Gross vs. Net Investment, Pre- and Post-TCJA

In other words, the sugar high fades. Brad Setser has noted how weak nonresidential fixed investment has been, despite the TCJA. It looks worse when looking at net — rather than gross — investment.

Figure 1: Gross nonresidential fixed investment (blue, left log scale), and estimated net nonresidential fixed investment (brown, right log scale), all in billions Ch. 2012$, SAAR. Net investment calculated by subtracting quadratically interpolated annual capital consumption, and assuming annualized 0.08% increase throughout 2019. Light orange denotes TCJA in full effect, very light orange denotes TCJA application to equipment. Source: BEA and author’s calculations. 

23 thoughts on “Gross vs. Net Investment, Pre- and Post-TCJA

  1. macroduck

    The latest slowing in net investment is not yet of the magnitude of the oil-driven slowdown in 2015 and 2016, but that should be small comfort to supporters of the tax cut. The slide in oil prices didn’t reflect badly on either party’s key legislative achievement. The best Republicans can honestly say in defense is “if it weren’t for Trump’s trade policy…”

  2. pgl

    “Pick your preferred cause — Fed, trade uncertainty, fall off in drilling, overestimation of the impact of US investment of giving US firms a tax cut on their offshore profits, no doubt others”.

    One would hope Brad would write a paper on this rather turning to The Twitter. Of course, the Twitter is the only thing Trump seems to read.

    1. Moses Herzog

      He writes a lot of articles. And when he writes he’s not “phoning it in”. One recently on Ireland. They are long ones where he gets into minutia. I’ve said before, next to maybe Michael Pettis he’s the best white dude on breaking down China economically, I’m sure there are Asian analysts better at the details on China but I seem to be unable to find them, probably because they write/post in their native tongue. I think Menzie knows a couple Hong Kong guys that are up there on the mainland situation.

  3. spencer

    I like to compare the change in net per employee investment lagged one year
    to the change in productivity. It clearly demonstrates that a primary
    factor in the weak productivity of recent decades is the much slower
    growth in the net capital stock per employee.

  4. dilbert dogbert

    His money base got theirs and his boob base don’t know or care what you are talking about.

  5. AS

    Professor Chinn,
    Is it reasonable to test the effectiveness of TCJA, by regressing dlog(gross private investment) on dlog(grosspi(-1)), economic uncertainty(FRED series: USEPUINDXD, and a dummy variable for TCJA for 2018 and 2019?
    Using prices indexes for GDP, I computed real gross private domestic investment (grosspi) back to 1985.
    The resultant model:
    dlog(grosspi) c dlog(grosspi(-1)) log(usrepuindxd) TCJA

    log(usrepuindxd) appears stationary
    Data used: 1982Q1 to 2019Q3
    Nobs: 137 after adjustment
    Rsqd .38

    1. dlog(grosspi(-1)) is significant at the 1% level
    2. urepuindxd is significant at the 5% level using.
    3. TCJA not significant
    4. Huber_White was used due to residual issue

    I used the FRED uncertainty index since I could not remember your favored uncertainty index.

    1. 2slugbaits

      AS Isn’t the uncertainty index you cited published at a daily frequency in FRED? And it looks like it only goes back to Jan 1985.

      1. AS

        Yes, the index is published daily, and reported as a daily index and monthly index. EViews averages the daily data to use as quarterly data.

        1. macroduck

          And you can ask FRED to turn the series into quarterly (monthly/semiannual/annual) data.

          You know that, but newbies might miss that frequency-adjusting capacity. I did when I first began using FRED.

    2. Menzie Chinn Post author

      AS: Commendable attempt. I haven’t had chance to look closely at the data, but real private fixed investment (as well as nonresidential fixed investment) appear to be I(1) series, so lagged dependent variable is problematic…

      1. AS

        Professor Chinn,
        Thanks for the response. I used the first log difference, dlog(real private fixed investment) and lagged dlog(real private fixed investment). I guess I need an “in-service” on this. I also used a model:
        dlog(real private fixed investment) = c log(usrepuindxd) AR(1) +error.
        Both models seem to show the same conclusion.

        I appreciate learning the correct form on this effort.


  6. Moses Herzog

    It’s looking more and more like a recession is the only thing that can save this nation’s soul. I present to you what I have long suspected is going on with rural trump voters still highly underrepresented in political polls (Nate Silver is a slow learner, who knew??). I am still quoting the number I have quoted for months: I give donald trump a 60% probability of winning in 2020 (not electoral vote count, probability of winning). <1.5% GDP is a crucial number, the lower you go on decimal points you shave off the 5%–10% he probably wins by with electoral college at this point.

    If you are counting on/ depending on donald trump to be voted out, you are being especially cruel in fooling yourself.

    1. Barkley Rosser


      Oh, and why is 1.5 percent growth such a crucial cutoff?

      Actually I think people are overstating how important the condition of the economy is. Yeah, if we really go down into a recession, it will hurt Trump. But we have now seen several presidential elections where there has been a reasonably good performing economy that did not save a candidate of an incumbent party. One was in 2000, although Gore did barely win the popular vote, and there had been a few indicators of impending downturn before the election. But mostly the late 90s was as boomy as we have seen in many decades, and it did not save Gore. Many said what did him in was lingering unhappiness over Clinton’s conduct that led to him getting impeached, even though he did not get convicted.

      Then we had 2016, although clearly there was some hangover still from the Great Recession. But Obama got reelected in 2012, while Clinton could not pull it off in 2016. Yes, she had a number of problrems, and like Gore she actually won the popular vote. But for all the things she was accused of, pretty much all of them Trump was far guiltier than she was. In any case, not too badly performing economy (a bit slower than a year or two earlier) did not save her.

      So I do not see a good performing economy saving Trump. It has slowed down from its peak in early 2018 after the tax cut came in, and while we may not have a recession or even get below 1.5 percent, it is not going to be booming (or highly unlikey it will be) more than it is now, which is pretty so-so. And even with his better performing economy in early 2018 Trump could never get poll numbers above 50 percent (hit it a couple of times in some polls, but never above it).

      I am not saying he cannot win. He certainly can as he did in 2016. But an economy doing better than 1.5 percent growth is not at all going to put him back into the WH, although it will be better for him than an outright recession, which certainly would just completely tank him.

      1. Barkley Rosser

        Just to add a bit more. The latest growth rate is 1.9 percent. If election were held today, would Trump have greater than a 60 percent chance of winning? Doubt it, although it would be close in a lot of the swing states.

        1. Willie

          My take is that economic performance, or lack thereof, in 2015 and 2016 was one of the foundations of Trump’s narrow victories there. There were other factors as well. Those other factors are unlikely to repeat in a similar way, in spite of undoubtedly upcoming Russian attempts on our society’s life. The economic stress in Pennsylvania, Wisconsin, Michigan, and Ohio is back and will have an effect. It may not be sufficient to unseat the incumbent.

          I have no problem fearlessly and unscientifically forecasting recession or other noticeable downturn in mid2020 with unwarranted precision. Election results are another issue. I will not touch any electoral predictions.

  7. Moses Herzog

    It is said that Juncker enjoys his drinky. Well I’m sympathetic to that, but drink or no drink, making love in public (if you saw the video you wouldn’t think I was exaggerating much, it was 1/4 step away from a french kiss) to the man who is the biggest threat to Europe does not spell out I-N-T-E-L-L-I-G-E-N-C-E.

    Juncker says “it didn’t hurt”?? I guess nothing commands respect like wet kissing tyrants in public.

  8. Moses Herzog

    Two important stories which could be semi-justifiably missed during all the impeachment buzz:

    Take note that the 2nd story linked above is co-authored by Jason Leopold, who I have stated before on this blog is one of the best journalists working in America today.

Comments are closed.