Saturday, October 3, 2015

So many monetary policy shocks!

Scott Sumner shows the fall in the Euro (vs the Dollar, I believe) after Draghi's announcement to do whatever it takes at the beginning of September from 1.12 to 1.11. Here's a broader view from Bloomberg (September 3rd is marked):

This graph covers the region from about 1.25 to 1.05 -- about 20 times the range of the graph in Scott's post. By the standard of that graph, there appears to be several monetary policy events of equal size each month over the past year.

Another interpretation is that there are people in forex markets who trade on random dupes that try to trade on that "new" information. It doesn't sound like a good basis for a macroeconomic theory. The fall from the Draghi announcement vanishes into the noise in a few days, therefore it probably didn't matter that much.

In general, it appears that markets make "mistakes" -- going in some direction based on macro policy announcements regardless of whether that macro policy will have the outcome it purports to have. This seems particularly significant in forex markets.


  1. I'm glad somebody thought to zoom out.

  2. Replies
    1. Thanks for the alert. Basically, I think this is what is happening:

      But that might have been too confrontational.

    2. Thanks for the link. I think that is the crux of the disagreement... now looking at Sumner's latest response (invoking the EMH and not leaving $100 bills on the table) to your response.

    3. I responded ... I was actually a bit confused. That prices are a random walk is the EMH, so information in past price changes must 'evaporate' ... otherwise it wouldn't be a memory-less process.

    4. Hi Jason, thanks, but now I'm confused. I think I know what you mean by random walk (integrated white noise), and memoryless (like a Markov process, correct? I.e. that a future state's dependence on past states can be summed up completely by the single most recent past state?).

      Are you saying that because the EMH implies memorylessness and a random walk that thus this further implies that the counterfactual (no QE and no QE announcement) would have resulted in an indistinguishable sequence a few months out, not 1% lower? But then in what sense is the QE announcement "priced in" a few months later? I guess I would have thought that "memorylessness" just implied that a few months later, it wouldn't matter what the particular sequence was that got you to that price (i.e. there were many different sets of announcements and price paths that could lead to the same price months away): the only thing that will be important at that time is the actual price then, not it's history. But that seems like a different concept than the counterfactual would be indistinguishable. If a few months out the particular intervening noise sequence (between now and then) were identical on the price process, then it seems like the counterfactual would still be different.

      Also, I don't know much about the EMH or how it relates to memorylessness or random walks, but I read Scott's comment about the EMH as him claiming that the Market is not wrong on average, in the sense that on average, it's impossible to profit from these sorts of announcements by detecting that the market was wrong in it's reaction and betting accordingly. That seemed to tie into your link to your post on the market being wrong... in other words, it seemed to me that you and he were on the same page at least: with you implying that the market can be wrong, and him implying that the market cannot be wrong on average else it would violate the EMH.

      But, as usual, my confidence in my understanding evaporated in the subsequent noise, just like the effects of that QE announcement. Lol.

    5. Hi Tom,

      You asked:

      "Are you saying that because the EMH implies memorylessness and a random walk that thus this further implies that the counterfactual ... would have resulted in an indistinguishable sequence a few months out, not 1% lower?"


      "But then in what sense is the QE announcement 'priced in' a few months later?"

      I was saying the impact of QE was priced in *before* the announcement since the announcement didn't produce something that could be distinguished from the counterfactual. We either a) already knew what impact QE would have or b) the announcement of QE didn't have any effect.

    6. And I don't think you could realistically profit from this market error. The price at time t is (incorrectly) low by δ, but the price at time t+1 is uncertain with a much wider margin than δ.

      It is possible that if these announcements were made frequently enough and you put enough money on it each time, you could average out the noise and raise the SNR of this error. It's a 1% error, so you'd need ~ 10,000 announcements to beat back the noise.


Comments are welcome. Please see the Moderation and comment policy.

Also, try to avoid the use of dollar signs as they interfere with my setup of mathjax. I left it set up that way because I think this is funny for an economics blog. You can use € or £ instead.

Note: Only a member of this blog may post a comment.