Most shocks don’t matter

Over the past 18 months we’ve seen three major shocks, each of which were expected to have a significant impact on growth:

1.  The late 2015 Chinese yuan devaluation/capital outflow shock

2.  Brexit

3.  The Trump election

Many people expected the Chinese economy to slow sharply in 2016.  It didn’t.

Many people expected the UK economy to slow sharply after mid-2016.  It didn’t.

Many people expected growth and inflation to rise significantly after Trump was elected.

The third prediction still might come true, but this FT article suggests that investors are moving away from the “reflation” trade.

We are nearing the anniversary of a great market turning point. Like most such turning points, it was not obvious as such at the time, but in early July last year markets reached the lowest point of their fear of deflation — falling prices amid a stagnating economy — and started to position for reflation.

The shock of last summer’s Brexit referendum brought bond yields, the market’s most direct expression of its belief in deflation, to a historic low. The rebound started as the effects of China’s economic stimulus were felt, while it grew clear that Brexit had not sparked a financial crisis.

It gathered momentum after Donald Trump’s victory in the US presidential election. The theory was that the Trump administration would pick up the baton from China, which looks over-leveraged and will soon need to ease off its stimulus, and would bring in its own growth-friendly policies, including tax cuts and infrastructure spending. From “reflation off”, we moved emphatically to “reflation on”.

Stock markets have risen this year, but 2017 has seen a gradual shift back to “deflation-off”. Short-term inflation expectations have moved sharply in recent weeks. The bond market’s implicit forecast for US inflation over the next two years, once at 2.15 per cent, has dropped to 1.35 per cent.

In each case I expected some effect, but less than the consensus.  In the case of China and Brexit, even the very mild slowdowns I predicted proved too bearish. I’ve learned my lesson from Brexit, and in the future will pay no attention to “uncertainty shocks”.  Always a skeptic, I am now convinced that uncertainty has virtually no significant business cycle effects.  And never bet against Chinese growth. Someday it will falter, but no one can predict when.

[In macro, the forecast most likely to be true is, “Not much will change”.  However the way to build a reputation is to forecast dramatic changes.  You will usually be wrong, but the public will forget those mistakes and give you undeserved praise on the few occasions when you are correct.]

I’m sticking with my prediction that Trump’s policies might lead to a couple tenths of a percent faster NGDP growth.  Monetary offset will keep inflation close to 2%, and Trump’s supply-side reforms are likely to be modest (at best).  An extra couple of tenths of one percent NGDP growth is so small it will be almost impossible to tell if I am correct—especially given that growth is likely to slow as we approach full employment.  If there is a recession, however, or 5%-6% NGDP growth persisting for a few years, then I will clearly be wrong.

Maybe the French will give us another shock later today.

PS.  I still believe that Brexit itself will reduce UK growth.  This post discusses the effect of pre-Brexit uncertainty, not Brexit itself, which is still years away.

PPS.  I believe the US should withdraw from the IMF.  Not because the IMF is not good enough for the US, rather because America is not good enough for the IMF. If we don’t leave, I hope the other IMF members expel us:

Finance ministers and central bankers from around the world have dropped a pledge to resist protectionism, in a further sign that the new US administration’s stance on trade is shifting the global debate.

The group from International Monetary Fund member countries issued a statement on Saturday saying they would “promote a level playing field in international trade” but did not reiterate a previous commitment to “resist all forms of protectionism”.

The change of stance mirrors a similar move made by the finance ministers and central bank governors of the G20 countries after they met in Baden-Baden in March. On that occasion, the US was unwilling to endorse forthright language on protectionism.





25 Responses to “Most shocks don’t matter”

  1. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    23. April 2017 at 05:12

    The IMF was created at Breton Woods to be the agency for the administration of the fixed exchange rate system. After Nixon it really lost its raison d’etre.

  2. Gravatar of flow5 flow5
    23. April 2017 at 06:46

    The FX market is a black market. “China only appears to lower tariffs. They use regulatory measures and their banking system to accomplish the same economic goals as a high tariff regime”

    “The Chinese government is through the banking system lending money to Chinese companies that export high labor content products to move the production arm offshore. That is why Vietnam and the Philippines are booming. Those are Chinese companies or fronts for Chinese companies. They laugh at us.”

    The Chinese “advantage was his engineering costs. He could put engineers on a project at $20,000 a piece versus the U.S. cost was over a $100,000 including benefits.”

    “Chinese companies that export..that export and use large numbers of engineers have unlimited access to capital borrowing at reasonable rates for long term investment from the Chinese banking system. This is a major competitive advantage. On top of that the educational system turns out increasing numbers of engineers with salaries substantially below the U.S. This, along with the undervalued currency, creates an economic powerhouse.”

  3. Gravatar of ssumner ssumner
    23. April 2017 at 11:45

    Patrick, I agree.

  4. Gravatar of Jerry Brown Jerry Brown
    23. April 2017 at 11:51

    I hope we leave or are expelled from the IMF also. That should free up hundreds of economists to do some kind of productive work. Kind of a small scale supply side reform. Assuming economists can do productive work of course…

  5. Gravatar of Benjamin Cole Benjamin Cole
    23. April 2017 at 16:09

    Excellent blogging.

    Re protectionism: the globalists need to face the issues of exploding house prices in nations with chronic current account deficits, and the results of globalized wages.

    Young buy homes in Mexico and China…but in the US? Not so much….

  6. Gravatar of Benjamin Cole Benjamin Cole
    23. April 2017 at 17:56

    Unsettling question: in the name of enlightened globalism, should the French commit to a future of being a majority Muslim and nonwhite nation?

    This is not only a theoretical scenario; I read that 25% of French schoolchildren are Muslim.

    Do the prerogatives of globalism preclude any neighborhood, ethnic group or nationality from preserving itself?

    If 59 million Han Chinese wish to move to a Thailand or Vietnam, is Thailand or Vietnam obligated to accommodate the migrants?

  7. Gravatar of Benjamin Cole Benjamin Cole
    24. April 2017 at 01:34

    Meanwhile, there is good news is some parts of the world….

    “Shinzo Abe eradicates last hawks from the Bank of Japan

    All 9 BoJ board members now PM’s choices who back aggressive monetary stimulus

    Shinzo Abe has filled the BoJ board with supporters his Abeonomics stimulus programme © Reuters

    APRIL 18, 2017 by: Robin Harding in Tokyo

    Supporters of monetary stimulus will dominate the Bank of Japan after Prime Minister Shinzo Abe nominated two new members to replace the last hawks on its policy board.”


    In the (altered) words of Leonard Cohen, “First we take Tokyo, and then we Washington, D.C.”

  8. Gravatar of Major-Freedom Major-Freedom
    24. April 2017 at 03:19

    Speaking of shocks, the United Nations voted last week to place Saudi Arabia on the Commission on the Status of Women.

    In 2015, Saudi Arabia reduced a Sri Lankan woman’s sentence for adultery from execution by stoning to three years in prison.

    The UN should be abolished

  9. Gravatar of flow5 flow5
    24. April 2017 at 03:45

    “Australia’s property market has been on fire and has been increasing some 17% year-over-year in Sydney alone.”

    That’s because of the FX black market.

  10. Gravatar of Matthew W Waters Matthew W Waters
    24. April 2017 at 05:07

    The 25% number for French school children is completely wrong. It was a study of one school in a region with more Muslims than rest of France. The real percentage is around 13%. The second generation will also be more assimilated than the first generation in general.

    The “globalists” do not believe in completely unrestricted and unregulated immigration. It is a strawman. The only two choices aren’t just completely open borders and nativist isolationism.

  11. Gravatar of Kevin H Kevin H
    24. April 2017 at 08:16

    Does the (non) effects of the Trump victory make you more skeptical of the merits of Higgs’ “regime uncertainty” theory for partially explaining the Great Depression?

  12. Gravatar of Christian List Christian List
    24. April 2017 at 12:57

    That’s hardly a shock, that’s just business as usual. In the “United Nations Human Rights Council” for example you find countries like Egypt, Iraq, Saudi Arabia, Cuba, and China.

    And they are really tough investigators, believe me. For example during all those years they always find the one country that’s the worst human right violator worldwide in every year. Which tiny country on the southeastern shore of the Mediterranean Sea might it be? It’s extremely hard to guess.

  13. Gravatar of Scott Freelander Scott Freelander
    24. April 2017 at 13:49


    I think you’re right. I was always skeptical of the “uncertainty” hand waving that many pundits and politicians would offer as possible outcomes of policy changes.

    To bring up a real shock, I’m curious as to whether you ever studied the 1990-1991 recession. I began studying the US recessions going back to ’82 recently, and ’90-’91 really stands out in the sense that stock and treasury market implicit forecasts were relatively very slow to recognize the degree to which Fed policy would tighten going forward.

    The ’90-’91 recession is the only one I’ve looked at so far in which the drop in the S&P 500 and Treasury yields does not match the drop in NGDP from trend, via predictions from our simple model.

    It’s remarkable how well our model matches data for the other recessions over this period, and the Great Depression.

    Change in S&P 500 = 1 – [1/(1 + change in NGDP from trend x P/E ratio)]

    The model for U3 unemployment works similarly well, which is reduced to the below for sudden shocks(expanded version exists to account for non-shocks):

    U3 = 1 – [1/(1+ U3b + change in NGDP from trend)],

    where U3b is the U3 baseline rate. Mike Sandifer is going to do a blog post on this soon.

    Of course, this is just a rule of thumb model. In reality, we have to model market expectations directly, taking the temporal dimension of expected return as it relates to GDP growth into account. When we do that, the 1990-91 recession looks fine in the model.

    Addtionally, we model the effects of changes in NGDP and monetary policy on gold by treating gold as a negative coupon perpetual bond. The price rises with the prospects for negative real interest rates, as gold increasingly becomes a competitive store of value. Of course gold also has its own supply and demand dynamics.

    Our forex model is still under development, but there’ve been great strides. We’re now including the effects of balance of payments changes and trying to perfect that part of it.

    We’ll be releasing some presentations soon.

    Lastly, looking at these recessions has made me begin to understand your point about being concerned about the Fed overshooting its inflation target. There’s a clear pattern in the data that fits your thinking that periods of overshoot are followed by periods of overly tight money.

  14. Gravatar of Scott Freelander Scott Freelander
    24. April 2017 at 18:45

    Here’s a link to the blog post mentioned above:

  15. Gravatar of George George
    25. April 2017 at 11:41


    problem with this is that it is necessarily ex post and lagging. Nominal GDP gap peaked in Q2 2009, which was reported in July. But for you to know it has peaked, you would need to wait at least one more quarter (typically one more quarter) for Q3 GDP which was reported in late October 2009. Coincidentally the unemployment rate peaked the same month.

    With the stock market it is even worse, by the time your formula told you the size of the decline, sp500 had rallied by over 40%! Everyone knew the worst was over. + Then you have countless revisions etc…

    Nice thought experiment but real time not too useful (perhaps it can help with getting the peak unemployment couple weeks beforehand if you are lucky and there are no big revisions afterwards). What is fascinating though is that the formula implies that very overvalued markets can crash a lot without the economy experiencing huge NGDP deviation from the trend (and thus high unemployment). SO perhaps stock market “bubbles” are really benign even when they deflate big time.

  16. Gravatar of ssumner ssumner
    25. April 2017 at 12:09

    Kevin, I have not read that book, but I am generally skeptical of “uncertainty” theories of the business cycle.

    Scott, Markets are often slow to predict recessions, although they are often ahead of the Fed itself.

  17. Gravatar of George George
    25. April 2017 at 12:22

    Just ran through it, Sp500 2007 top was 1577, 2009 bottom was 666.8 that gives you 57.72% decline not 54%. Haven’t checked other numbers yet.

  18. Gravatar of Jeff G. Jeff G.
    26. April 2017 at 09:39

    Hi Scott,

    I’m wondering if you are thinking of doing a post on how the fiscal theory of the price level have fared through these shocks?

  19. Gravatar of Mike Sandifer Mike Sandifer
    26. April 2017 at 10:25


    Yes, I should apologize for that less than user-friendly post. I thought it’d be okay to be lazy and post it that way, since I was only talking about a rule-of-thumb that works better than I’d expect. I’ll have to create some more comprehensive graphs when I have time to better display the relationships.

    I agree that as a forecasting tool, it’s of limited value, except that inverted yield curves are normally a good clue that one should begin hedging.

    I see what I’m developing as allowing for portfolio stress-testing, helping to measure risk and reward prospects with some precision. Looking forward, one wants to look at implicit market expectations of changes in the NGDP growth path going forward.

    On your second point, it’s certainly possible I made a mistake in identifying the top of the S&P 500 in ’07. I’ll check later when I’m at a computer. If I did, it’s still not far off and the rule of thumb works pretty well. I point out myself that it doesn’t work for the 1990-’91 recession.

  20. Gravatar of George George
    26. April 2017 at 11:13


    regarding the SP500, funnily enough, you made a little mistake but the real numbers are even closer. Take 2007 top value (not highest monthly close) and 2009 lowest value (not the lowest monthly close) and you should get the aforementioned number. Then I realized I was not sure where your P/E came from. Looks like last quarter average ahead of the recession onset. In that case you need to use Q4 2007, not Q3 2007. Q4 P/E was higher due to decelerating/falling revenue and EPS. Hence the final number from your formula is even closer, something like 57.3%.

    And yes I agree that 10yr to FFR spread is a very valuable advance indicator. However bear in mind that the yield curve did not invert in the Eurozone in 2011/12. Unless the US FFR rises significantly, you may no get an inversion the next time around. Right now it is hard for me to see 2%+ FFR anytime soon without deliberately causing a recession. Uncharted territory.

    I think the value of your formula, if robust, is that one could gauge the size of the output and employment contraction based on the stock market. When SP500 bottomed out in March ’09 you could basically get the bottom in GDP and top in unemployment months before they are released.

  21. Gravatar of Mike Sandifer Mike Sandifer
    26. April 2017 at 12:13


    Thanks for the update. As you can tell, I’m not real good at blogging about this stuff yet. I was up late last night, struggling with Excel to produce some better presentations. I hadn’t used Excel so much in a long time. I’m trying to figure out the best way to present the fit to data, both in terms of the rule of thumb and the actual model I’d use if serious, involving definite integrals, etc. I have to admit to being rustier at calculus than I expected. I had only used it sparingly since I learned it over 20 years ago, and then only the occasional simple derivative.

  22. Gravatar of ssumner ssumner
    26. April 2017 at 13:16

    Jeff, I’m not a fan of that theory; I don’t see how it explains anything, at least regarding the US.

  23. Gravatar of Jeff G. Jeff G.
    26. April 2017 at 15:09

    That’s kind of where I was going. Much of the news today signaled higher deficits and yet inflation expectations barely moved. I suppose you could argue today’s news was already priced in but inflation expectations have been slowly falling since the beginning of January.

  24. Gravatar of Major-Freedom Major-Freedom
    27. April 2017 at 03:29

    Christian List, re: Saudi Arabia voted to UN Women’s Rights Commission

    Want to make €10,000?

  25. Gravatar of Scott Sumner Scott Sumner
    29. April 2017 at 07:03

    Jeff, I agree.

Leave a Reply