obviously reported quarterly, but tax-specific changes should be pretty easy to spot in terms of effects on prices when passed. Of course, even the announcement of potential tax cuts will have an effect on prices, when they represent news, but the same principle applies.

    Also, one can always take such effects in the context of other asset price movements, such as commodities, TIPS, bonds, etc. And, there are also plenty of forward P/E estimates out there, and adjusting those for tax changes is not necessarily the heaviest lift for analysts. I don’t think that would be necessary though.

    I envision seeing an index change temporally associated with relevant tax cuts as a situation in which you temporarily target the new earnings yield, until the next earnings report.

    The Fed doesn’t typically intervene with FFR changes all that often anyway, and such tax changes don’t come along very often. They occur, perhaps every several years, usually?

  13. Gravatar of Michael Sandifer Michael Sandifer
    30. November 2020 at 09:27

    Another idea, though this would be controversial, would be to require more frequent reporting of earnings estimates. Many companies already provide earnings guidance between quarterly reports.

  14. Gravatar of Richard A. Richard A.
    30. November 2020 at 09:40

    How about nominal GDP real time targeting? This would require getting weekly nominal GDP figures in a timely fashion much like we get the money supply figures. This would allow the FED to adjust the monetary base weekly to keep nominal GDP growth path stable.

  15. Gravatar of rayward rayward
    30. November 2020 at 10:29

    I did not read his book (Prosperity for All) but I did read his blog for a while and my impression from reading his blog was that support for stock prices (by the Fed’s purchase of stock as well as government bonds) was a result of a policy of monetary stimulus if growth fell below target. Reading the part of Cloud Yip’s interview of Farmer posted by Sumner at Econlib (2017) confirms Sumner’s description: Farmer really does propose asset (stock) price targeting. His idea is even worse that I suspected (see my comment above about the reliance on rising asset prices for prosperity). The irony is that Farmer would describe rising asset prices as prosperity for all: no it’s not, it’s prosperity for the owners of assets. Who might they be?

  16. Gravatar of ssumner ssumner
    30. November 2020 at 10:34

    Ted, My approach focuses on the supply and demand for base money. Purchases made with credit rather than cash tend to reduce the demand for base money. Other things equal that’s inflationary, with two big caveats:

    1. The hoarding demand for currency has been rising even faster than the transactions demand has been falling, hence total currency demand is rising.

    2. The Fed uses open market operations to adjust the money supply as required to offset changes in currency demand. Thus as a practical matter things like credit card purchases aren’t much of a problem.

    As far as foreign holdings of Treasuries, that could slightly impact the demand for base money, but the effect would be tiny. Its impact is bigger in other areas, such as trade balances.

    Michael, Stocks could be a part of my “hybrid” model.

    Richard, We don’t seem to be improving the timeliness of economic data, even compared to 100 years ago, despite the computer revolution. I’m not sure why.

    But yes, more timely data would help. But I’d still target NGDP a year ahead, as you don’t want to offset every little up and down in NGDP.

  17. Gravatar of Thomas Hutcheson Thomas Hutcheson
    30. November 2020 at 11:03

    @ W Peden,

    “in the case of the Fed’s mandate, where hitting 2% inflation will often require deviating from full employment and vice versa.”

    It theoretically could happen, although since at least 2008 we have not hit the price level increase target. But if you want to translate the mandate into a single number then Congress could specify the tradeoff. NDGP is one such tradeoff.

  18. Gravatar of migaglop migaglop
    30. November 2020 at 13:28

    Sumner’s objections to some MMT ill defined assumptions seem to be trivially correct (in previous posts). However, in terms of policy, why not combine recommendations from MM and MMT? In particular, targeting NGDP via central bank funded job guarantee/basic income programmes, instead of OMOs? This would be a bottom-up approach, directly benefiting the poorest spectrum of society (as opposed to driving up stock prices, whose direct beneficiares are companies and stock owners?). But I understand that legally and politically this may be a hassle 🙂

  19. Gravatar of ssumner ssumner
    30. November 2020 at 14:07

    migaglop, I don’t agree that using OMOs to engage in NGDP targeting would drive up stock prices. Lots of progressives favor NGDP targeting.

    I certainly favor a strong labor market, but I’ve never understood how the job guarantee would work. What sort of jobs would these workers do, at what pay rate? Would workers be fired if they did not do a good job? It all seems kind of vague to me. Have other countries had success with this policy?

  20. Gravatar of Ted Durant Ted Durant
    30. November 2020 at 14:49

    Scott –

    Thanks for the quick response to my earlier note!

    You wrote:
    “Thus as a practical matter things like credit card purchases aren’t much of a problem.”

    You also wrote in post 8 of your intro course:
    “[Each time I make this point a few commenters try to argue that aversion to nominal wage cuts is not irrational, because of factors like nominal debt obligations. Unfortunately this argument doesn’t work unless all of one’s expenditures are repayment of nominal debts, which is obviously not true.]”

    I would suggest that, for a large share of households in the USA, it actually is the case that the vast majority of expenditures are repayment of nominal debts. Though, I’m not sure of the implications for your theory if aversion to nominal wage cuts is rational.

    Also, while I understand that foreign holding of US Treasury debt has at most a small impact on money demand, what about the status of the US Dollar as the reserve currency for much of the world?

  21. Gravatar of Ricardo Ricardo
    30. November 2020 at 16:01

    Ted Durant: Regarding “…for a large share of households in the USA, it actually is the case that the vast majority of expenditures are repayment of nominal debts.” I doubt that’s true. In aggregate, that’s definitely not true according to Fed data for “Household debt service payments and financial obligations as a percentage of disposable personal income”:

  22. Gravatar of ssumner ssumner
    30. November 2020 at 17:22

    Ted, Interesting comment. I’d make two points:

    1. I originally was thinking of your question in terms of the effect of credit purchases on inflation, via changes in the demand for base money. The nominal debt stickiness might be an issue, but the question is how empirically important? If it is a problem, it shows up more in an extra risk of financial crises, rather than affecting the path of things like inflation.

    2. The nominal debt stickiness problem is biggest for longer term debt like fixed rate mortgages, or perhaps 7 year car loans. I’d want to see data on how the share of household expenditures on those longer term debts is changing over time. Credit card debt is often short term, although I suppose it varies from one individual to another. So I have an open mind on this issue. The data Ricardo links to seems to show a rise then a fall. The rise may have links to the 2008 banking crisis.

    In general, I view nominal wage stickiness as relating most strongly to unemployment and nominal debt stickiness as relating most strongly to financial crises.

  23. Gravatar of migaglop migaglop
    30. November 2020 at 22:50

    Scott, thanks for your reply, and you’re right – got confused with the definition of OMOs, which I mistakenly considered to include quantitative easing – which was the object of my objection. Regarding job guarantee, those reservations are valid (even though I am sympathetic with the motivation and believe it’s a matter of intelligent implementation), but what about other measures affecting consumption directly – eg. (variable) basic income, consumption checks?

  24. Gravatar of ssumner ssumner
    1. December 2020 at 10:13

    migaglop, I favor wage subsidies for low wage workers. Not a fan of UBI for the able-bodied, but wouldn’t be opposed to a modest UBI program that was not large enough to discourage work effort.

    Not sure what you mean by “consumption checks”. On the margin, we’d probably be better off favoring investment over consumption right now, especially infrastructure and housing investment.

  25. Gravatar of Spencer B Hall Spencer B Hall
    1. December 2020 at 11:29

    Targeting N-gDp means you missed the R-gDp target. There’s actually no need to “catch up” because you already know in advance whether you are missing the output gap or not.

    All these comments are out of the sides of mouths. None of them can forecast so their input is irrelevant.

  26. Gravatar of David Hansell David Hansell
    2. December 2020 at 00:25

    > 6. NGDP futures targeting … doesn’t really have any downsides (as you’d expect of a policy that I favor) except that central banks are reluctant to adopt such a radical policy.

    I may not agree with everything you say, but I can’t remember when I didn’t love the way you said it.

  27. Gravatar of ssumner ssumner
    2. December 2020 at 08:52

    Thanks David.

  28. Gravatar of Spencer B Hall Spencer B Hall
    2. December 2020 at 10:02

    The experts aren’t so good. The economy fell off a cliff beginning in July 2008. Bernanke said he couldn’t see that happening. In Ben Bernanke’s book, pg. # 56: “The Courage to Act” he opined:

    “Unfortunately, beyond a quarter or two the course of the economy is extremely hard to forecast”.

    “At the same time, because economic forecasting is far from a precise science, we have no choice but to regard all our forecasts as provisional and subject to revision as the facts demand. Thus, policy must be flexible and ready to adjust to changes in economic projections.”

    That wasn’t true.

    See also author, Dr. Richard G. Anderson: “former St. Louis Fed’s technical staff: “Although the evidence is mixed, the MSI (monetary services index), overall suggest that monetary policy WAS ACCOMMODATIVE before the financial crisis when judged in terms of liquidity.”

    There is “no fool in the shower”. Economic prognostications within a year are almost infallible. So targeting N-gDp is an anachronism.

  29. Gravatar of Spencer B Hall Spencer B Hall
    2. December 2020 at 17:19

    “We are drawn to stories that are simple to understand, and concrete — rather than abstract.”

  30. Gravatar of Spencer B Hall Spencer B Hall
    3. December 2020 at 15:20

    Economic policy is backwards. Lending by the Reserve and commercial banks is inflationary, whereas lending by the nonbanks is non-inflationary.

    Markets are about to get hit by a dose of inflation.

Leave a Reply