The difficulties of measuring money arise not only because it is hard to decide what
is the best definition of money, but also because the Fed frequently later revises earlier
estimates of the monetary aggregates by large amounts. There are two reasons why
the Fed revises its figures. First, because small depository institutions need to report
the amounts of their deposits only infrequently, the Fed has to estimate these amounts
until these institutions provide the actual figures at some future date. Second, the
adjustment of the data for seasonal variation is revised substantially as more data
become available. To see why this happens, let’s look at an example of the seasonal
variation of the money data around Christmas-time. The monetary aggregates always
rise around Christmas because of increased spending during the holiday season; the
rise is greater in some years than in others. This means that the factor that adjusts the
data for the seasonal variation due to Christmas must be estimated from several years
of data, and the estimates of this seasonal factor become more precise only as more
data become available. When the data on the monetary aggregates are revised, the seasonal
adjustments often change dramatically from the initial calculation.
Table 2 shows how severe a problem these data revisions can be. It provides the
rates of money growth from one-month periods calculated from initial estimates of
the M2 monetary aggregate, along with the rates of money growth calculated from a
major revision of the M2 numbers published in February 2003. As the table shows,
for one-month periods the initial versus the revised data can give a different picture
of what is happening to monetary policy. For January 2003, for example, the initial
data indicated that the growth rate of M2 at an annual rate was 2.2%, whereas the
revised data indicate a much higher growth rate of 5.4%.
A distinctive characteristic shown in Table 2 is that the differences between the
initial and revised M2 series tend to cancel out. You can see this by looking at the last
row of the table, which shows the average rate of M2 growth for the two series and
the average difference between them. The average M2 growth for the initial calculation
of M2 is 6.5%, and the revised number is 6.5%, a difference of 0.0%. The conclusion
we can draw is that the initial data on the monetary aggregates reported by
the Fed are not a reliable guide to what is happening to short-run movements in the
money supply, such as the one-month growth rates. However, the initial money data
are reasonably reliable for longer periods, such as a year. The moral is that we probably
should not pay much attention to short-run movements in the money supply
numbers, but should be concerned only with longer-run movements.
is the best definition of money, but also because the Fed frequently later revises earlier
estimates of the monetary aggregates by large amounts. There are two reasons why
the Fed revises its figures. First, because small depository institutions need to report
the amounts of their deposits only infrequently, the Fed has to estimate these amounts
until these institutions provide the actual figures at some future date. Second, the
adjustment of the data for seasonal variation is revised substantially as more data
become available. To see why this happens, let’s look at an example of the seasonal
variation of the money data around Christmas-time. The monetary aggregates always
rise around Christmas because of increased spending during the holiday season; the
rise is greater in some years than in others. This means that the factor that adjusts the
data for the seasonal variation due to Christmas must be estimated from several years
of data, and the estimates of this seasonal factor become more precise only as more
data become available. When the data on the monetary aggregates are revised, the seasonal
adjustments often change dramatically from the initial calculation.
Table 2 shows how severe a problem these data revisions can be. It provides the
rates of money growth from one-month periods calculated from initial estimates of
the M2 monetary aggregate, along with the rates of money growth calculated from a
major revision of the M2 numbers published in February 2003. As the table shows,
for one-month periods the initial versus the revised data can give a different picture
of what is happening to monetary policy. For January 2003, for example, the initial
data indicated that the growth rate of M2 at an annual rate was 2.2%, whereas the
revised data indicate a much higher growth rate of 5.4%.
A distinctive characteristic shown in Table 2 is that the differences between the
initial and revised M2 series tend to cancel out. You can see this by looking at the last
row of the table, which shows the average rate of M2 growth for the two series and
the average difference between them. The average M2 growth for the initial calculation
of M2 is 6.5%, and the revised number is 6.5%, a difference of 0.0%. The conclusion
we can draw is that the initial data on the monetary aggregates reported by
the Fed are not a reliable guide to what is happening to short-run movements in the
money supply, such as the one-month growth rates. However, the initial money data
are reasonably reliable for longer periods, such as a year. The moral is that we probably
should not pay much attention to short-run movements in the money supply
numbers, but should be concerned only with longer-run movements.
No comments:
Post a Comment