The SNB’s currency failure in 2015 proves Fed can’t do h money

Half of wrong macro commentary is just people getting excited about words and phrases and forgetting that those words and phrases are supposed to relate to something stable.

In choosing the title of this post, I am intentionally trying to sound ridiculous, to grab your attention. But please hear me out.

Yesterday I ribbed some of the talking heads of Wall Street for abusing the idea of “sigmas” and for finding 4+ sigma events everywhere they look.  In my view, they fail to realize that sigma has no meaning without reference to a distribution, which they unwittingly assume is N.

If you see a lot of 4+ sigma events, you may be assuming the wrong distribution.  Sure you can compute sigma without reference to a distribution, but you cannot thus assign meaning to it.  That people get excited about words or phrases and then forget what they mean is a strongly recurring theme in my work.  Half of wrong macro commentary is just this, in my view.

My favorite example of this was somebody putting around the chat rooms that the euro-swiss move after the SNB gave up was a 14-sgima event.  Because efforts to peg exchange rates never fail!

I mentioned as an aside that that “peg” breaking was actually predictable as a possibility, even though the conventional wisdom – at least among a lot of economists – was that no central bank could ever have trouble holding its currency down.  After all, the central bank could just print up the funding to intervene against their currency without limit, and without fear of inflation, because what they are presumably fighting is deflationary pressure.

That view turned out, predictably, to be wrong.  And I think it provides a rare indirect empirical test of a related bad idea, that central banks can nudge inflation desirably higher with helicopter money.  The parallel is closer than you may realize, but I would say the practical importance of this post is zero in the short-run and perhaps pretty low in the long-run too, because h money is not really a hot topic anymore.  The idea that central banks have “unlimited” balance sheet power has taken quite the hit in the past couple years.

Let me start by discussing the problem what the Swiss National Bank faced back in late 2014 and early 2015.  Market forces were inclining CHF to rally against EUR, because the EUR was generally under pressure and CHF was viewed a safe haven. This was a problem for the SNB because Switzerland was experiencing deflation and the SNB did not believe it had room to cut its policy interest rate much below zero.

The solution was believed to be for the SNB to use its “unlimited” ability to produce base money to intervene against CHF in a battle with market forces. And a lot of people figured that the SNB could – on purely logical grounds – not fail to win this battle.  A central bank’s power to prevent depreciations was limited, as the British had shown in the early 1990s.  But a central bank trying to hold its currency down held all the aces.  That premise turned out to be false and in a way that holds a lesson for the overconfident advocates of h money, the few that remain.

In my view, the first thing you need to do to recognize what is actually going on in a situation like this is to recognize that the central bank has an inflation objective.  It does not matter to the logic of the argument at all what that inflation objective is (beyond tweaking the arithmetic slightly).  All that matters is that the central bank has an inflation objective that is exogenous to the instruments it chooses to pursue that objective.

For the purposes of demonstrating the argument, it is easiest to assume that the inflation objective is a single number or narrow range, say 1-2% on average over the longer haul.  So I will do that. But the argument should (I think) carry through qualitatively even if you were to assume that the central bank believed that some policy instruments require allowing a little slippage on the inflation objective even in the long run.

For example, the SNB might say, ok if we are going to use the currency intervention as our main short-run anti-deflationary tool, then I guess we are going to have to tolerate a slightly higher inflation objective, like 3%, to allow that to work.  That would make the analysis more complicated and nuanced, so I will leave that to others. But it would not change the basic point.

So let’s say that the SNB has an inflation objective of precisely 1.5% a year on average over the long haul.  Presumably, they need to consider then the implications of a policy that would work at driving inflation above zero and towards that objective.  It is important that they cannot assume money market conditions consistent with deflation as they assess the implications of achieving their objective. When the policy works, they get all the effects of it working, which is where the trouble with unlimited money-financed forex intervention starts.

Let’s assume the SNB is looking out to success over ten years.  In the next ten years, the general price level is going to rise a cumulative 16%.  The escape from deflation will presumably allow real output growth to rise more quickly, on the standard Keynesian logic, which means the nominal GDP will be, say, 40% higher.

You can use whatever numbers you like and assume whatever probability distribution you want, just so long as that distribution has a mean that the central bank perceives, even if wrongly.   The point is that the central bank has in mind some level of nominal GDP ten years forward.

There are two important implications of which.  The first is that the path of nominal GDP should be strong enough to pull the economy out of liquidity trap. Second, and conditional on the first, the escape from liquidity trap implies a unique an exogenous demand volume of demand for the non-interest bearing liabilities of the central bank, which are mostly central banks notes or “currency.”

The key point here is that the central bank does not get to pick both its inflation objective and the forward demand (away from liquidity trap) for its non-interest-bearing liabilities.  Accordingly, in long-run equilibrium, from which the central bank must backwards solve, the firepower of the SNB is indeed limited.  (The short-run disequilibrium conditions are irrelevant to the central bank’s ultimate financial position, which people definitely have trouble internalizing.)

Given that the SNB has an inflation objective, roughly irrespective of what it is, and given further that their firepower is limited, the decision to intervene against their own currency represents a speculation that the real foreign exchange value of the CHF in the spot and forward markets is too strong.  That speculation might be right and it might be wrong. But it is certainly a speculation that has the potential to generate huge opportunity costs or outright losses for Swiss taxpayers in the event that the SNB turns out to be wrong.

Just before the “peg” broke in early 2015, the SNB apparently and predictably (at least as a possibility) decided that their confidence in their speculation was too low to allow the size of the related bet to continue mounting rapidly. So they were forced to give up, just like the BoE had to in 1992.

The macroeconomic dynamics of the path to that result and its aftermath were quite different, in Switzerland compared with Britain. So in many ways these two outcomes are not comparable and it is probably fair to say that the SNB had more policy room than the BoE did, largely for that reason.  But contrary to what you may have heard, the SNB’s power was limited, even though they were trying to depress their currency and even though inflation was not a proximate threat.

This has important lessons for – or really against – the idea of helicopter money in the sense that in both cases, we need to recognize that the central bank has a long-term objective for the inflation rate, that this implies or indeed imposes an exogenous path for the long-run demand for its non-interest bearing liabilities, and that intervention can never be “free”, whether that intervention be in the forex markets directly or in the real economy via, say, a money-financed infrastructure spending program.

I won’t elaborate on that here because I have already written far too much on h money, and if you can take the pain you are fully welcome to go read my backgrounder.  My point here is that the SNB episode is actually an indirect empirical test of the central premise behind h money, which failed spectacularly.  Switzerland did not fail spectacularly, just the premise.