« Bernanke confirmation hearings | Main | How much do we know about China? »
November 16, 2005
What's going on with non-M2 components of M3?
The discontinuance of M3 has been getting attention, notably from The Prudent Investor and The Big Picture. Mark Thoma and David Altig also mentioned it. Altig, for his part, is not too worried. He looks at the percentage changes in M1, M2, and M3 together.
Let's probe further. Here's a chart, similar to Altig's, but showing only M2 and M3 (M1 behaves a lot differently) just over the last 20 years (a period in which the two series have varied in a fairly stable manner).
Yep. They move together. Now, there looks like there could be a lead/lag relationship that asserts itself from time to time. Might be worth looking at. There is also a small gap that appears to be opening up. That bears watching.
But it would be useful to see if any particular non-M2 component of M3 is driving things. Consider the following chart:
Large time deposits are red "ltd" and repurchase agreements are in black "rp". Eurodollars are in blue. Money market mutual funds in M3 are in green.
First of all, note that some of these components are pretty small, and most are quite volatile. Raw data is from H.6 historical tables from the Fed.
Repurchase agreements jumped in 2002 and 2003, but that coincides with falling M2 and M3. It is possible, of course, that policy lags come into play. However, at the moment, it appears that if there is a divergence in M2 and M3, it's large time deposits causing the divergence in the data. Interesting, but not sinister.
I'm never really happy to see a data series disappear. Continuity of data is important to researchers. But at the same time, I see no evidence that they are covering up something going on with repurchase agreements. After all, system open market holdings, as well as temporary and permanent operations are available on a daily and a historical basis from the New York Fed. RPs along with other factors affecting reserve balances are also published weekly in the H.4.1 release. Nothing says that this will change, so I am under the impression that you will still be able to find the data there.
UPDATE: A commenter asks why I think they are discontinuing M3 (a question I did not address directly above). The truth is, of course, I don't know. As I said above, I'm pretty sure that it's not to try to hide anything, particularly anything nefarious going on with repurchase agreements. If I had to speculate, I'd say it was probably determined that the value of that particular formulation of a monetary aggregate to the Fed was no longer was worth the cost of producing it on a monthly basis and benchmarking it annually.
This has happened before, you know.
I offer the dedicated reader three papers that describe the monetary aggregates from a historical perspective. These papers detail, among other things, the old monetary aggregates (M4, M5, ...) that are no longer reported. As a research point, I think it would be interesting to see some investigation into the behavior of the non-M2 M3 components as I described above. If there is anything useful in there, let's focus on it, and then the world will little remember the precise definition of M3, just like we no longer remember what was in M4 or M5.
The papers:
Walter, John. Monetary Aggregates: A User's Guide, Richmond Fed Economic Review, Jan/Feb 1989.
One from the 70's, one from the '80s, and one from the '90s. Read them together and you get a very nice historical overview of the evolution of the aggregates we have today. Maybe the time is right for one from the '00s to explain why M3 is no longer worth publishing. I, for one, would read such a paper with interest.
Posted by William Polley at November 16, 2005 12:23 AM
Trackback Pings
TrackBack URL for this entry:
http://www.williampolley.com/cgi-bin/mt-tb.cgi/435
Listed below are links to weblogs that reference What's going on with non-M2 components of M3?:
» Blogosfera from De Gustibus Non Est Disputandum
* Por que ainda sou evolucionista? * Olha o Hal Varian ali! Sim, ele mesmo. * Pergunta de prova: Utilidade é comparável interpessoalmente? Se não, explique. Todo aluno fala bobagem aqui. Mas uma resposta correta, bem construída e esclarecedora, é... [Read More]
Tracked on November 17, 2005 2:54 AM
Comments
Why do you think they decided to stop publishing the data?
Posted by: Mark Sullivan at November 16, 2005 1:19 PM
Mark,
See my update above.
Posted by: William Polley at November 16, 2005 4:45 PM
Thank you for all those links and that summary William.
Posted by: calmo at November 18, 2005 2:24 PM
Mr. Polley,
Thanks you for the thoughtful and well-researched post. As long as the RP data remains available, then it seems we will continue to have real-time transparency as to the value of the USdollar.
I'm still not at all sanguine about this situation, particularly in light of the cost of servicing the federal debt and the impending arrival of the new Fed chairman.
Again, thanks!
Posted by: Idaho_Spud at November 19, 2005 4:05 AM
Mr. Polley,
Thanks you for the thoughtful and well-researched post. As long as the RP data remains available, then it seems we will continue to have real-time transparency as to the value of the USdollar.
I'm still not at all sanguine about this situation, particularly in light of the cost of servicing the federal debt and the impending arrival of the new Fed chairman.
Again, thanks!
Posted by: Idaho_Spud at November 19, 2005 5:07 AM