Step #3 In My Dispute With Beckworth: Moving From Davie on Interest Rates to David on Interest Rates
First, a recap:
==> In Step #1 I made the obvious point that if an employer in late October says that the paycheck will be $1000 higher in November than previously expected, and the paycheck in December will be $700 higher than previously expected, that clearly this is a more generous policy, relative to the path expected the moment before the announcement. The fact that workers now expect a $300 pay cut in December, relative to November, is misleading and irrelevant. What matters is that the workers are getting more money in each month relative to the original baseline. The month-to-month movements in the new path per se aren’t important when deciding whether the information indicates a “looser” vs. “tighter” pay policy. Nobody gave me any argument here.
==> In Step #2, we tweaked the example so it was dealing with the employer’s interest rate policy, rather than paychecks. By analogy with wages, I argued that the important issue was whether interest rates were lower than the originally expected level in each unit of time (either month or the year as a whole). The fact that workers might expect interest rates to rise more quickly later in the year is misleading and irrelevant. What matters is that the workers are paying lower interest rates in each month (or year for the whole) relative to the original baseline.
After some clarification, again, nobody gave me any argument here, in the hypothetical scenario as I constructed it. To be sure, David R. Henderson was getting ready to say that interest rates aren’t a good indicator of monetary policy, but even he conceded that *if* we are going to say “lower interest rates means looser money” and “higher rates means tighter money” then the correct baseline is to look at the originally expected path. It is clearly nonsensical to merely look at the change down the road, if that change still puts us at a lower interest rate. That would be analogous to our Step #1, when Davie argued that the $300 wage cut in December meant he was poorer, even though he actually was making $700 more than he originally expected.
==> Now we are finally at Step #3. (And sorry for the delay, I got sick during my holiday traveling and missed a window when I thought I’d knock out this post.) I am going to argue that David Beckworth in the real world has made a mistake very similar to what Davie in Step #2 committed.
Specifically, Beckworth back on December 10 made two empirical arguments to defend Ted Cruz’s claim that the Fed tightened monetary policy in the summer of 2008. In this post (Step #3), I will tackle the first of Beckworth’s empirical arguments. In the final Step #4, I will tackle Beckworth’s second empirical argument.
Now then, Beckworth’s first piece of evidence involved the following text and graph:
The [figure below] shows the 1-year treasury interest rate minus the 1-month treasury interest rate. Standard interest rate theory tells us that this spread equals the expected average short-term rate over the next year. That is, if the spread goes up in value then the market is expecting the short-term treasury rate to rise over the next year and vice versa. This figure shows a sustained surge in the expected short-term interest rate over the next year from April to November 2008. It especially intensifies in the second half of the year. Only in December does the spread really begin to fall. For most of the year, then, the market increasingly expected a tightening of policy going forward.
Now does everyone see how that is exactly analogous to Davie’s argument in my hypothetical scenario in Step #2? Specifically, Beckworth has shown–correctly–that the future increase in short-term rates expected by the market went up sharply during the fall of 2008. (It’s not as clear to me what happened before the fall, since the lines bounce violently up and down.) Beckworth seems to believe that his demonstration was the same thing as demonstrating that the market expected higher short-term rates later in the year, compared to the originally expected path.
But we know that in general, those two claims are not equivalent. That was the whole point of my (obviously contrived) scenario in Step #2; I wanted to make it crystal clear that in principle, those two things are not equivalent. If you have lost me here, I urge you to go back to Step #2 and look at the two graphics boxes, and see the distinction. Specifically, the market can simultaneously (A) expect rates to increase more during the year than previously expected and (B) expect future short-term rates to be lower than previously expected.
How is this possible? Simple: If all expected short-term rates drop for the next year, but the closer ones drop more than the further distant ones. Then you get the result that (A) the market expects rates to climb more during the year than it expected as of yesterday and (B) the market expects rates in, say, 8 months to be lower than it expected as of yesterday.
So why do I think that in the real world, this is what happened with Beckworth’s numbers? Well, in the chart below I decompose the spread. That is, I show not just the spread between the 1-year and 1-month Treasury bill yields–which is the red line in Beckworth’s chart above–but I also show the actual values of the 1-year and the 1-month yields:
However, my FRED chart has more information than Beckworth’s. In particular, the spike in the green line in September occurred because short-rates dropped (my red line) sharply, while one-year rates (my blue line) only dropped modestly. So since one-month rates collapsed while one-year rates fell more gently, I think this is similar to what happened in my contrived Step #2 example. Yes, the spread widened and we can interpret that as saying, “As of September 2008, the market expected one-month rates to rise more rapidly over the next 12 months than the market expected as of August 2008,” but that’s not the same thing as saying, “As of September 2008, the market expected the one-month Treasury yield in August 2009 to be higher than the market expected as of August 2008.”
==> In my Step #4–the final in this series–I’ll deal with Beckworth’s second empirical argument, and in so doing I will also shore up my position for this Step #3. Specifically, we will look at futures contracts to back out actual implied market forecasts of certain types of short-term rates in future months, at various points along 2008.
==> Last point for purists: Even my FRED graph is consistent with the claim that the Fed led markets to believe it would tighten and/or was actually tightening, from about March 2008 through June 2008. (That’s because the blue and red lines are both rising during this period, meaning that short rates were rising and the market thought future short rates would follow suit.) But that pattern turned around in June, when the blue line began its gradual fall for the rest of the year. So this is the exact opposite of what Ted Cruz needs for his version of history to make sense. In other words, if you want to argue that the Fed was giving winks to the markets to lead them to think tighter policy was in the works from March to June 2008, but then in June the Fed said, “Ha ha fooled you, actually we’re loosening,” then my FRED chart is consistent with that narrative. Yet that’s the exact opposite of what Cruz said happened, and thus the opposite of what David Beckworth said he was explaining with the Treasury yield curve data.