Scott Sumner  

Never reason from an interest rate spread

Hurray for the Imminent End of... Posner and Weyl on Sponsoring ...

Here's the Financial Times:

Analysts have struggled to explain why the yen is rising when the gap between US and Japanese interest rates is widening further, with the US Federal Reserve tightening policy while the BoJ is pinning 10-year bond yields at zero.

Possible answers include a belief that the BoJ will soon tighten policy, reduced confidence in the dollar because of the rising US budget deficit, or technical factors relating to the end of the Japanese fiscal year in March.

That final suggestion is interesting, as massive US budget deficits are the standard explanation of the strong dollar of 1983-85.

When thinking about the relationship between any two macro variables, I always start with the extreme cases, much like physicists like to do experiments under extremes of heat and pressure.  It exposes a lot of relationships that are more difficult to see under ordinary conditions.  Let's start with recent trends in Argentine interest rates:

Screen Shot 2018-02-16 at 11.15.26 AM.png
So Argentine interest rates are far higher than US rates, and trending upward.  How would you expect that to impact the value of the Argentine peso? Over that same period, the peso has fallen from 20 cents to 5 cents:

Screen Shot 2018-02-16 at 11.17.56 AM.png
Now you might argue that I've cherry picked the Argentine example.  But actually this example is typical of countries with very high interest rates.  Doesn't it seem plausible that if very high interest rates cause a currency to depreciate sharply, then modestly higher interest rates might cause a currency to depreciate modestly?

Of course this is one of those "never reason from a price change" cases.  Yes, there really are cases where rising rates in the US are associated with a stronger dollar---the FT is not stupid.  The actual correlation entirely depends on what causes rates to change in the US, and in Japan.

Notice that the FT article cites a "tightening" of monetary policy in the US.  But they provide no data to support their claim that monetary policy has become tighter.  I'll provide data for the opposite claim.  NGDP prediction markets show 4.7% growth up through Q1, and 4.4% expected growth over the next 12 months.  Those figures are a bit higher than recent trends, which suggests to me that the Fed has been easing monetary policy.

Is there any data suggesting that the Fed is tightening monetary policy?  (And please, don't even go there . . . )

Their comments about expected Japanese tightening are plausible.  It would be crazy for the BOJ to tighten right now, but weirder things have happened in central banking.  It would probably help to look at the yen vs. the euro, and other exchange rates, to get a sense of whether this recent move in the exchange rate is more about the dollar or more about the yen.

Comments and Sharing

COMMENTS (4 to date)
Alec Fahrin writes:


This FT article’s juvenile analysis of interest rates relationship to exchange rates is widespread in the economic news community.

There is a lack of understanding about second and third order effects as well. Many people simply don’t understand that interest rates are a reflection of other factors. Although the causal flow can sometimes reverse, interest rates are almost always determined by real economic changes.

That is why nations like Russia, Brazil, and Argentina can raise interest rates 10% and still see their currencies weaken. Likewise, the interest rates are decreased (in Brazil and Russia) and the currencies still appreciate.

A real concern here is what this widespread ignorance of the “don’t reason from a price change” principle shows. These “analysts” influence investors and political opinion, and they are often wrong.

A final thought, you point to rising US nominal GDP growth rates as evidence of a lack of monetary tightening.
Well, Japan, China, and the EU all had their nominal GDP growth rise about 50% relative to 2016. The US, only 35%.
So is it actually just a growth boom in the World Economy?

Kirill writes:

Scott, may i ask u to be more specific about what exactly in yr opinion constitutes tight/loose monetary policy that is attributable to central bank.. as ngdp seems to me rather the outcome of the policy rather than monetary policy itself.. ft is specific.. they cite higher rates at which the fed does borrowing/lending and lower rates targeted by the boj as a specific actions/policies that cud be attributed to cb policies.. ngdp projections as i said look more like the outcomes of higher/lower rates and other policy measures by the cb's.. so what are these measures that make u think that us monetary policy is loose? And i suspect these measures shud be more important than tightening via higher interest rates channel already expected by the market

Scott Sumner writes:

Alec, Yes, monetary policy is becoming more expansionary in many countries.

Kirill, Higher interest rates are also an outcome of policy. If higher rates represented tighter money, then Argentina's monetary policy would be contractionary in recent years. America's policy would have been contractionary during the 1970s. No one seriously believes that.

Like Ben Bernanke, I believe that NGDP is one of the most useful measures of the stance of monetary policy. There are of course alternative measures, some more useful, some less useful.

Kirill writes:

Scott, this is exactly true.. higher rates are the outcome of specific policy measure - sitting together at the fed hradquarters and ultimately after a long heated debate deciding to increase the price of borrowing funds from the central bank.. and in this sense higher rates are the outcome of the policy and maybe reliably attributed to the actions by cb.. but what are specific loosening policy measures by cb that are more powerful than this specific tightening policy measure (raising rates) to be reflected in higher ngdp noy lower ngdp?


Return to top