Re: Reengineering the Fed

[Basil Moore <bmoore@xxxxxxxxxxxx>]

Per

Very well put. I agree wholeheartedly with everything you say. I was just
trying to put forward the expectations theory as a sort of tautology for
equating holding period yields. The key problem, which you highlight, is
that, like all neoclassical theorems, it assumes perfect information.

In your view consols become like gold: The expected return is primarily the
expected capital gain or loss. I agree. But, unlike gold, there is also a
pecuniary income stream at the root of their valuation. If the market
believes short term rates will be raised in the future, long term bond
prices will fall, today. I don't have an answer here. Does anyone one the net???

Basil Moore










t 02:18 PM 10/7/97 +0200, you wrote:
>In reply to Basil Moore:
>
>Thanks for another intriguing post. I would like to add a little to your
>exposition about long-term rates, since I fear that you are thinking in
>terms of "buying for keeps", which is not really relevant to the speculative
>world we live in.
>
>SHORT-TERM AND LONG-TERM RATES
>
>The long-term rate r(L), as you conceive it here, is defined by the ratio of
>the annual coupon AC and the present value PV of a perpetuity:
>
>r(L) = AC/PV
>
>You suggest that r(L) depends on expectations about the r(S) at some
>(distant) future date. In the perpetuity case, this expectation must be
>taken to refer to some _infinitely_ distant future date. Clearly,
>expectations are not of this kind. The distant future is so vague and
>uncertain that we cannot possibly conceive any expected value at all -- the
>subjective probability distributions will not even be defined for such
>distant dates.
>
>What we should take into account is not the ultimate r(S) at the maturity
>date, but the _speculative_ changes in PV due to near-future changes in
>r(L). The relevant magnitudes for comparison and portfolio "equilibrium"
>between short and long paper is not r(L) versus r(S), but the _total yield_
>of long paper versus r(S) -- the latter is relevant because we may overlook
>the very minor "lever" between short rates and PVs of short paper.
>
>The speculative "lever" is however highly relevant for long paper, which
>means that we must add to _direct yield_ DY = r(L) the capital gains CG =
>dPV/dt  to obtain the total yield TY:
>
>TY = DY + CG = r(L) + CG
>
>The expected total yield should be set equal to r(S) to ensure "portfolio
>equilibrium":
>
>r(L) + CG = r(S)
>
>Now, the capital gains will depend on the _rate of change_ in r(L), in an
>inversely quadratic fashion:
>
>dPV/dr(L) = - AC / r(L)^2
>
>which yields the well-known elasticity
>
>e[PV; r(L)] = [dPV/dr(L)] [r(L)/PV] =
>
>= - [AC / r(L)^2] [r(L) / PV] = - [PV / r(L)] [r(L) / PV] = -1
>
>A one-percent increase in r(L), from say 5.00% to 5.05% will reduce the PV
>by one percent, a loss (insofar as it is expected to occur during the year)
>which must be compensated by the level of r(L) exceeding the level of r(S)
>by one percentage-point. In this case, thus, r(S) must be at 4.00 percent to
>ensure "portfolio equilibrium".
>
>The expectations about what is going to happen to r(L) five years or ten
>years hence are basically irrelevant in a speculative setting. What really
>matters is how r(L) is expected to develop the next few months, or weeks or
>days, the closer time-perspective being increasingly more relevant than the
>distant, the less the market agents tend to buy "for keeps". Today, hardly
>anything is bought "for keeps", and the speculative horizon has become very
>short indeed.
>
>The connection between long rates and the "underlying" AC (or the net
>profits as in your example) is very loose nowadays, so loose that I doubt
>there is any use at all in reasoning "as if" there existed a firm relation.
>You are surely right in saying that the goal for monetary policies should be
>to keep (real) rates low and to stabilise PVs, but I doubt that your
>"long-perspective" mode of reasoning will be of much guidance for actual
>policies.
>
>Best,
>Per
>
>
>Per Gunnar Berglund
>Lilla Sallskapets vag 60
>127 61  SKARHOLMEN
>SWEDEN
>
>Voice/fax +46-(0)8-883065
>
>
>
>
>-----Original Message-----
>From: Basil Moore <bmoore@xxxxxxxxxxxx>
>To: POST-KEYNESIAN THOUGHT <pkt@xxxxxxxxxxxxxxxx>
>Date: den 7 oktober 1997 01:35
>Subject: Re: Reengineering the Fed
>
>
>
>Low interest rates are it is true the essence of Keynesian monetary policy.
>The CB directly controls the nominal ST rate. Since the inflation rate is
>predetermined in the short run, It is also realistic to say that the CB
>directly controls the real ST rate.
>
>But the CB cannot directly control the LT rate. On the expectations theory,
>LT rates are a weighted average of expected future ST rates, so as to
>achieve equal expected holding period returns on assets of diferent
>maturity, plus or minus some term premium. It is impossible to refute the
>expectations theory empirically, since expected future rates cannot be
>measured directly. But the expectations theory is a tautology: it is the
>condition for successful holding period arbitrage, just like MR=MC is the
>condition for profit maximization. The CB can be viewed as indirectly
>attempting to target the nominal and real LT rates, but it will be
>successful only if it has sufficient credibility in the capital markets.
>
>Now although targeting a low ST or LT real rate is Good, i.e. Keynesain, it
>is not true that the real rate should be held BELOW the LT real growth rate.
>The reason is that this situation, if maintained over a period of time, will
>lead to rational bubbles.
>
>The present value of an infinite growing income stream (e.g. land) can be
>formulated:
>
>PV = R/(r-g), where r is the discount (interest) rate and g is the growth
>rate of real GDP. When the CB keeps  r < g, this causes the discounted value
>of land to rise indefinitely.  Land values, and stock values if companies
>own land, can then increase indefinitely.
>
>This is the explanation of the Japanese bubble in the late 80's. The Bank Of
>Japan lowered long term nominal rates to 2 % , while the grwoth of real GDP
>was 5-6 percent, and the inflation rate had been reduced to zero. The value
>of Japanese land became indefinitely high, which also pushed up stock values
>to 100x earnings.
>
>Such bubbles are dangerous, since if loans are made against such assets as
>security, when asset prices later fall, it is as if an atomic bomb went off
>in the financial industry. Many FI's become "zombies", with negative net
>worth. The Japanese have still not yet cleared away the rubble from the 89
>fallout.
>
>Basil Moore
>
>
>Basil Moore, Department of Economics
>Wesleyan University
>685-2363
>
>
>
>
>
>


Basil Moore, Department of Economics
Wesleyan University
685-2363