My thoughts on monetary policy

Each spring, the Hoover Institute has a conference on monetary policy.
Over at Econlog, I recently published a book review of the 2019 conference volume.

Recommended if you are looking for a change of pace from the usual political nonsense on this blog. It’s definitely higher quality than what I post here. (I wrote it several months ago, so I don’t have any discussion of the recent adoption of average inflation targeting.)

I also recommend my new Econlog post on how we should teach monetary policy.



6 Responses to “My thoughts on monetary policy”

  1. Gravatar of Michael Sandifer Michael Sandifer
    2. November 2020 at 23:35

    Seems like you’re much too easy on state of thinking on monetary policy in your review, and though you do a thoughtful job of covering the various points, I disagree with some of your analysis.

    First, I think you’re much too dismissive about the idea of using the stock market to guide monetary policy. I’m referring to statements such as

    “Clarida is somewhat dismissive of the stock market as a guide to monetary policy, however, with some justification. Stock prices move around for many different reasons, some of which are difficult to explain.”

    Almost all stock price movements are related to changes in NGDP growth expectations. It’s certainly not most days that tax or regulatory changes are adopted that disprortionately affect stock prices, though such days do occur. And these exceptions are not only rare, but pretty unimportant for our purposes here, given that divergences in the NGDP growth rate and broad stock index earnings yields indicate the degree of monetary disquilibrium clearly and precisely at any given time, as do differences in earnings growth and stock price changes.

    To illustrate, even if you don’t think e/p should equal NGDP growth rates in equilibrium, a decline in NGDP growth expectations of, say 1%, when the e/p of the S&P 500 is, say 5%, will lead to drop in the index of 16 and 2/3 percent. The calculation looks like this:

    Change in S&P 500 Index = [.05/(.05+.01)] – 1 = ~16.67%

    This means the 1% drop in NGDP growth expectations is translated into a 1% increase in the discount rate on future expected earnings, as e/p = the discount rate. To check, integrate(with t 0 to infinity) the most recent S&P 500 earnings times e^(-rt) = S&P 500 index level. My little equation above is just an algebraic shortcut.

    It may quickly become clear how well this simple math fits top of mind historical record. In the Great Depression, for example, NGDP fell roughly in half, at a time when e/p was roughly 3%.

    Change in S&P 500 Index = [.03/(.03+.5)] – 1 = ~.94.33%, which is awfully close to the decline in broad stock indexes of the time, and the S&P 500 index as constructed retroactively.

    This also means that the discount rate roughly doubled during the Great Recession, leading to a bit more than 50% decline in stock prices, which is what is observed in the data.

    This approach works well for nominal shocks, or nominal components of mixed shocks, but not well for very temporary real shocks. But, since we’re talking monetary policy, the nominal shock is all such policy can address anyway.

    While the discount rate (e/p) increases, inititally in line with negative nominal shocks, it decreases as economic recovery begins, and will actually fall below its pre-shock level if the recovery is relatively weak, meaning below the previous trend. The degree to which it is lower is the degree to which NGDP growth expectations are below the pre-shock trend. So, for example, if the pre-shock e/p was 5%, and is 4.5% during the recovery, then NGDP growth is .5% below the pre-shock trend. And again, one can also see the evidence of such monetary disequibrium in the asymmetric changes in p versus e.

    And, it turns out that since 1948, there’s only a roughly .3% difference in average annual NGDP growth and the S&P 500 e/p. Coincidence? Could be, but it’s unlikely, and this same approach seems to work really well when predicting the change in real estate values given cap values and changes in NGDP growth rates. Hence, it’s highly possible that the Fed could conduct NGDP level targeting by referring to e/p.

    You’ve stated in the past that while this might work, it could be risky, but is it riskier than inflation targeting using interest rate targeting/signaling, especially near the ZLB? Look not only at the sad results of monetary policy under the current regime, but also at the complexity of the discussion around monetary policy, most of which is due to the Rube Goldberg “concrete steppes” approach to policy transmission.

    The Fed doesn’t need to buy any assets at all. They can just create money electronically, and get it into the economy in any number of ways, including by just giving some to each actor in the economy. Monetary policy could be run by a computer that can gently adjust long-run trend growth in the money supply in future years to make up for any over or undershooting. There’s certainly no reason this shouldn’t work, at least on paper.

    Even better, the more I learn about monetary policy, the more I favor abolishing the Fed and privatizing it completely. Selgin helped win me over. Our monetary policy is a shameful mess.

  2. Gravatar of xu xu
    3. November 2020 at 07:29

    The reason Sumner requests that you go to econblog is because the platform censors the comment section. This way he can help to reinforce the Groupthink that already exists within academia.

    Academia’s love for socialism will come to an extraordinary finish today as millions of patriots vote for freedom, and take back the USA from the corrupt globalists.

    Sumner loves to talk about nationalism, but fails to understand the difference between nationalism and patriotism. Trump is a patriot! He’s a Teddy Roosevelt. He’s an Andrew Jackson. And it’s about time that we return to our core values of individualism and love for freedom and liberty. America is the last beacon of hope for the free world. Don’t let the apparatchiks take away the American Dream.

    Join the movement. Reject Sumner’s advocation for communism, restore patriotism and love, over hate and despair. Trump has promised to remove these apparatchiks from American classrooms and university administration panels. Join the likes of Glenn Loury, Thomas Sowell, and true American Patriots in thwarting Sumner and his communist agenda.

    Save Hong Kong. Save Australia. Save India. Save France. Save the Free World from the corrupt globalists. Trump 2020

  3. Gravatar of harry harry
    3. November 2020 at 07:47

    America please vote Trump.
    We are being strangled by these communists in Europe. They are even talking about regulating our food now. It never ends with these socialist creeps.

    If Trump wins, we are safe. if biden wins it will only get worse. these people are organized. they meet at their davos cocktail parties and follow in lockstep over an agenda that is anti human.

    everyone attacks trump because he’s standing up for you. not because he’s bad. please please set us free. vote trump.

  4. Gravatar of ssumner ssumner
    3. November 2020 at 10:00

    Michael, I’ve already addressed this issue in other comment sections.

  5. Gravatar of Michael Sandifer Michael Sandifer
    3. November 2020 at 10:47


    Yes, my purpose of posting this here now was to be a bit more explicit, because I think people often miss some points that I see as obvious. For example, that the inverse of the p/e ratio is also the discount rate. When I first presented these ideas, I was using p/e ratio which confused people, to my surprise. Apparently, even many in economics and finance never actually play with the numbers, perhaps seeing that the math is simple and assuming only trivial insights are to be had, at best.

    Also, in monetary equilibrium, particularly under an NGDP level targeting regime, broad stock index appreciation should match the discount rate, which should match the economic growth rate. Since the S&P 500, for example, includes 80% of the US market cap and obviously is dominated by large cap stocks, on average industries for these companies should be saturated, and hence growth depends on macroeconomic growth.

    And, I underline again the quote from you I offered above, because it is overwhelmingly the case that movements in stock prices reflect changes in NGDP growth expectations. Tax and regulatory policy changes do occur, as do flash crashes, but their relatively quite rare, and don’t represent a large problem for targeting the discount rates for broad stock indexes.

  6. Gravatar of Spencer B Hall Spencer B Hall
    4. November 2020 at 13:11

    @Michael Sandifer Re: “Almost all stock price movements are related to changes in NGDP growth expectations.”


    Re: “A permanent doubling of the money supply will lead to a permanent doubling of the price level, in the long run.”

    That’s patently false. If the preponderance of additional bank deposits reflect shifts into savings’ account classifications then inflation will fall.

    Dr. Philip George’s equations prove this. But actually he doesn’t understand the underlying accounting changes. Both of his forecasts were accurate based on his formula.

    I.e., MZM velocity will fall in response.

Leave a Reply