From an email from Dennis E. Farley from the FRB's Division of Monetary Affairs:
M3 does not appear to convey any additional information about economic activity that is not already embodied in M2. Academic papers have occasionally used M3 in empirical work, but these studies have not concluded that M3 is an important financial indicator.I have listed the components of EU and US monetary aggregates below.
In addition, the role of M3 in the policy process has diminished greatly over time. M3 is not closely tracked by policy makers nor is it routinely analyzied by Federal Reserve System staff.
Consequently, the costs of collecting the data and constructing and publishing M3 now appear to outweigh the benefits.
You also mentioned that the ECB looks at M3. It is hard to do a straight mapping from euro-area M3 to a U.S. monetary aggregate. If we tried to do a mapping, euro-area M3 would be somewhere between M2 and M3 in the U.S., with a few components not in U.S. M3. Two components of euro-area M3 that are not in their M2 (repurchase agreements, debt securities issued with maturity up to 2 years) are pretty small compared with other parts of their M3. Money market funds are sizable, but we are not eliminating them from our H.6 release. In a sense then, euro-area M3 has a great deal of overlap with U.S. M2.
= M1 (currency in circulation + overnight cash deposits + demand deposit balances + travelers cheques + checking accounts)
+ M2 (time and savings deposits below $100,000 + retail money market accounts)
+ time deposits above $100,000 + repurchase agreements + institutional money market accounts + Eurodollars held by US residents.
=M1 (currency in circulation + overnight deposits)
+M2 (deposits with agreed maturity up to 2 years + deposits redeemable up to 3 months' notice)
+ repurchase agreements + money market funds and papers + debt securities issued up to 2 years.
All in all the reasons given have a more rational basis than the de-emphasization argument Barry Ritholtz was fed and the cost argument The Capital Spectator got when researching the matter.
Now I need to find the time to find some reading material about the fact that the share of M3 has risen significantly over the years as the chart done by Mark Thoma shows.
GRAPH: Mark Thoma, Economist's View.
So far I found one speech by former Fed governor Laurence Meyer who compared the Fed's monetary policy with the ECB's two-pillar strategy in 2001:
"The first pillar is a reference value for money growth. The ECB sets a reference value for a single monetary aggregate, the M3 definition that is essentially the same as the M2 definition for the United States. The ECB reference value is the rate of M3 growth consistent with achieving its inflation target over an intermediate term, based on estimates of trend growth in potential output and velocity. The second pillar considers the appropriate setting for the policy rate in terms of the wide range of information available and the prospect for inflation over the medium term.
The ECB rationale for the reference value for M3 is the long-run stable relationship between its rate of growth and inflation. The reference value provides a second check for policymakers to ensure that monetary policy, set in terms of the ECB's policy rate and in consideration of pillar 2, is consistent with price stability. The ECB is very explicit about the fact that, in light of the short-term volatility of velocity, short-run deviations of money growth from the reference value might provide little useful information that would help policymakers adjust the stance of monetary policy. But in light of the more stable longer-term relationship, continued deviations would raise significant questions and should, at the least, require a careful reassessment of whether the prevailing monetary policy is consistent with the inflation objective.
He outlined why the Fed is following an low-inflation policy without a money supply target as a second check, questions en vogue again in the light of the discussion whether the Fed should aim for an inflation target.
"The ECB uses the term "reference value" rather than a target to make clear that deviations from the reference value will not necessarily result in policy adjustments to encourage a return of money growth to the reference value. Each year the ECB updates its estimate for potential output growth and, if necessary, updates the reference value to ensure that it is lined up on the inflation target.The learning curve continues.
The ECB approach to the reference value for M3 is very close to the way in which the Fed was setting its benchmark range for M2 until the recent revision to the Federal Reserve Act. The major differences are that the Fed was perhaps somewhat less transparent about how it derived the range for M2 and did not update it regularly to maintain an estimated consistency with an unchanged trend inflation rate objective. At any rate, the recent change in the Federal Reserve Act removed the requirement that the Federal Reserve report to the Congress on growth ranges for M2 and other money and credit aggregates. My final topic is whether setting a reference value for money growth would be constructive for the FOMC and, if so, how to implement such an approach.
To move in this direction would have the advantage of allowing money growth once again to play a role as a failsafe, or second check, on the consistency of monetary policy with FOMC's medium-term inflation objective. On the other hand, moving in this direction would require other significant changes in the conduct of policy. The FOMC - presumably in consultation with the Congress - would have to establish an explicit inflation target and would have to reveal its estimate of the rate of growth in potential output. This direction would itself be even a more significant step than setting a reference value for money growth. An intermediate approach might be to set a reference value based on implicit assumptions about both the target inflation rate and the rate of growth of potential output - without explicitly identifying either. This would be similar to how the benchmark range was set for M2 in the last few years before the benchmark ranges for the monetary aggregates were abandoned."