Is Trump “Special”? Economic Policy Uncertainty Levels in Perspective

The answer is “yes”.

Reader Ed Hanson asserts asks if that economic policy uncertainty is no higher under Mr. Trump than during other periods of election, transition and new administration.

Your post always have a point, so what is it? Are you showing that for the Trump administration, and despite the constant anti-trump rants we read and see in the main media, that uncertainty index is generally falling? And that means what, that Trump seems better than most recent politicians to mean his promises and to follow through on them and people are now realizing this and uncertainty is reduced?

Sure wish you had extended your plot to cover other administrations so we could see if the pattern of rising uncertainty just before gaining office, followed by reducing uncertainty is the normal trend. Is it?

I have plotted the entire span of the economic policy uncertainty index from Baker, Bloom and Davis in several previous posts [1] [2] [3]. However, to formally and comprehensively address this query, first consider the data:

Figure 1: Economic policy uncertainty index (news) (blue), NBER defined recession dates shaded gray, vertical solid lines at change of administrations, light orange shading denotes Trump Administration. Source: and NBER.

Economic policy uncertainty as measured by Baker, Bloom and Davis does seem elevated before and after new administrations, but it’s hard to tell. I estimate the following OLS regression, over the 1985M01-2018M02 period:

epu = 102.55 + 5.23election + 6.73Trump + 11.85recession + 40.38Lehman + 44.309/11 + 34.52Brexit

Adj-R2 = 0.12, SER = 29.85, N = 398, DW = 0.38. bold denotes significance at 10% msl, using HAC robust standard errors.

Where election is a dummy variable that takes on a value of 1 12 months before an election through 12 months after inauguration, except when the president is re-elected, in which case the dummy takes on a value of 1 in the 12 months before to up to the month of the election; Trump is a dummy variable for the Trump administration; recession is a dummy variable for NBER defined recession dates; Lehman is a dummy variable taking a value of 1 for 2008M09-2009M06, 9/11 for 2001M09-M11, Brexit for 2016M06-M07.

The proportion of variation explained is fairly low; in addition the residuals look highly non-Normal.

The Jarque-Bera tests strongly rejects the null hypothesis of Normality.

I re-estimate using quantile regression (Least Absolute Deviation), which estimates the conditional median (rather than conditional mean as in OLS).

epu = 93.61 + 6.74election + 10.39Trump + 11.35recession + 37.13Lehman + 61.439/11 + 34.52Brexit

Pseudo-R2 = 0.08, SER = 31.03, N = 398. bold denotes significance at 10% msl.

In other words, the Trump administration is indeed “special”; economic policy uncertainty has been higher than expected, even conditioning on the newness of the administration (and not including the elevated uncertainty associated with the Section 232 tariff announcement on 3/1).


40 thoughts on “Is Trump “Special”? Economic Policy Uncertainty Levels in Perspective

  1. Moses Herzog

    Hahahah, FUNNY. No one can say Menzie doesn’t give reader questions their “fair due” or a “fair hearing”. Unrepresentative of whether they deserved the extra effort on his part, or not…….

    As they say, “it’s all good” though. It gives us another white paper, some math examples to tinker with, and another blog reference. This is what exemplary teachers and Uni profs do.

    Is the VSG “special”?? His Mother Mary Anne MacLeod always said he was “special”:

    1. pgl

      The excerpts Grumpy gave us are fine as they go. As I noted in the comment section – people should refer to both Modigliani and Miller propositions. He has left out a lot of his article as the WSJ so maybe this is premature but nothing in his post indicates that the new tax law did a single thing to promote more investment.

  2. joseph

    “Trump is a dummy variable for the Trump administration”

    Should have just stopped right there.

      1. Moses Herzog

        The preferred nomenclature is “senile variable” and we’re working on a copyright for that at this very moment.

  3. PeakTrader

    None of the major newspapers endorsed Trump. Even moderate and many conservative newspapers declared Trump unfit for office, since he had no prior experience. And, the news hasn’t been fair – it’s been overwhelmingly negative, hysterical, and to a large extent fake. There’s likely a huge disconnect between the media and Main Street. Democrats want Trump impeached for no reason. Many Republicans endorsed Hillary. All of Trump’s policies are reported in the most negative way.

    1. PeakTrader

      And, regarding Trump’s trade policies, everyone knew his position during the campaign, including renegotiating NAFTA – he’s just following through. The media hypes it up, and European trading partners are talking retaliation only as a scare tactic. It’ll likely be much ado about little. Yet, the media will run around with their hands up in the air, like Mel Brooks in that movie 🙂

    2. pgl

      Once again Trump’s #1 cheerleader shows that he has no clue what Menzie’s post was even about!

      1. PeakTrader

        Pgl, obviously, you have no clue what you’re talking about, as usual. Look at the probability distribution for a clue.

        1. Menzie Chinn Post author

          PeakTrader: I’m unsure what the distribution for the residual shows, besides the fact that a linear, least squares regression was not optimal.

          1. PeakTrader

            Menzie Chinn, it doesn’t show the Trump Administration in itself causes the higher economic policy uncertainty.

          2. Menzie Chinn Post author

            PeakTrader: Yes, that’s right. Without taking a stand on exogeneity, the reported coefficients indicate correlation.

            A Granger causality test rejects the null that Trump doesn’t cause uncertainty at 24%, while it fails to reject the null EPU doesn’t cause Trump (p-val 75%). Take that as you will.

    3. Moses Herzog

      @ PeakIgnorance
      That’s really inconsiderate and discriminatory to major newspapers. There was at least one major paper that endorsed Donald “The VSG” Trump:

      When I read this newspaper editorial, as a current dues paid member of the Tea Party, the Alex Jones Fan Club, and the John Birch Society, it brought tears to my eyes that someone is willing to defend Donald “The VSG” Trump. I also believe that a chemical cocktail of chemtrails, fluoride, black helicopters, and New Coke all combined together caused Donald “the VSG” Trump to involuntarily molest women’s ladyparts, and a strange green-colored Canadian Sasquatch hypnotized Trump’s lawyer Michael Cohen to payoff porn stars to keep quiet about affairs that never happened.

      I plan to unveil all of the proof, documentable evidence, and corroborating support TONIGHT, on the premier episode of my new podcast: “PeakIgnorance, KookyTown Guys, and the Women Who Love Them”. If you order it within the next 16 hours, I will also throw in our special, one-time only discount, streaming rate of $19.95 per half-minute listening time, buffering problems included in the time fee.

  4. Ed Hanson.


    Before continuing the discussion of the uncertainty index, three things need to be said.

    One. I appreciate you taking the time to answer my questions.

    Two. But note, I said questions, not assertions. You are wrong to write, “Reader Ed Hanson asserts that economic policy uncertainty is no higher under Mr. Trump than during other periods of election, transition and new administration.” I had asked questions because your chart did not extend back far enough historically to put it in perspective. When I make assertions, it will not be in the form of interrogatives, but in a declarative form.

    Three. And this is most important. It must be noted that you, Menzie,, changed the variety of uncertainty index used in this post from the one in the previous post. It is the previous index, not this one, that generated the questions and discussion. This will cause readers of the following the continuing discussion to be forced to “click” back and forth between post. This should not be. This problem is similar to the one you caused rtd in a previous post when you unfairly edited the very lines that he was discussing. You admirably corrected that problem in a subsequent post. It would be best if you edited this post by placing the original uncertainty index within it (extending to the complete time period) and explaining the change and difference between the two varieties.

    I will post this and then continue in a separate reply so as not to make this too long .

    1. Menzie Chinn Post author

      Ed Hanson: OK – sure seemed like you felt that Trump era EPU was not elevated, conditioned on newness of the administration, but I’ve changed the text.

      You do know that the link in the legend goes to graphs of both the daily and monthly indices? You could have spent 30 seconds to click on two links to get there…

      Finally, the plot in Figure 1 of the new post is the average of daily figures used in the Figure 1 of the original post. I don’t know what you mean by the assertion that the uncertainty index is different…Or do you want me to plot daily EPU for the 32 years of data? That would sure be messy, but I can do it if you want such a graph.

      1. Moses Herzog

        @ Menzie
        I think Ed Hanson attended the same “I was was busted for being WRONG, so now I am going to back-pedal and say my former factually incorrect statement was ‘just a question’ ” “School of Debate” as “rtd”

        I have admitted at least once on this site I was wrong. It’s really not that hard. You lick your ego wounds in private moments and the doctor tells you to drink more “Monster” drink so you can read those white papers late into the night. I’m told, as part of my “long-term rehab”, that if I drink enough vodka on weekends I’ll forget I was ever wrong.

        1. rtd

          I’m not going to pay you a dime of rent despite occupying the space inside your skull.

          I’d love for you to point out an instance where i have been “factually incorrect” and back-peddled claiming I was asking “just a question”. I may follow up, but my interest in your comments has become practically non-existent.

      2. Ed Hanson.

        Menzie you write,

        “Finally, the plot in Figure 1 of the new post is the average of daily figures used in the Figure 1 of the original post.”

        No, The previous post used the ‘7-day moving ave’ Economic Policy Uncertainty Index, while this post uses what looks like the “Monthly’ Economic Policy Uncertainty Index. The numbers per date are completely different. A needless complication for the discussion and once again an occurrence where you make substantial changes without so noting so to your readers. It is a disservice to your readers. As described by me in another post, the historical “7-day” is quite available at and should have been used in this post.

        1. Menzie Chinn Post author

          Ed Hanson: You have the advantage of me. I have been downloading data from for years; I don’t think I know where the link to the 7 day moving average series is — can you provide (the 7 day trailing moving average is plotted, but I don’t know of a link).

          Thank you in advance.

  5. Ed Hanson.


    nice Sudoku as my friend Mr. Fox might say. But we had already established that since 2008 the index has been “special” being twice as high as previously historically. I suspect a similar fair mathematical treatment of the Obama years would find them also “special.”

    Note to readers. The following refers to the original variety of uncertainty index found, not in this post, but the previous one.

    Looking at the index, the value for approximately 3 months before and 1 month after the Trump inauguration was about 140. By the end of 2017, the last 3 months, the value was approximately 90. The index shows a generally decreasing uncertainty. But Menzie would like us to look at the recent spikes of the index and their cause. Fair enough.

    Concerning the spike, thereal2slugbaits wrote, “The top chart is telling us that up to the Nov 2016 election there was only one significant spike in the uncertainty index.” Mike v wrote, “Can you really not see the giant spike at the end due to the tariff announcement? And how uncertainty, while trending lower until the tariff announcement, has been higher on average since the election than in 2016? Both excellent observations from the chart provided by Menzie, but as I assert, historically not particularly relevant or even correct. Perhaps by the time this reply can be read, Menzie will have made the edit requested, but if not the readers will have go to and read the chart labeled ‘US daily news-based Economic Policy Uncertainty (7-day moving ave)’ and click “All”. This will be Menzie’s original chart extended historically. You will see that “up to the Nov 2016 election” there were many spikes, that they are not unusual for the index, and the particular spike being discussed is not “the giant spike” but actually rather small and mild in comparison, historically.

    One last observation, before the spike being discussed began another larger spike occurred beginning about De. 10, peaking about Jan. 6. This peak, I repeat, was larger than the current one and it had nothing to do with tariffs, but measured the uncertainty deriving from the Tax Bill passed and signed. That is how the index works, it might spike reasonably close to events then settle back down. If interest in the index remains after Menzie tires of it, it will be interesting to see at what level it settle to.


    1. Menzie Chinn Post author

      Ed Hanson: The original post has not been altered except as requested *by you*, as shown by the strikeout and new text in italics.

      You have not addressed my point: The index is the *same index* in the two posts. In the first post, it’s the daily index which is news based; in the second post, it’s the monthly average of daily data on the news based index.

      1. Ed Hanson.

        I have addressed it , Menzie, but you are just faster than me, and my post are somewhat out of order do to the length of time it takes me to compose.

        But just to be clear, and repeating, the ‘Monthly’ and the ‘7-day’ are not the same. Although derived from the same data, it results in substantially different values corresponding to the same day. This makes apple to oranges, and needlessly complicates the discussion.


        1. Menzie Chinn Post author

          Ed Hanson: Legend to Figure 1 reads “Figure 1: US daily Economic Policy Uncertainty index from Baker, Bloom and Davis (dark blue), and seven day centered moving average (red). Orange shading denotes Trump Administration. Source:, and author’s calculations.” I stated that the series plotted in Figure 1 of ‘Is Trump “Special”?’ reads: “Figure 1: Economic policy uncertainty index (news) (blue), NBER defined recession dates shaded gray, vertical solid lines at change of administrations, light orange shading denotes Trump Administration. Source: and NBER.” is the monthly average of the daily data series. That series is downloaded from, while the daily data is downloaded from . By the way, the 7-day average is not downloadable as far as I know. Furthermore, I used a centered moving average, not a trailing.

          I really don’t know what you are arguing about. The monthly average of the daily data is a simple transformation. But, you know what, I’m going to plot the daily data just for you.


      1. Ed Hanson.


        Yes still awaiting, and am surprised how far I missed. While Wisconsin has done alright, certainly in terms of unemployment rate, as well as continued state overall growth, I am surprised it was not more robust. I guess I could go to the leftist excuse for the slack performance and high debt load of Obama years, and just say how much worse it would have been without the Walker initiatives, but that’s not my style.

        Now how about your turn and your defense of the terribly inaccurate GDP predictions of CBO. Still hoping that the country reverts to the anemic growth of the Obama years? I wouldn’t even if Clinton was in the White House.

        And yes with the Fed gov pulling maybe 2018 will finally be the year. Oh Well


      2. Moses Herzog

        Most likely what happened here, Ed Hanson was watching a 2013 infomercial for Sprott Gold, and mistook the image as a block of Colby cheese.

    2. Menzie Chinn Post author

      Ed Hanson: Given tonight’s announcement regarding the head economic policy adviser, I’m guessing it won’t settle down to a really low level.

  6. rtd

    Interesting approach to play around with, yet in doing so I’m (again) unable to replicate your results. I believe that I understand your explanation of specifying the dummies and that’s really all I can think of that I may be doing differently. As such, I’m (again) interested if you’re interested in sharing these eviews workfiles with your readers.

  7. Erik Poole

    Nice use of the Jarque-Bera diagnostic test.

    Menzie: Questions follow. Just curious.

    Do you examine the residual chart of every regression you post on

    I am not that familiar with quantile regression (Least Absolute Deviation) though it certainly appears to make sense in this context. Are the magnitudes of the estimated coefficients in a LAD regression comparable?

    If so, the Trump contribution to economic policy uncertainty (epu) is not that great compared to other sources of epu.

    Frankly, I suspect that Trump is in reality having a much more negative impact on near-term growth than your LAD regression suggests. But it will not become fully apparent until well after the fact.

    1. Menzie Chinn Post author

      Erik Poole: No, but if the LHS variable is very spiky or seems poorly characterized as a Normal shaped distribution, I consider alternatives (usually Tobit for censored, quantile for those with outliers, and of course probit for binary dependent).

      1. Erik Poole

        Thanks Menzie Chinn. You get major points in my book for using a Jarque-Bera test in a blog like this.

        joseph: The anonymous comments re: Cochrane were mine. I had accidentally erased all my cookies!

  8. Bruce Hall

    I’m beginning to be uncertain about what all this means. 😉

    But perhaps we will get a better idea when this index reflects the tariff “event”:

    And this from two weeks ago:
    27 Feb. 2018

    Index at Highest Level Since 2000

    The Conference Board Consumer Confidence Index® increased in February, following a modest increase in January. The Index now stands at 130.8 (1985=100), up from 124.3 in January. The Present Situation Index increased from 154.7 to 162.4, while the Expectations Index improved from 104.0 last month to 109.7 this month.

    Again, we’ll have to see if the Trump tariff event… and higher interest rates… have an impact.

  9. joseph

    “Cochrane tosses up a softball.”

    Yikes, for a finance economist to be that wrong/misleading is unfathomable. He claims that the stock buybacks prove that tax cuts work. How’s that? By preventing companies from spending their new tax windfalls on wasteful projects. Yes, it’s just that stupid. If they don’t have any ideas for good projects, why give them a tax windfall?

    And he goes on “Not all companies should make new investments, and some of the best investments come from new companies that don’t have profits yet.” Well, companies that don’t have profits yet also don’t pay taxes, so they don’t benefit from the Republican tax cuts at all. That’s right, Cochrane admits that the best investments aren’t helped by tax cuts at all.

    Cochrane seems to be that same type of economist as Navarro. When younger that seemed to be smart and show promise. But they went over to the dark side, selling their souls as loyal Republican propagandists and writing drivel.

  10. Anonymous

    joseph: Interesting comments. I find Cochrane a compelling read on lots of policy issues but on this one he has me scratching my head.

    Share buy-backs can be very useful for senior managers who want to issue large numbers of generous options without unduly diluting the number of issued and outstanding common shares. Share buybacks often end up enrichening top tier management and to a far lesser extent ordinary shareholders.

    As an investor, I tend to view share buy-backs with a jaundiced eye. There are exceptions where share buybacks make sense from the perspective of ordinary shareholders but not often in my experience.

  11. Douglas Araujo

    Prof. Menzie, interesting post. One question: shouldn’t there also be dummies for the other Presidents? It shouldn’t change the Trump Beta, but could help put it in perspective to other administrations. Maybe also increase r2 and improve residuals distribution.

    1. Menzie Chinn Post author

      Douglas Araujo: Yes, depending on your null and alternative hypotheses. Probably will increase R2, but given the pattern of EPU, doubt it’d change the attributes of the OLS residuals substantially.

  12. rtd

    Here’s what I came up with:

    epu = 104.09* + 3.20election + 28.71Trump* + 17.51recession + 45.41Lehman* + 106.049/11* + 91.91Brexit*
    Adj-R2 = 0.18
    *denotes sig @10%


Comments are closed.