The Rebound Slows: Business Cycle Indicators, August 28th

Several of the five key indicators referenced by the NBER’s Business Cycle Dating Committee are decelerating, as shown in Figure 1.

Figure 1: Nonfarm payroll employment (dark blue), Bloomberg consensus for August employment (light blue square), industrial production (red), personal income excluding transfers in Ch.2012$ (green), manufacturing and trade sales in Ch.2012$ (black), and monthly GDP in Ch.2012$ (pink), all log normalized to 2019M02=0. Source: BLS, Federal Reserve, BEA, via FRED, Macroeconomic Advisers (8/3 release), NBER, and author’s calculations.

We will get August employment numbers in a week. Manufacturing and trade industry sales rebounded smartly in data going through June.

Income and outlay data were released today. While real disposable income is above pre-Covid-19 levels, it is declining. Consumption on the other hand has slowed its ascent, and is some 5% below NBER-defined peak levels.

Figure 2: Consumption in Ch.2012$ (blue), and real retail sales in CPI-deflated 1982-84$ (brown), both in logs, 2020M02=0. Source: BEA, Census, BLS via FRED, NBER, and author’s calculations.


41 thoughts on “The Rebound Slows: Business Cycle Indicators, August 28th

    1. rjs

      it won’t be a “V” when you graph it, but it could very well sound like a “V” when the numbers come out..

      that’s because an increase from a very low level is a much higher percentage change than an equivalent decrease from a high level…

      best be careful, because Kudlow may be able to claim his V yet….the October surprise will be on the 29th, when the advance 3rd quarter GDP report is released…

      1. Barkley Rosser


        Look at Menzie’s most recent post. The shape shown there for the GDP pattern, not in this one, but there, is a V (or close to it) that has since flattened. This is what I forecast back in June, and it is what has happened. Don’t get misled by numbers or some game that the Trump people might pull in October. Just look at the figures. By failing to cut a deal with Congress the Trump admin has just about killed any chance for overcoming seriously the slowdown that is now going on, although today I saw reports that Mnuchin is trying again to jumpstart talks with Pelosi. He gets it, even if Meadows, Kudlow, and Trump do not..

        Also, it is, of course, hilarious to see Moses the Troll somehow thinking that I am in agreement with Kudow and crew. This is just another lie. But he has simply gone to whole-hog Rich Stryker level lying. My position is clear, and it has proven correct, much as Moses the Troll continues to refuse to recognize that fact.

  1. Moses Herzog

    I’m pretty certain this news broke tonight (firings late night on a Friday are pretty common in Government).

    Now, I’m not going to be crying any tears for this woman since she worked for Tom DeLay and Ted Cruz–two of the most amoral men you’ll find on planet Earth, one–Cruz— groveling at the shoes of a U.S. President who basically spit in his wife’s face. But it is interesting to note that the action which instigated Miller being fired from her job was her advising the head of donald trump’s FDA to “walk back” his false statements, which by the way, have now turned the FDA into a national joke.

  2. Moses Herzog

    Michael Irvin has engaged in various behaviors which probably do some damage to a person’s immune system. However considering at one time Irvin was one of the greatest athletes in the world, and is now only age 54, I’m wondering if those who think wearing masks is about their “freedoms” have anyone they know that fits into Michael Irvin’s category?? For example Big 12 college football coaches with a big mouth and a small brain who think the world revolves around them: <— former Oklahoma State QB claims Mike Gundy ruined his life. Kind of hard to argue looking at Gundy's actions. This is what is often lost in the video and beer commercial jokes about Gundy's "I'm 40!!! I'm a man!!!" toddler tirade. That Gundy demonstratively ruined a kids' life whose parents he promised to protect. Not to mention the heavy rumors that Gundy was the one "leaking" the descriptions to Jenni Carlson about Bobby Reid.

    But Gundy is now telling all of his players (many of them African American kids from low income families) they need to get their butts out on the field and play. I'm curious if one of these players' older relatives dies before they find a vaccine, what is Mike Gundy or his athletic director going to say then?? "How could we foresee this happening??" Really?!?!?!?!?! How's that going to play out at a press conference or a PR release out of Stillwater in late Autumn of 2020??

  3. The Rage

    The populace I95/15 areas begin a quiet reopening this September, which should help personal services while durables will sharply decline.

    1. Menzie Chinn Post author

      The Rage: Much as I have a soft spot for the I95/I5 (not 15) area, I can’t see this having a really big impact on national level numbers.

      1. Moses Herzog

        I-5 Oregon and Washington is a little better than the Cali stretch in my opinion. Although if I am remembering right the stretch along the Edmund Brown viaduct has a certain charm to it. I always liked north California way better than south, even Sacramento and Fresno but that part might make some people puke. I probably have actually been more along the East West routes in California, but for awhile I had a dedicated route that went between Cincinnati (small town in proximity) to LA then to San Mateo (north I-5 obviously) and then back to Cincinnati again. We always picked up in San Mateo and never had any idea what the H we were picking up, I always assumed it was an expensive tech product or input but it was run by an anal retentive old grump and the trailer was always loaded when we went to pick it up. I always hated that guy, and it always seemed like forever to get the load. And I remember one time they announced Sinatra had died over the little radio he listened to in the office, and I said “Wow…… Frank Sinatra, he was one of those guys seemed like he would live forever”. and the guy, who never gave me eye contact, looked me straight square in the eyes with a dead expression and said “Everybody’s gotta die sometime”. My co-driver was a toothless redneck from Tennessee who literally could not read who claimed he had been in the Vietnam war and had an obsession with sprinkling the most vulgar terms related to the female anatomy in every other sentence of his conversation.

        “Joyous” times for me. But I-5 going north of San Fran (which I experienced on other random routes I drove) has some of the most beautiful places in the nation (at least as viewed from the interstate highway system). For truck drivers the West part of USA is exponentially more enjoyable than the East part of USA. Although both have their charms.

        Don’t dare me to drone on about anything, OK Menzie??

        1. Alan Goldhammer

          As one who was born and raised in California, I look for any opportunity I can to correct faulty state geography. Unless the big earthquake has come and picked up the fine city of San Mateo and relocated it 50 or so miles to the east, there is no way on this earth that you can drive on I-5 from Los Angeles to San Mateo. Now if you took the 101 north from LA you would in fact get to not only San Mateo but also help Dionne Warwick get back to San Jose.

          If you intend to drone on, drone on with facts and not ‘fake news.’

          1. Moses Herzog

            @ Mr Goldhammer
            San Mateo is in essence a suburb of San Francisco, or part of the San Francisco metro area if you prefer. If you know a better way than I-5 with a 53′ trailer attached to a semi-truck get to San Francisco I’m all ears to your MAGA revelations on traveling America’s highways. It may also shock you to know, semi-trucks do not dump their shipments in the middle of interstate traffic.

            What is it with people on this blog wanting to play an incessant game of “Gotchyeh!!!” about topics they have no clue about??

            BTW–you can call me someone who rambles—that may be very accurate—-you may not however call me a LIAR about something I did for f’king years.

          2. Barkley Rosser

            So, Mose, perhaps you would take I-5 most of the way, but then cut over, probably indeed going through San Jose on the cutover.. San Mateo is just north of the US city with the least variable temperature, Redwood City.

          3. Moses Herzog

            @ Barkley Junior
            “perhaps” I would take I-5 most of the way?? I tell you what—-you take 101 all the way north together with our resident shit-for-brains geography/navigation major Goldhammer all the way up to San Mateo, I’ll take I-5, to 580 and the associated branch offs to San Mateo. If we’re both going in a 4-wheel automobile I’ll be drunk at the bar by the time you two idiots park your car.

          4. Alan Goldhammer

            @Moses Herzog – you are constantly calling BS on other posters and hijacking threads with meaningless commentary. Someone calls you out on making a gaffe and you go through the ceiling. I knew there was a good reason for ignoring your ramblings. I should have adhered to my rule about feeding trolls. Needless to say, you can be comforted that I won’t be responding to you in the future.

          5. noneconomist

            OK, time for my two cents worth. All the N-S routes in California have their own charm and worth and provide examples of the breadth of the state and its economy.
            I’ve done I-5 (Sacramento to SoCal) more times than I care to count, but I still marvel at the Central Valley and the abundance of agriculture therein.Whatever is in your diet, much of it likely comes from there.
            I-5 to Portland-Seattle is more scenic, and less nerve rattling (for me anyway).
            101 SoCal to the Bay Area (then east to the Sierra foothills) is a preferred route if I have time with plenty more examples of watching what you eat (and, if you’re a wine drinker, drink) growing everywhere. Double that with a jaunt up 101 to Portland.
            Great example of the importance of tourism is any drive down California 1. We’re very partial to the drive from Half Moon Bay to San Luis Obispo. Last time we drove it was late February, a few years ago, with temps in the mid 70’s.
            Have to admit, though. my favorite N-S and S-N drive is 1-15, 395, San Diego to Reno (or family wise, San Bernardino to Reno) then west to the foothills. If you’re a fan of westerns, you know you’re in a familiar place:when you hit Lone Pine: High Sierra/Mt. Whitney to the W, desert to E. We’ve even been fortunate enough to do Spokane to Reno as well
            Whatever hell Ms. Guilfoyle was describing at the Republican Convention is largely absent from these areas and a figment of vivid imaginations, assisted perhaps by the ingestion of illegal substances and/or beverages imported from Kentucky.

          6. Barkley Rosser

            Unlike Alan G. I was not born in CA, but have spend lots of time there and have lots of family living there. I have taken I-5 and 101 and Cal 1 in both directions between the LA and SF metro areas numerous times. So I simply tried to help out with noting that indeed you could do I-5 if you cut over, which you know but did not write in your long drone. When Alan pointed out that what you described in your overly long drone did not work Because I-5 does not go to San Mateo, you erupted back, not at all admitting that he was right and then putting in emboldened caps how you are not a “liar.” Well, maybe you are not a “liar,” Moses, but you have said a lot of incorrect things here and continued to do so even after being corrected multiple times.

            So, at least you have agreed that my correction of what you said is correct, but you had to adopt a tone as if I was making a mistake. I did not advocate taking 101 or anything in particular. I simply agreed accurately with Alan that you could not take I-5 itself all the way to San Mateo, which you now admit, although indulging in all kinds of whining and back biting that now have Alan saying he is not going to waste time responding to you.

            As for scenic, I agree that north of the Bay Area I-5 can be pretty scenic, but it is not at all scenic on that route you drove many times between the LA and SF areas, although it is certainly faster most of the time than 101. Of course the most scenic route by far, way ahead of either I-5 or 101, although definitely slowet than either, is Cal 1 on the coast, widely agreed to be one of the most scenic roads in the entire nation, if not the most scenic.

            BTW, you seem to be doing a lot of this lately, Mose. On another thread I agreed with you that it looked like we are having a K pattern and then mildly noted that was not a pattern about aggregate outcomes. This set you off on a string of insulting posts where you revived your unbelievably ignorant and incorrect rant about how I was just so wrong to call the aggregate pattern “a V that flattens,” even though virtually everybody here who has expressed an opinion on that matter has agreed with me on the matter with not a single person siding with you, not one. I can’t even provide a helpful remark like one can indeed do I-5 but will need to cut over at some point without getting a bunch of insulting baloney out of you. You really need to get your act together here, and, yes, Alan is right, you post way too much irrelevant and extraneous commentary and links. You do not own this blog, Moses, so cool it, please.

          7. Moses Herzog

            @ Alan Goldhammer
            If that’s your way of apologizing because after living in California multiple years you never figured out !-5 interstate is the fastest route from Metro LA to San Fran and suburbs I accept your apology. I could find a 12 year old who lives in Cali who could probably inform you why the interstate highway system was created. That 12 year old would tell you it’s not an optimal way to transport goods over long distances or between states when they have to stop at 3+ stop lights in every town traveling north twisty zig-zag road on 101. Now if you are too DUMB to figure that out after residing in the state many years I’m very “sorry, not sorry”.

            I give people their just due when they deserve it, even those I hate. Such as donald trump being better at campaigning towards the latter part of the 2016 race than Hillary Clinton.

            Does that reality make me happy?? No, it does not. If you talk out of your a$$ on this blog, yeah, I will call you out. Even those who can dish it out but cannot take it.

            Now, in extra summary here, I’m pretty sure BOTH the authors of this blog are familiar with California. If EITHER Menzie or Professor Hamilton want to mediate or arbitrate this little dispute and make a child happy by telling him 101 from LA to San Mateo is more efficient time and resource wise than I-5 is, either in a car, or transporting goods from point A to point B, I’ll be very entertained to see either of those intelligent gentlemen present that one to us. Maybe one of those two men would like to label my thoughts on this particular navigational quarrel “fake news”?? I tend to doubt it.

          8. Menzie Chinn Post author

            Moses Herzog: I-5 is faster but more soul-crushing than 101. And Hwy 1 more magical, but you have to have the right vehicle to traverse, of course.

          9. Barkley Rosser

            You are now completely out of line, Moses. I find it absurdly ironic that you claimed somehow that Alan G. was accusing you of being “a liar.” What is indeed ironic is that in fact you have lied here. You are a liar, Moses, liar liar liar. You somehow decided that AG said that 101 was the fastest way to go from LA to San Mateo, and have gone apoplectic over that. But he said no such thing. Go read what he said, unless you have gone blind or become utterly illiterate. You will see he did not say what you excoriated him for allegedly saying. That makes you a liar, liar, pants on fire!

            What he said very clearly and accurately was that you could not get to San Mateo by using just I-5 but can by using 101, and he accurately noted that I-5 goes a good 40-50 miles or so east of San Mateo. You could have just accepted this correction with good humor, but no, you had to start ranting about him accusing you of being a liar while you ironically spouted lies about what he said, and it did not get much better when I weighed in with my accurate correction, which you did finally grumpily accept.

            Oh, and no, I do not think Alan G. was apologizing to you. Just the opposite. He was calling you out for unacceptably bad conduct on your part. You are the one who should be apologizing to him, not the other way around, although you have a long and bad record of very rarely apologizing when called upon to do so. As it is, your conduct towards Alan here is just execrable and inexcusable.

            As it is, looks like Menzie agrees with my view on the various roads: I-5 generally faster than 101 generally faster than Cal 1, but in terms of scenery it is the reverse, with Cal 1 the most beautiful and I-5 the least, even if noneconomist likes all that Central Valley ag.

          10. Moses Herzog

            @ Menzie
            Going long ways over the interstate can be rough going. Although I will say if you have travelled over most of America’s interstate highway system (as I have) you will find California’s (outside of major cities, arguably even in those cities, but San Fran is a hellcat in a semi truck) more merciful, even on I-5. I had imagined in my mind that state route 1 and U.S 101 would have included places where many scenes from one of my favorite films of all time, “Sideways” had been filmed. There are a few. Unfortunately (or fortunately??) I never got to live the life of Miles Raymond while driving the semi.

          11. Moses Herzog

            @ Stuart
            I guess this makes me “Devin” because I keep wanting to crack up in the middle of this comedy skit. “Duuuuude, my GPS says I should take I-5 up most of the way to Napa Valley, then switch to the I-580W, the I-680N, and then the I-80W because there’s a guy there that sells black grapes on the off ramp. But Dr. Goldhammer says I should, like, take state route 99 north and then, like, connect to state route 4 going west around Stockton, then, like, to I-80, because I-80 goes right through Napa Valley and I-5 is ’50 or so miles’ east of that, so, like, you can’t, like, take I-5 up to Napa Valley from LA or south California. Duuuuude, so I-5 is dumb man, it’s tooooooh-tuhlly dumb to take I-5 all that distance, because I-5 is, like, ’50 or so miles east’ of Napa Valley. Duuuuuude!!!!!!….. you need to talk to Goldhammer, he’s lived in Cali, like, a long time and he’s liiike a navigational genius and stuff. Like tooooooooh-tuhlly duuuuuuuude!!!!”

        2. Barkley Rosser


          Oh my, incorrigibly smearing Alan Goldhammer again are we? I guess this means you have been listening to Sinatra too much on the road and so we must hear you trying to croon, “I Did It My Way.”

          BTW, I suppose your “dedicated route” did not go through San Jose but rather went from 580 to and across the San Mateo Bridge, going north anyway.

          1. Moses Herzog

            @ Barkley Junior
            NO, you were going to take the 101 with Goldhammer remember?? Uh-Oh….. changing your goal posts again. Did the “mathematical economist” finally figure out how to do the 3rd grade math on mileage and route time from LA to San Mateo?? Call the grandkids, they are going to be so proud of you!!!!

            Phone up Goldhammer if you get confused on how to find the Golden Gate bridge. My personal preference would be if you took your personal vehicle (without the grandkids) and see if you can make it via 4-wheels to Alcatraz. I believe in you.

          2. Willie

            All I can say about this is that I5 is mind numbingly dull in Oregon between Porland and the mountains. The mountains are interesting and the the Central Valley is dull unless you run into a Tule fog. Then it is terrifying. One other thing – do not drive south on 1 in a loaded van with a nervous passenger. The co-pilot demands for ever slower speed gets old.

        3. Barkley Rosser

          Well, I am not going to comment on this further as it has now become utterly tireseome and Troll Moses is totally out of control with his “humor,.”
          But, Mioses, you are getting worse with your liar liar pants on fire routine. I moved no goal posts, liar, I never recommended taking 101, liar, and neither Alan G. nor I ever said taking 10i1 was the fastest route, liar yet again. Are you trying to imitate Rick Stryker here, liar?

          You might as well simply agree that I have pointed out, given what you added later about your cutover, the fuller details of your “dedicated route.” Again, you could have simply accepted that Alan G, was right that I-5 does go fairly far east of San Mateo so that you cut over on 580 and the San Mateo bridge to get there and also reverse to return, without all the pointless insults and lies.

          Gag, you may think you are funny, Moses Troll, but actually you are thoroughly disgusting. Some have said Menzie should ban Rick and CoRev and Bruce H., but actually you are the one who should flushed down a toilet out of here for good, liar liar, pants on fire.

  4. dd

    Mini data point: everyone in our family structure is taking an economic hit. We’ll all be fine but spending is decelerating at a rapid pace.

  5. AS

    Professor Chinn,
    Given the large shock in data due to the pandemic, is it a legitimate time series forecast method to use the past few data points of PAYEMS to forecast August 2020 change in PAYEMS?

    Using the the past four data points, Microsoft Excel and the log trend function, I find that the next data point implies a change in PAYEMS from July to August of about1.2 million jobs. This change agrees closely with the consensus forecasts from various sources as mentioned in this Blog.

    Using changes in ICSA, CCSA and WEI, I find the change near zero or a bit negative.

    Perhaps a transfer model is preferable, but I am having difficulty implementing the transfer function properly.


    1. Moses Herzog

      @ AS
      Not getting into models, of which I only have very surface knowledge, I think you need to keep in mind a lot of the data started to look worse again after mid-July.

      Can you humor me?? Are you using ARMA or ARIMA model and how did you decide which to use??

      1. AS

        Hi Moses,
        My understanding of the difference between an ARIMA and ARMA model is that the data related to an ARIMA model are not stationary and must be differenced in some fashion to become stationary. An ARMA model uses stationary data.
        I have tried several models to forecast PAYEMS and the data are non-stationary. So, I assume some type of ARIMA or ARIMAX model will be needed. ARIMAX models include independent variables
        Since there is an abrupt decline in PAYEMS in April 2020, I have tried to implement a version of an Intervention Model as described by Bowerman et al, “Forecasting, Time Series and Regression” in chapter 12. The authors demonstrate a model in which Cincinnati Bell Telephone begins to charge for information services. Requests dropped dramatically with an abrupt start and the interaction seems to have a permanent effect. I was able to model this, and my forecast agreed with the authors. However, there is another model that may better fit the decline in PAYEMS. The pandemic interaction had an abrupt start but may have a gradual decay. I may not be interpreting the transfer function properly that helps to transfer the model prior to the pandemic to after the pandemic. The forecast shows an increase of about 2.0 million, which makes me think I am not interpreting the transfer function correctly.
        Another group of models that I have tried uses mixed frequency. ICSA, CCSA and WEI are reported weekly, while PAYEMS is reported monthly. EViews has a model called, MIDAS, which allows one to use mixed frequencies, in this case, weekly and monthly data. One MIDAS model forecasts 0.5 million change in PAYEMS, another shows a change of -0.6 million.
        EViews also has a forecast evaluation function that compares forecasts for a selected period, using RMSE, MAE, MAPE, SMAPE, Theil U1 and Theil U2. Using this function to compare forecasts from January 2020 to July 2020, the function says I should average my EQ02 and EQ03 to minimize the forecast error. Averaging these two models shows a forecast of about -0.1 million change in PAYEMS. Hope this helps and I assume interested experts will correct any misinterpretation I have presented.

        1. 2slugbaits

          AS With a transfer model you need a purely exogenous variable with no feedback effects from the dependent variable. Good luck with that. It sounds like you might be planning on using pandemic cases (or deaths???) as the exogenous variable and then use the latest cases and deaths to forecast out-of-sample employment. That could be tough because there won’t be a lot of pandemic observations, so the relationship in the transfer function is apt to be pretty weak.

          I think the bigger problem is that employment drops faster than it recovers, which is another way of saying that the differences are non-linear. An ARMA model assumes differences are linear. Testing for non-linearity is pretty straightforward (e.g., McLeod-Li test or a RESET test). But trying to find the appropriate non-linear model can be a job.

          1. Barkley Rosser

            Indeeed, while first differencing often induces stationarity when it was not there in the original series, it does not always do so. HIgher order differencing might induce it, but some series simply cannot be made stationary by any sort of manipulation or transoformation.

          2. AS

            Thanks for the comments.
            It is difficult to determine from the Bowerman text if a nonlinear model is used since the text uses a SAS program we discussed in the past. The SAS program has an automatic function to evaluate intervention models. EViews does not have an automatic function for intervention models. The Cincinnati Bell Telephone problem is modeled as a linear function. Since the data are seasonal, the model needed both non seasonal and seasonal differencing to make the data stationary. After finally deciphering the mathematical symbols, I saw that the model could be expressed as: d(calls,1,12) = d(dum,1,12) ma(12). This seems simple, but it took me a while to understand what was being written. As mentioned, the authors use SAS and the model was not stated clearly for software users other than SAS, in my opinion. In this case the Ramsey RESET test shows a t-statistic of 0.65. This model is an example of an abrupt start and permanent effect.

            An abrupt start and gradual decay model seems to be nonlinear, but the text is not clear on this intervention model. I defer to your knowledge.

        2. Moses Herzog

          @ AS
          Very educational, Thanks for the thorough and methodical reply, I learned much from this comment, as I often do reading your own thoughts and queries to Menzie. Comment sections can add a lot of value to blogs—you are the ideal contributor in this attribute and one of the gems of this blog.

          1. AS

            Thanks for the kind comments. I’m just an old duffer who enjoys time series models and finds them endlessly fascinating and challenging.

Comments are closed.