Last week, the stock market dove in part because President Trump appeared to be plunging ahead with new tariffs; on Monday, the market recouped that loss (and then some) as the conventional wisdom over the weekend was that Congress would never let that happen and so it is unlikely that tariffs will be implemented.

I’m always fascinated by market behavior around events like this. Investors seem to love to guess right and to put 100% of their bet on an outcome that depends on being right. Here’s what I know about tariffs – prior to last week if there was a risk that tariffs would be implemented that risk was not priced into the markets. And markets are supposed to price risks. Regardless of what you think the probability of that outcome is, surely the probability is non-zero and, therefore, ought to be worth something on the price. Putting it another way: if I was willing to pay X for the market when I wasn’t worried about the possibility of the detrimental effect of future tariffs, then assuredly I will pay less than X once I start to consider that possibility. Although the outcome may be binary (there will be increasing tariffs and decreasing free trade, or there won’t be), the risk doesn’t have to be either/or.

This is one of the things that irritates me about the whole “risk on/risk off” meme. There is no such thing as “risk off.” Risk is ever-present, and an investor’s job is not to guess at which risks will actually present themselves but to efficiently preserve as much upside as possible while protecting against downside risks cheaply. Risk management is really, really important, but it often seems to get overlooked in the ‘storytime’ that 24-hour market news depends on.

To be sure, the risk of tariffs coming out of the Trump Administration is not new…it’s just that it has been ignored completely until now. Right after Trump’s election, in our Quarterly Inflation Outlook, I wrote about which elements of Trump’s professed plans were a risk to steady inflation. The one area which I felt could be the real wildcard leading to higher inflation as a result of policy (as opposed to higher inflation from natural dynamics, which are also a risk as interest rates normalize) was the possibility of a Trump tariff. Here is what I wrote at that time – and it’s poignant today:

Policies Which Will Erect Trade Barriers of Some Kind

This is the area in which we would be most concerned about the possible upward pressure on inflation from a Trump Administration. The President-Elect got to this point partly by pledging to “make better deals” with trading counterparties such as China and to work to “bring jobs back” to the US. While this may be at least partly bluster, Mr. Trump was consistent enough during the campaign on this topic that it is hard to imagine him doing an about-face and strengthening NAFTA, rather than weakening it.

And here is why it matters. We have written in the past that a big part of the reason for the generous growth/inflation tradeoff of the 1990s was the rapid globalization of many industries following the end of the Cold War. Parameterizations of inflation models, in general, cannot be consistently calibrated on any period that spans 1992-1993. That is to say that for any model that we have seen, the parameters if the model is fit to 1972-1992 are different than if the model is fit to 1994-2014. Specifically, models calibrated to the former period consistently over-estimate inflation in the latter period, while models calibrated to the latter period consistently under-estimate inflation in the former period. The Federal Reserve believes that this is because inflation expectations (which we cannot measure very well) somehow became “anchored” in 1993. On the other hand, we believe that the culprit was globalization. In the 2014Q3 QIO, we illustrated that assertion with this chart of Apparel prices, set against domestic apparel production.

The chart (Source Bloomberg) is updated to 2016. We think it illustrates clearly the inflation dividend brought by globalization – as production was moved to cheaper overseas manufacturers, apparel price increases first leveled off, and then actually declined. Prices continued to go sideways or down until apparel production in this country was essentially gone – and thus, there was no further gain in production costs to be passed on to consumers. In late 2012, apparel prices started to rise again, although it has still been only in fits and starts. (We think this is because manufacturing is being moved further downstream, to manufacturers located in even cheaper countries – but this can only go on for so long of course.)

What we haven’t been able to find before, until recently, is more general evidence that there was a dramatic shift in the globalization dynamic in general, rather than in this isolated case. We found the evidence recently in a Deutsche Bank piece, in a chart that plots the number of free trade agreements signed per year. The chart is printed below (sources: as cited).

This chart is the “smoking gun” that supports our version of events, in terms of why the inflation dynamic shifted in the early 1990s. The apparel story is supporting evidence about the next step in the chain, illustrating how free trade helped to restrain prices in certain goods, by allowing the possibility of significant cost savings on production.

The flip side of cost savings on production, though, is a loss of domestic manufacturing jobs; it is this loss that Mr. Trump took productive advantage of. We believe that Mr. Trump is likely to move to increase tariffs and other barriers to trade and to reverse some of the globalization trend that has driven lower prices for the last quarter-century. We view this as potentially very negative news for inflation. While there was some evidence that the globalization dividend was beginning to get ‘tapped out’ as all of the low-hanging fruit had been harvested – and such a development would cause inflation to be higher than otherwise, it would have been – we had not expected the possibility of a reversal of the globalization dividend except as a possible and minor side-effect of tensions with Russia over the Ukraine, or the effect the Syrian refugee problem could have on open borders. The election of Mr. Trump, however, creates the very real possibility that the reversal of this dividend might be a direct consequence of conscious policy choices.