virtue of the multiplier approach – its simplicity without obtaining equivalent benefits in the area of their greatest deficiency, their lack of behavioural content, for such multipliers generally do not illuminate the behavioural process whereby people and institutions adjust their overall portfolio to arrive at some general equilibrium. Yet to be able to express changes in the money stock, or in the total holdings of liquid financial assets, in terms of only three variables has considerable advantages of brevity and simplicity, though even these advantages may be lost those circum- stances where there is a plethora of differing kinds of banks or intemediaries, and of deposits, each involving separate reserve ratios. None the less, the lack of any innate theoretical, or behavioural, content in the multiplier approach, per se,¹ may be realised more easily by noting that multiplier identities can be constructed over a virtually limitless range of cases. Take any aggregate, X, which can be decomposed into two parts, Y and Z, so that X=Y+Z then if the provision of Z is constrained by the identity U=W + Z one can construct the identity X = U. (1 + Z/Y) (W/Y+Z/Y) and various other mathematically-equivalent identities. Consider, for example, the potato multiplier of total personal expenditure, E. Personal expenditure is used for the purchase of potatoes (P), and of other goods and services (O), so that E = P₂ + O The total production of potatoes (P) either is sold to persons (P₁) or goes to other uses (P), so that P = P₁+ P₁ then E = P. (1 + P/O) (P/O +P/O) In short, total personal expenditure is a multiple of the value of the production of potatoes, with the size of the multiplier depending on only two ratios, the ratio of expenditure on potatoes to expenditure on other goods and the ratio of sales of potatoes for uses other than personal consumption to other personal expenditure. One can spread one's wings in the construction of exotic multipliers. Take, for example, the academic multiplier of total wealth. Let W represent total wealth, which is defined as comprised of human and non-human wealth, so that W = WN + WH 'It is possible, of course, to add behavioural content by analysing the derivation of the elements in the identity in terms of the underlying behavioural relationships. But that leads back towards specifying the full structure of the system, as we did in Chapter 5. If it is necessary to specify the structure of the system in order to understand why the multiplier works as it does, it is difficult to see what advantage is to be gained from using it as an analytical tool in the first place.
というのも、乗数計算の最大の欠点である行動学的内容の欠落は、乗数計算では一般に、人々や機関がポートフォリオ全体を調整して一般均衡に到達する行動過程を明らかにすることができないからである。しかし、マネーストックや流動性金融資産の保有残高の変化をたった3つの変数で表現できることは、簡潔で単純であるという点でかなりの利点がある。ただし、銀行や仲介業者、預金の種類が多数あり、それぞれが個別の準備率を伴う状況では、これらの利点も失われるかもしれない。しかし、乗数法それ自体に理論的・行動的な内容がないことは、乗数の恒等式が事実上無限の場合について構築できることに注目すれば、より容易に理解できるであろう。X=Y+Z となるように Y と Z の二つの部分に分解できる任意の集合体 X を考え、Z の供給が U=W+Z という恒等式で制約されるなら、X=U. (1+Z/Y) (W/Y+Z/Y) やその他の数学的に同等な恒等式が構成可能である。例えば、個人支出総額のジャガイモ乗数Eを考えてみよう。個人支出はジャガイモ(P)と他の財・サービス(O)の購入に使われるので、E=P₂+O ジャガイモの生産総量(P)は人に売るか(P₁)他の用途(P)なので、P=P₁+P₁ ならE=P.である。(1+P/O) (P/O+P/O) 要するに、個人の総支出は、ジャガイモの生産額の倍数であり、その倍数の大きさは、他の財への支出に対するジャガイモへの支出の比率、他の個人支出に対する個人消費以外の用途のジャガイモの売却の比率の2つだけに左右されるのである。エキゾチックな乗数を構築することも可能である。例えば、総資産に関する学術的な乗数を考えてみよう。Wは総資産を表し、それは人間と人間以外の富から構成されると定義され、W = WN + WH「もちろん、根本的な行動関係の観点から恒等式の要素の派生を分析することによって行動的内容を加えることは可能である。しかし、それは第5章で行ったように、システムの完全な構造を指定する方 法に戻ることになる。乗算器がなぜそのように機能するかを理解するためにシステムの構造を特定することが必要であるなら、そもそもそれを分析ツールとして使うことにどんな利点があるのかがわからない。
134 MONEY, INFORMATION AND UNCERTAINTY H Let Wµ/W= d; let the value of human wealth be related to educational input (E); assume, for simplicity only, a linear relationship, so that WH = a +bE Let the ratio of expenditure on academic lectures (A) to total education expenditures be g - i.e., A/E=g - then it follows that b w=9+ d gd W A Total wealth is thus some multiple of expenditure on academic lectures, plus a term, a/d, which may in the short run be taken as constant. Moreover, both g and d are between 0 and 1, since they relate a component of an aggregate to an aggregate, while we may presume that b is greater than unity, since education is widely regarded as a worthwhile investment for society. Total wealth is therefore related to expenditure on academic lectures by some very large multiplier. Apart from the simplifying assump- tion of the linear relationship between education expenditures and human wealth, and of the approximate short-term constancy of the a/d ratio, it is all true by definition. These examples are intended to persuade readers either that it will be of enormous benefit to the nation to increase the salaries of academic lectures multifold or, alternatively, that the multiplier approach needs cautious handling if it is to be used as a basis for explaining something large (say, national income, or the money stock) from movements in a much smaller component of that aggregate, say autonomous investment, or high-powered money, grossed up by some function which involves ratios connecting the small total to the large.¹ In order to use such an approach to 'explain' variations in the larger total as contrasted with describing such move- ments, which definitional multipliers, however ridiculous, can always do some further conditions are necessary. The primary condition is that the ratios linking the two variables should be predictable, say a stable function of other variables which will also be predictable, under all prospective circumstances. Clearly, if the ratios vary unpredictably, then information on the value of the small variable will not allow one to forecast how the large variable will change. But even if the ratios seem stable under one set of circumstances, they may not be so generally. Consider again the previous example of the academic multiplier of total wealth. It may be that the ratios in that example (b, g and d) have remained fairly stable over time, reflecting the preferences and structure of the system. If the authorities, however, should alter that system by doubling (or halving) expenditure on academic lecturers at a stroke, the values of the ratios would alter quite sharply, so that the previous stability would be shattered. Obviously, in this case, the authorities cannot rely on the continued stability of this multiplier to vary total wealth by adjusting expenditure on academic lecturers. If the ratios remain stable and predictable under all feasible states of the world, then if you know the value of the small variable you will be able to predict with a high degree of confidence what the value of the large variable will be. In this sense, and this sense only, it is possible to say that variations in the small variable (say, the high- 'The multiplier approach is, alas, used indiscriminately by all schools of macro-economic analysis. The Keynesian multiplier, relating changes in incomes to changes in autonomous expenditures, is of this genre. The theory of distribution developed by N. Kaldor (1955) pp. 83- 100 is essentially based on a definitional multiplier of this kind.
●●● CONTROLLING THE SUPPLY OF MONEY 135 powered money base) explain the variations in the large variable (say, the money stock). It should be noted that this usage of the term 'explanation' does not depend on whether the small variable is being determined exogenously (say, as a control variable by the authorities) or endogenously, when these are alternative possible states of the world. All that is necessary is that the relationship should be predictable and stable in all possible states; indeed the bank multiplier will in some respects, perhaps, be most useful if the relationship remains unaltered whatever the state of the world, in our previous example whether the high-powered money base is, or is not, being fixed consciously by the authorities. On the whole, the bank multiplier relating the money stock to the high-powered money base has been quite useful in the above manner.¹ As long as the probability distribution of deposit inflows and withdrawals, and the penalties resulting from a cash shortage, remain fairly constant,2 the banks' reserve ratio is likely to remain a stable function of relative yields on alternative assets. Similarly, the public's desired currency - deposit ratio usually remains fairly constant over time, changing gradually in response to slow-moving institutional factors, and also perhaps to the relative yield attractions of holding deposits. It is possible, however, that if the authorities should suddenly change their system of monetary control, say by moving from a system in which they fixed interest rates in the market as their prime target to one in which they concentrated upon determining the monetary base, the subjective uncertainties - and penalties facing the banks might change, resulting in possibly unforeseeable changes in their desired reserve ratios.³ And if developments in the banking system should make the public revise their views of the safety of their deposits, the public's desired currency - deposit ratio could change very sharply, as in the United States in the 1930s. Nevertheless, as Friedman and Schwartz, and Cagan, have shown in the case of the United States, the relationships involved in the bank multiplier have remained fairly stable over a long run of years and during several different monetary regimes e.g., gold standard with no Central Bank, Federal Reserve System in its differing phases of operation. So if you know what the change in the monetary base has been over some period, it is likely that you will be able to forecast what the change in the money stock will have been over the same period with reasonable accuracy, irrespective of the nature of the monetary regime or of changes in that regime. In this respect, the bank multiplier has some useful explanatory content, as compared with the other examples of the potato multiplier or the academic multiplier, which have none. Moreover, this ability to forecast the larger aggregate (e.g., the money stock) from movements in the smaller variable (e.g., the high-powered money base) requires that only a few ratios be predictable. If one instead looks at the wider canvas, at the complete portfolio adjustment of the sectors concerned, involving the interplay of all the asset preferences, the number of behavioural functions which have to be included ¹ See, in particular, the work of Cagan (1965). 2 On this subject, see Morrison (1966). 3 Changes in official reserve requirements will naturally alter the desired reserve ratio; but the response should be fairly predictable. The difficulty of predicting how the banks, and indeed how the remainder of the financial system, might respond to the introduction of monetary base control (MBC), and the the problems of operating such a system in the transitional period, were arguments deployed against a move to MBC in the UK in the early 1980s, see HM Treasury, Monetary Control (1980) and Bank for International Settlements, Monetary and Economic Department, The Monetary Base Approach to Monetary Control (1980).
貨幣の供給を制御する 135)により、大きな変数(例えば、マネーストック)の変動が説明される。この「説明」という言葉の使い方は、小さな変数が外生的に(例えば、当局によるコントロール変数として)決定されているか、内生的に決定されているかには関係なく、これらが世界の代替可能な状態である場合、注意する必要があります。必要なのは、その関係があらゆる可能な状態において予測可能で安定していることです。実際、銀行乗数は、世界の状態がどうであれ、その関係が変化しない場合、ある面では、おそらく最も有用でしょう。先ほどの例では、ハイパワードマネーベースが当局によって意識的に固定されているか、されていないか、です。全体として、マネーストックとハイパワードマネーベースとの関係における銀行乗数は、上記のように非常に有用であった¹ 。預金の流入・流出の確率分布と現金不足によるペナルティがほぼ一定である限り2、銀行の預金準備率は代替資産の相対利回りの関数として安定的に推移する可能性が高い。同様に、国民が望む通貨と預金の比率は、通常、時間の経過とともにかなり一定に保たれ、動きの遅い制度的要因や、おそらく預金保有の相対的利回りの魅力に反応して徐々に変化する。しかし、当局が金融管理システムを突然変更した場合、例えば、市場金利を最重要目標として固定するシステムから、マネタリーベースの決定に集中するシステムに移行した場合、銀行が直面する主観的不確実性や罰則が変化し、結果として、望ましい準備率に予測できない変化が生じる可能性がある³。しかし、フリードマンとシュワルツ、そしてケーガンが米国のケースで示したように、銀行乗数に関係する関係は、長い年月の間、そして中央銀行のない金本位制、運用段階が異なる連邦準備制度などいくつかの異なる通貨体制において、かなり安定している。したがって、ある期間のマネタリーベースの変化を知っていれば、同じ期間のマネーストックの変化を、金融レジームの性質やその変化にかかわらず、合理的な精度で予測できる可能性が高いのです。この点で、銀行乗数は、他の例である芋づる式乗数や学問的乗数が何もないのに比べ、有用な説明的内容を持っています。しかも、このように小さな変数(例えばハイパワードマネーベース)の動きから大きな総体(例えばマネーストック)を予測する能力は、いくつかの比率だけが予測可能であることが必要である。その代わりに、より広い視野で、すべての資産選好の相互作用を含む関係部門の完全なポートフォリオ調整に目を向けると、含まれなければならない行動関数の数が多くなる¹ 特に、Cagan (1965) の研究を参照のこと。2 このテーマについては、Morrison (1966)を参照。3 公的準備率の変更は、当然ながら望ましい準備率を変化させるが、その反応はかなり予測可能であるはずである。1980 年代初頭の英国では、マネタリーベース・コントロール(MBC)の導入に際して、銀行、ひいては他の金融シス テムがどのように反応するかを予測することの難しさや、過渡期におけるそうしたシステムの運用の問題が、MBC への移行に反対する論拠となった(HM Treasury, Monetary Control (1980) および Bank for International Settlements, Monetary and Economic Department, The Monetary Base Approach to Monetary Control (1980) を参照)。
133
CONTROLLING THE SUPPLY OF MONEY
virtue of the multiplier approach – its simplicity without obtaining equivalent
benefits in the area of their greatest deficiency, their lack of behavioural content, for
such multipliers generally do not illuminate the behavioural process whereby people
and institutions adjust their overall portfolio to arrive at some general equilibrium.
Yet to be able to express changes in the money stock, or in the total holdings of liquid
financial assets, in terms of only three variables has considerable advantages of
brevity and simplicity, though even these advantages may be lost those circum-
stances where there is a plethora of differing kinds of banks or intemediaries, and of
deposits, each involving separate reserve ratios.
None the less, the lack of any innate theoretical, or behavioural, content in the
multiplier approach, per se,¹ may be realised more easily by noting that multiplier
identities can be constructed over a virtually limitless range of cases. Take any
aggregate, X, which can be decomposed into two parts, Y and Z, so that
X=Y+Z
then if the provision of Z is constrained by the identity
U=W + Z
one can construct the identity
X = U.
(1 + Z/Y)
(W/Y+Z/Y)
and various other mathematically-equivalent identities. Consider, for example, the
potato multiplier of total personal expenditure, E. Personal expenditure is used for
the purchase of potatoes (P), and of other goods and services (O), so that
E = P₂ + O
The total production of potatoes (P) either is sold to persons (P₁) or goes to other
uses (P), so that
P = P₁+ P₁
then
E = P.
(1 + P/O)
(P/O +P/O)
In short, total personal expenditure is a multiple of the value of the production of
potatoes, with the size of the multiplier depending on only two ratios, the ratio of
expenditure on potatoes to expenditure on other goods and the ratio of sales of
potatoes for uses other than personal consumption to other personal expenditure.
One can spread one's wings in the construction of exotic multipliers. Take, for
example, the academic multiplier of total wealth. Let W represent total wealth, which
is defined as comprised of human and non-human wealth, so that
W = WN + WH
'It is possible, of course, to add behavioural content by analysing the derivation of the
elements in the identity in terms of the underlying behavioural relationships. But that leads
back towards specifying the full structure of the system, as we did in Chapter 5. If it is necessary
to specify the structure of the system in order to understand why the multiplier works as it does,
it is difficult to see what advantage is to be gained from using it as an analytical tool in the first
place.
134
MONEY, INFORMATION AND UNCERTAINTY
H
Let Wµ/W= d; let the value of human wealth be related to educational input (E);
assume, for simplicity only, a linear relationship, so that
WH = a +bE
Let the ratio of expenditure on academic lectures (A) to total education expenditures
be g - i.e., A/E=g - then it follows that
b
w=9+
d gd
W
A
Total wealth is thus some multiple of expenditure on academic lectures, plus a term,
a/d, which may in the short run be taken as constant. Moreover, both g and d are
between 0 and 1, since they relate a component of an aggregate to an aggregate, while
we may presume that b is greater than unity, since education is widely regarded as a
worthwhile investment for society. Total wealth is therefore related to expenditure on
academic lectures by some very large multiplier. Apart from the simplifying assump-
tion of the linear relationship between education expenditures and human wealth,
and of the approximate short-term constancy of the a/d ratio, it is all true by
definition.
These examples are intended to persuade readers either that it will be of enormous
benefit to the nation to increase the salaries of academic lectures multifold or,
alternatively, that the multiplier approach needs cautious handling if it is to be used
as a basis for explaining something large (say, national income, or the money stock)
from movements in a much smaller component of that aggregate, say autonomous
investment, or high-powered money, grossed up by some function which involves
ratios connecting the small total to the large.¹ In order to use such an approach to
'explain' variations in the larger total as contrasted with describing such move-
ments, which definitional multipliers, however ridiculous, can always do some
further conditions are necessary.
The primary condition is that the ratios linking the two variables should be
predictable, say a stable function of other variables which will also be predictable,
under all prospective circumstances. Clearly, if the ratios vary unpredictably, then
information on the value of the small variable will not allow one to forecast how the
large variable will change. But even if the ratios seem stable under one set of
circumstances, they may not be so generally. Consider again the previous example of
the academic multiplier of total wealth. It may be that the ratios in that example (b, g
and d) have remained fairly stable over time, reflecting the preferences and structure
of the system. If the authorities, however, should alter that system by doubling (or
halving) expenditure on academic lecturers at a stroke, the values of the ratios would
alter quite sharply, so that the previous stability would be shattered. Obviously, in
this case, the authorities cannot rely on the continued stability of this multiplier to
vary total wealth by adjusting expenditure on academic lecturers.
If the ratios remain stable and predictable under all feasible states of the world,
then if you know the value of the small variable you will be able to predict with a high
degree of confidence what the value of the large variable will be. In this sense, and this
sense only, it is possible to say that variations in the small variable (say, the high-
'The multiplier approach is, alas, used indiscriminately by all schools of macro-economic
analysis. The Keynesian multiplier, relating changes in incomes to changes in autonomous
expenditures, is of this genre. The theory of distribution developed by N. Kaldor (1955) pp. 83-
100 is essentially based on a definitional multiplier of this kind.
CONTROLLING THE SUPPLY OF MONEY
135
powered money base) explain the variations in the large variable (say, the money
stock). It should be noted that this usage of the term 'explanation' does not depend
on whether the small variable is being determined exogenously (say, as a control
variable by the authorities) or endogenously, when these are alternative possible
states of the world. All that is necessary is that the relationship should be predictable
and stable in all possible states; indeed the bank multiplier will in some respects,
perhaps, be most useful if the relationship remains unaltered whatever the state of the
world, in our previous example whether the high-powered money base is, or is not,
being fixed consciously by the authorities.
On the whole, the bank multiplier relating the money stock to the high-powered
money base has been quite useful in the above manner.¹ As long as the probability
distribution of deposit inflows and withdrawals, and the penalties resulting from a
cash shortage, remain fairly constant,2 the banks' reserve ratio is likely to remain a
stable function of relative yields on alternative assets. Similarly, the public's desired
currency - deposit ratio usually remains fairly constant over time, changing gradually
in response to slow-moving institutional factors, and also perhaps to the relative yield
attractions of holding deposits. It is possible, however, that if the authorities should
suddenly change their system of monetary control, say by moving from a system in
which they fixed interest rates in the market as their prime target to one in which they
concentrated upon determining the monetary base, the subjective uncertainties - and
penalties facing the banks might change, resulting in possibly unforeseeable
changes in their desired reserve ratios.³ And if developments in the banking system
should make the public revise their views of the safety of their deposits, the public's
desired currency - deposit ratio could change very sharply, as in the United States in
the 1930s. Nevertheless, as Friedman and Schwartz, and Cagan, have shown in the
case of the United States, the relationships involved in the bank multiplier have
remained fairly stable over a long run of years and during several different monetary
regimes e.g., gold standard with no Central Bank, Federal Reserve System in its
differing phases of operation. So if you know what the change in the monetary base
has been over some period, it is likely that you will be able to forecast what the change
in the money stock will have been over the same period with reasonable accuracy,
irrespective of the nature of the monetary regime or of changes in that regime. In this
respect, the bank multiplier has some useful explanatory content, as compared with
the other examples of the potato multiplier or the academic multiplier, which have
none.
Moreover, this ability to forecast the larger aggregate (e.g., the money stock) from
movements in the smaller variable (e.g., the high-powered money base) requires that
only a few ratios be predictable. If one instead looks at the wider canvas, at the
complete portfolio adjustment of the sectors concerned, involving the interplay of all
the asset preferences, the number of behavioural functions which have to be included
¹ See, in particular, the work of Cagan (1965).
2 On this subject, see Morrison (1966).
3 Changes in official reserve requirements will naturally alter the desired reserve ratio; but the
response should be fairly predictable.
The difficulty of predicting how the banks, and indeed how the remainder of the financial
system, might respond to the introduction of monetary base control (MBC), and the the
problems of operating such a system in the transitional period, were arguments deployed
against a move to MBC in the UK in the early 1980s, see HM Treasury, Monetary Control
(1980) and Bank for International Settlements, Monetary and Economic Department, The
Monetary Base Approach to Monetary Control (1980).
0 件のコメント:
コメントを投稿