This Time Was Not Different

Revising notes on recession prediction for my course, I find the 10yr-3mo spread has preceded 7 of the last 8 recessions (8 if the 14 bps near inversion of mid-1989). And back in June 2019, recession seemed not unlikely.

Figure 1: Ten year minus three month Treasury spread, % (constant maturity) (dark blue), ten year minus two year (red). NBER defined recession dates shaded gray; assumes last trough in April. Source: Federal Reserve Board via FRED, NBER, and author’s calculations.

Of course, there is a false positive if you go back to 1965…

Using the most plain-vanilla of specifications, a probit regression on the 12 month lagged 10yr-30mo spread estimated over 1960M01-2019M01period  (for recession indicator variable going up to 2020M01), I obtain the following recession probabilities up to 2020M09:

Figure 2: Probability of recession for indicated month. NBER defined recession dates shaded gray. Source: NBER, author’s calculations.

Of course, had the pandemic not occurred, maybe 2019 might’ve been a false positive…

Interestingly, using the term premium adjusted spread, the probability of recession peaks around 4%, much as in this post.

Course slides on recession prediction, here.

32 thoughts on “This Time Was Not Different

  1. Moses Herzog

    I registered (for the La Follette panel on trade) !!!! Socially/emotionally stunted Menzie Chinn fanboy reporting for duty. : )

    Is this the part where I have to click on 3 tractor pictures to prove I’m not a bot??

  2. Not Trampis

    come on Chinny,
    The inverted yield curve had nothing to do with this recession unless it was in league with covid

      1. Not Trampis


        an inverted yield curve usually works when it is the Central Banbk pushing up the short end of the curve not when the longer end falls

        1. Moses Herzog

          @ Not Trampis
          That may or may not be true, but I don’t think it conclusively tells us if there would have been a recession sans Covid-19. I would argue rates tend to lower when people are looking for safe places to park their money/assets, that’s not totally related to central bank moves. The central bank is usually reacting to the market and/or economy, not the other way around. Market expectations is what drives this, not QE or the discount window.

        2. Moses Herzog

          @ Not Trampis
          Menzie can also correct me here if I’m wrong, but I think the last time the fed raised rates was 2018. And even the oddball open market operations they have done since late 2018 were to lower rates, not raise them, so how “the fed pushed up the short end of the curve” to create the inversions of 2019 and 2020, I’m at a loss to see. I’m happy to hear that one explained though. Happy to soak up and absorb some of that great Aussie education system.

    1. Willie

      Covid made the downturn worse, but it was about to happen or was already happening before Covid.

      I am feeling deja vu while typing this.

  3. pgl

    Trump health alert:

    Trump continued to careen from topic to topic on Sean Hannity’s show, promoting conspiracy theories, peddling falsehoods, and launching attacks on Democrats and the media during a roughly 25 minute conversation. But it was arguably the President’s health that took center stage. During the interview, Hannity twice asked Trump if he had been tested for the coronavirus since he became ill last week. It was a question that the President apparently couldn’t answer. Trump instead said that he will “probably” be tested on Friday. Medhi Hasan quipped about the non-answer, “Mark the date, we have reached the point where even Sean Hannity is asking Trump simple questions that the president can’t or won’t answer.” It wasn’t just Trump’s comments on his health that drew attention. It was what the audience could hear at home. At least twice during the interview, Trump had to pause his sentences and audibly clear his throat and cough. The President insisted he was feeling great, claiming he saw his doctors earlier in the day and that they believe him to be in “great shape.” But he sure didn’t sound like he was cured, as he claimed to be.

    And yet he wants to hold rallies starting tomorrow evening. If his doctor OKs this – it is gross malpractice.

    1. John Smith

      “At least twice during the interview, Trump had to pause his sentences and audibly clear his throat and cough”

      You’re being silly. Even people who are entirely well have to pause and clear their throat from time to time, especially when they’re talking a lot.

      1. noneconomist

        Maybe someone who’s got the virus, who has received (by injection!) an “experimental cocktail “, an experimental anti virus medication, steroids, and oxygen multiple times SHOULDN’T” be talking “ a lot”? Or much at all?

  4. Wally

    The slowdown began before the Covid outbreak. You can go back to December, 2019, financial articles to see that. Of course the Covid slowdown swamped all else (except for the rich getting richer). In my mind the yield curve indicator gets a ‘win’. Put an asterisk on it if you want but the cold fact is that a recession followed as indicated.

    1. John Smith

      You’re being silly. The statistics depend entirely on how you collect them. In the US if you have COVID when you die then you’re a statistic. Even if you die in a car wreck or commit suicide, or have six other deadly illnesses. But would some poor guy dying in a remote part of India even be tested after death?
      As for Europe, the keep using insane lockdowns to try and stop the disease, despite increasing evidence that lockdowns cause MORE deaths than the illness does.
      The WHO now puts the infection fatality rate of COVID-19 at just 0.16% – about the same as seasonal flu.

      1. Menzie Chinn Post author

        John Smith: (or should I say Obergruppenfuhrer John Smith?) For the love of g**, provide a URL when you write:

        As for Europe, the keep using insane lockdowns to try and stop the disease, despite increasing evidence that lockdowns cause MORE deaths than the illness does.

        1. pgl

          Forgive him as Kelly Anne Conway sent him the text but omitted any links. After all – she is ill with the virus.

      2. pgl

        Second time you started a comment with “You’re being silly”. Of course both of your comments were really dumb. I guess we are forewarned the next time you start a troll comment with “You’re being silly”.

      3. 2slugbaits

        John Smith The WHO now puts the infection fatality rate of COVID-19 at just 0.16% – about the same as seasonal flu.

        You’re confusing case fatality rates (CFR) and infection fatality rates (IFR). The CFR for influenza obviously depends upon the specific influenza…the Spanish flu is not the same as last year’s seasonal flu. But with respect to recent influenzas over the last 10 years, the average has been around 0.13%. But that’s a CFR of 0.13%. If you convert that to the IFR for seasonal influenza it comes to something like 0.065%. As to COVID-19, in July the CDC revised it’s estimate of the IFR from from 0.26% to 0.65%. A meta-analysis of many studies estimates 0.68%. On the other hand, the CFR for COVID is pretty straightforward and works out to about 2.8%, which is down from the 5% range it had been at. Bottom line, COVID is around ten times deadlier than ordinary seasonal flu. It’s also a lot more contagious.

      4. macroduck

        So, the “John Smith” guy is apparently trying out “you’re being silly” as a method of dismissing ideas that his (or her – “John Smith” is obviously an attempt to rally faux-patriotic feeling) master don’t like.

        Dear “John Smith”, we have enough bouht-and-paid-for trolls. You should try peddling your rhetorical tricks elsewhere.

  5. ltr

    October 8, 2020



    Cases ( 7,833,763)
    Deaths ( 217,738)


    Cases ( 6,903,812)
    Deaths ( 106,521)


    Cases ( 799,188)
    Deaths ( 82,726)


    Cases ( 671,638)
    Deaths ( 32,521)


    Cases ( 561,815)
    Deaths ( 42,592)


    Cases ( 315,514)
    Deaths ( 9,667)


    Cases ( 175,559)
    Deaths ( 9,557)


    Cases ( 85,500)
    Deaths ( 4,634)

  6. ltr

    October 8, 2020

    Coronavirus   (Deaths per million)

    US   ( 657)
    Mexico   ( 640)
    UK   ( 625)
    France   ( 498)

    Canada   ( 253)
    Germany   ( 115)
    India   ( 77)
    China   ( 3)

    Notice the ratios of deaths to coronavirus cases are 10.4%, 7.6% and 4.8% for Mexico, the United Kingdom and France respectively.

  7. ltr

    October 9, 2020

    Chinese mainland reports 21 new COVID-19 cases

    The Chinese mainland on Thursday registered 21 new COVID-19 cases, all from overseas, the National Health Commission announced on Friday.

    This is the 54th consecutive day with zero domestic transmissions reported on the Chinese mainland. No deaths related to the coronavirus were reported over the previous 24 hours, while 15 patients were discharged from hospitals.

    The COVID-19 tally on the Chinese mainland stands at 85,521 infections and 4,634 fatalities, while 357 asymptomatic patients remain under medical observation.

    Chinese mainland new imported cases

    Chinese mainland new asymptomatic cases

    [ There has been no coronavirus death on the Chinese mainland since May 17. There has been no community or domestic coronavirus case for 54 days. Since June began there have been only 2 limited community clusters of infections, in Beijing and Urumqi in Xinjiang, both of which were contained with mass testing, contact tracing and quarantine, and both outbreaks ended in a few weeks. Imported coronavirus cases are caught at entry points with required testing and immediate quarantine. Asymptomatic cases are all quarantined.

    The flow of imported cases to China is low, but has been persistent. There are as a result 206 active imported coronavirus cases on the Chinese mainland, but of which only 2 cases are classed as serious or critical. ]

  8. pgl

    It seems our new troll John Smith is a Mike Pence Minnie Me. Let me explain by noting something about the VP debate.

    Every time Harris said something that was true that Pence could not handle he would smugly fire back “that’s your opinion”. Meanwhile Pence lied repeatedly but his words must be considered gospel.

    So John Smith starts each comment with “you’re being silly” followed by bat$shit insanity that he thinks is gospel. MAGA!

    1. macroduck

      While we are talking about how to distinguish between really good work and intellectual prostitutes, we got Louise Gluck (with an umlaut) and the World Food Organization. Where’s the really important Nobel?

      1. Moses Herzog

        @ macroduck
        From a surface glance both of those seem like good choices to me. The Economics prize will be announced very early in the morning, around 5:00am local time USA on the 12th, which I believe will be this Monday. I was trying to find rumors on the top 5 candidates wandering around online, but seems to be a void of info right now.

      2. Moses Herzog

        You know, you have to think another female choice is highly probable in this cultural time phase. Although her work is not highly technical or elaborately mathematical, I have to wonder if someone like Mariana Mazzucato would be on the Nobel committee’s radar at this juncture. Mazzucato strikes me as a timely choice, especially if we consider someone like Richard Thaler’s work wasn’t that computationally complex either.

        1. macroduck

          Could be too political, unless we are about to get another Omaba-style prize.

          By the way, I meant to include the Nobel winners in the “good wrk” category, as contrasted with Covid Kelly mini-mes. Gluck (with an umlaut) is a treasure.

        2. Barkley Rosser

          I think the reason there is not an obvious choice this year is because things have been in such a state of upheaval between the pandemic, the economic collapse the demonstrations over racial justice, with the latter suggesting a possible Black candidate. I do not think this will be it but a colleague of mine suggested an interesting combo would be Glenn Loury, Robert Putnam, and Charles Manski. I have heard some buzz that it might be for econometrics as there has not been one for it for awhile, and there are multiple possible candidates for that (includnig Manski).

          If it is a woman, Claudia Goldin is probably more likely than Mariana Mazucato. As it is, I have seen lots of names being thrown around. I personally would like to see Richard Easterlin get it, who is well into his 90s, but I doubt he will. Some years I have ideas about who it might be, and I have been right on several occasions, but this year I do not have a feeling about it because there are so many cross-currents and upheavals going on. There really is a large list of possible winners this time around.

          1. Barkley Rosser

            My colleague, who leans Austrian and libertarian, has revised his forecast to one about social interactions, replacing econometrician Manski with sociologist Mark Granovetter, to go along with political scientist Putnam and Black economist Loury.

            I note that if the committee does not want to follow any particular current movement or fad, they can always fall back on remaining highly cited war horses who have not gotten it. I note that of the ten most cited living economists according to RePec, five have not yet received the invitation to go to Stockholm. In rank order of their citations and with no accompanying comment they are Andrei Schleifer, Robert Barro, Ken Rogoff, Daron Acemoglu, and Robert Vishny.

  9. The Rage

    The repo mess was the end. That was the sign real profits weren’t translating financially. The next phase was inflation to stop real profit deterioration. Which undermine consumer spending. 2020 wasn’t going to be bit recession year, but 2021 was. Covid sped it up.

    Let’s also note October 1999 has a similar problem but the Fed told the banks to handle it and they did.

Comments are closed.