View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

IP1 1k\ aJ
c 1r
IFi f @1mCC
li

caJ

October 30, 1981

Penalty Discount Rate: I
In October 1 979, as part of its program for
bringing down the rate of monetary growth,
and hence inflation, the Federal Reserve
adopted new procedures for controlling the
stock of money. The new procedures control
money from the supply side by manipulating
the quantity of total bank reserves, rather than
from the demand side through interest rates.
Banks legally must hold reserves equal to
fixed percentages of various types of deposits
in the form of either deposits with the Fed or
vault cash. (By "banks," we mean all
depository institutions with transaction
deposits, since such institutions must hold
reserves with the Fed under the terms of the
1 980 Monetary Control Act.) Therefore, by
fixing the total quantity of reserves available
to the banking system, the new operating
procedures allow the Fed to control the total
amount of bank deposits, and hence stock of
money, that banks are able to create.
Two sources of reserves are available to
/ banks. Nonborrowed reserves are owned
outright by the banks and supplied by the Fed
through purchases of securities. Borrowed
reserves are supplied through loans from the
Fed on a temporary basis. The Fed can closely
control the amount of nonborrowed reserves
by means of open market operations in
securities. However, the Fed has only indirect
control over the small but potentially volatile
portion of total bank reserves which banks
borrow from the Fed at the discount window.
These borrowed reserves are created
primarily at the discretion of the borrowing
institutions. Consequently, some analysts
suggest that the precision of monetary control
would be strengthened under the new
operating procedures if the Fed kept the cost
6f reserves at the window (the discount rate)
above the cost of reserves in the private
financial market (the Federal funds rate) so as
to stabilize borrowed reserves at low levels.
As discussed in our next Weekly Letter, the
successful operation of a penalty discount
rate would require a switch to

contemporaneous reserve accounting from
the current system of lagged reserve
accounti'ng, which sets required reserves on
the basis of deposits two weeks earlier.
However, since the Federal Reserve Board of
Governors recently announced that it is
contemplating this change, it is interesting to
assess how well a penalty discount rate might
work if contemporaneous reserve accounting
were introduced.

Rationale
The Fed sets the basic discount rate (now 14
percent), as well as a surcharge (now 2
percent) for frequent borrowing by large
institutions. Banks then decide on the total
amount that they wish to borrow, subject to
the Fed's rules on frequency and duration.
While borrowed reserves constitute only 2 to
3 percent of total reserves on average,
fluctuations in borrowings can contribute
significantly to movements in total reserves at
any particular time. Some analysts thus have
proposed reforming discount-window
operations so that borrowed reserves might
be made more controllable and thus more
predictable. With greater predictability, the
Fed would be able to hit its targets for total
bank reserves more easily by controlling the
non borrowed component.

At present the discount rate tends on average
to fall below market rates of interest, although
the Fed adjusts it up or down in response to
movements in market rates. Moreover, the
spread between these rates tends to be
variable because of the discontinuous nature
of the adjustments in the discount rate, which
result in fluctuations in banks' incentive to
borrow from the Fed. Banks' "reluctance to
borrow" and the System's administrative
procedures tend to hold borrowing in check,
but the actual amount of borrowing varies
positively with the difference between
market rates and the discount rate (see chart).

15)
CE\1ffi CD)
1 1t
k

I f 1©\1fficell
f

CD)
,

Opinions expressed in this newsletter do not
necessarilv refiect the views of the rnanagernent
of the Federal Reserve Ban k of San Francisco,
or of the Board of Covernors of the Federal
Reserve System.
Overall, in the first year of the new operating
procedure, the absolute monthly deviation of
M l-B from the mid-point of its annual target
range averaged 1.40 percent. But M l-B
deviated on average by on Iy .78 percent from
the FOM C's desired
path. So the
Committee's "intentional" deviations were
quite significant.

Tying the discount rate to market rates would
eliminate fluctuations in the incentive to
borrow, and thus would help make
borrowings more predictable. Still, this
would not eliminate all of the variability in
borrowing. To keep such variability to an
absolute minimum, the Fed would also have
to set the discount rate at a penalty level,
somewhat above market interest rates. Banks
without alternative souces of funds for
maldng adjustments in reserve positions
cou Id still do so at reasonable cost. But the
amount of borrowing would be quite small
and therefore considerably more predictable
than it is now because banks would generally
find it less costly to borrow in the private
financial markets than at the window.

Three different types of operating errors also
caused deviations from the FOM C's desired
short-run growth path. For any given month,
these included misses due to 1) errors in the
projected "multiplier" relationship between
the monetary aggregate and total bank
reserves, 2) errors in hitting the targeted path
for nonborrowed reserves, and 3) errors in
,projecting the total amount of
discount-window borrowing. The first two
types of error were both relatively small
. compared to the third in the first year of the
new operating procedure. The average
absolute monthly deviation was .47 percent
for the M 1-B multiplier (in relation to its
predicted value), and was .43 percent for
non borrowed reserves (in relation to the
path for total reserves). In contrast,
the average absolute deviation of borrowed
reserves from projected levels (also as a
percent of the total reserve path) was. 94
percent -twice as large as either of the other
two types of operati ng errors.

Sources control errors
of
Whether a penalty discount rate would
actually help to improve monetary control
depends upon the size of the errors stemming
from the variabi Iity in borrowed reserves, and
also upon the system of reserve requirements
in use. To analyze those factors, we first
examine the major sources of error in
monetary control duringthe initial year of the
Fed's new operating procedure, and then
consider the implications of alternative
reserve-requi rement systems.
A monetary aggregate may deviate from the
mid-point of the annual target range for four
different reasons. The first of these is an
"intentional" deviation -a difference
between the System's targeted short-run path
for the monetary aggregate and the mid-poi nt
of its annual target range. Intentional
deviations occur when the Federal Open
Market Committee (FOM C) instructs the
Trading Desk to supply bank reserves
consistent with money paths that are closer to
the upper or lower bound of the annual target
range. For example, for a time in the spring of
1 980 when M 1-B was below its target range,
the FOM C chose a short-run path close to the
lower bound of the annual target range, to
avoid an excessive interest-rate decline that
could have weakened the dollar and
adversely affected inflationary expectations.

The size of the errors in projecting borrowing
would seem to support the case for a penalty
discount rate. However, if these errors
partially tend to offset the other two types of
errors, monetary control would not
necessari Iy be worsened by borrowed
reserve variability. A significant amount of
offsetting actually occurred, since the sum of
all three operating errors amounted to 1.84
percent -over twice the .78 percent average
deviation of the Ml-B from the FOMC's
short-run path. The major reason was an
inverse relation between errors in projecting
borrowed reserves and errors in predicting
the multiplier. Because of the tendency for
offsetting errors, we cannot tell from the data
alone whether or not the reduction in
2

Percent

$ Billions
3

12
10
8
6

,...

4
2
0
-2

Weekly Averages -4

-1

1979

1980

1981

borrowing errors obtainable through
introducing a penalty discount rate would
actually improve monetary control. Our next
Weekly Letter wi II present some evidence on
whether a penalty rate would be likely to
reduce control errors on balance.

discount rate. Of course, in the extreme this
wou Id be equ ivalent to tyi ng the d iscou nt rate
to current market rates. Unfortunately,
however, there are limits to the feasible
amountof short-run flexibility in the discount
rate under the present system of lagged
reserve accounting, where reserves are
required against deposits outstanding two
weeks earlier. Tying the discount rate to
current market rates, either at a penalty level
or otherwise, with lagged reserve accounting
would tend to produce a short-run ratcheting
in the level of interest rates, because the
amountof reserves desired by banks is almost
completely fixed within any two-week
period. For example, a discount-rate increase
produced by a rise in market rates would
drive up the Federal-funds rate as banks tried
to borrow reserves from each other instead of
the Fed, and this would then produce a
further increase in the discount rate, and so
on.

Alternatives to penalty rate
Even if a penalty rate were not adopted, errors
in monetaiy control arising from variability in
borrowings could be reduced in other
ways-first by making more forward-looking
projections of the level of borrowing. In the
first year under the new operating
procedures, the Fed generally set projected
borrowings each month at a level equal to the
average amount of borrowing in the previous
control period. This procedure generates
reasonably accurate forecasts of borrowing
when interest rates are stable. But the
procedure tends to underestimate borrowing
in periods of rising interest rates, and does the
reverse in periods of falling interest rates.
Although the Fed sometimes revises
borrowing projections within the control
period, errors still have generally been
associated with the direction of changes in
interest rates. However, the Fed can predict
interest rates to some extent, because their
movements flow partly from the adjustments
in total bank reserves required to keep the
monetary aggregates on path. Therefore, a
more forward-looking setting of the projected
borrowing level atthe beginning of a control
period, as well as larger adjustments within it,
might help to reduce borrowing errors.

In summary, we could obtain some reduction
in monetary-control operating errors by using
more forward-looking projections of
borrowed reserves and by maki ng more
frequent adjustments of the discount rate. But
with the present system of Iagged reserve
accounting, the apparently simple solution of
tying the discount rate to current market rates
(ateither a penalty level or otherwise), would
create very large instabilities in interest rates.
Moreover, even under contemporaneous
reserve accounting a tied or penalty rate
might not actually reduce monetary-control
errors by much, or at all, if borrowing errors
tend mainly to offset opposite multiplier
errors. In our next Weekly Letter, we will
detail the changes in reserve requirements
necessary for the successfuI operation of a
penalty discount rate, as well as the likely
resulting errors in monetary control.
JohnP. Juddand Adrian W. Throop

Errors in projected borrowing also could be
reduced by more frequent adjustments in the
discount rate, to keep it more closely in line
with market rates of interest. Borrowing levels
cou Id be made more predictable if there were
smaller fluctuations in the incentive to
borrow, through greater flexibility in the

3

SSV1 :l lSl:Il:I
U018u!4seM.4eln • uo8aJO • epet\aN • o4epi
!!eMeH • e!uJ0J!je:) • euozpv • e>jselV

(\J)

·J!IIl':l10:lSpUIl'J:!
UIl'S
ZS.l·ON IIWH:Jd
OIVd
:J9VIS Od ·s·n
llVW SSV1:lISHI:!

RANKINGDATA-TWELFTHFEDERAL
RESERVE
DISTRICT

.

(Dollaramounts miflions)
in

Selected ssets Liabilities
A
and
LargeCommercljil anks
D .
Loans
(gross,
adjusted) investments*
and
Loans
(gross,
adjusted) total#
Commercial industrial
and
Real
estate
Loans individuals
to
Securities
loans
U.s.Treasury
securities*
Othersecurities*
Demand
dePosits total#
Demand
deposits adjusted
Savings
deposits total
Timedeposits total
#
Individuals, & corp.
part.
(Large
negotiable
CD's)
Weekly Averages
of DailyFigures
r
eserve
POSitlOl1
Excess
Reserves )/Deficiency )
(+
(Borrowings
Netfreereseryes )/Netborrowed( )
(+
-

Amount
Outstanding
10/14/81
153,123
132,238
40,235
54,789
23,164
1,510
5,572
15,313
42,076
29,702
29,532
86,078
78,212
33,865

Weekended

Change
from
yearago
Dollar
Percent

Change
from
10/7/81
-

-

-

-

584
487
40
134
28
277
110
13
48
508
242
123
75
339

-

-

-

Weekended

10/14/81

10/7/81

81
13
68

85
3

82

10,997
12,114
5,152
5,920
722
380
1,029
84
6,328
5,510
406
21,178
21,984
9,635

-

7.7
10.1
14.7
12.1
3.0
33.6
15.6
0.5
13.1
15.6
1.4
32.6
39.1
39.8

Com arable
p
year-ago
period
88
94
7

* Excludes
trading
account
securities.
# Includes
itemsnotshown
separately.
Editorial
comments ay beaddressed the editor(William Burke) to theauthor•..• Free
m
to
or
copieS this
of
andotherFederal
Reserve
publications beobtained calling writingthe Public
can
by
or
Infonnation
Section;
Federal
Reserve of sanrancisco, Box7702, SanFrancisco4.120.Phone(415) 544-2184.
Bank
F
P.O.
9