III. IMPLICATIONS OF THESE CHANGES FOR MONET ARY POLICY
2. THE BASIC COMPON ENTS
The basic compon ents of a Keynesian model of invest ment are repres en ted in Figure 1. The Keynesian model postu lates that two sets of markets inter act
Figure 1
Accepting Accepting Accepting Accepting
Accepting Accepting Accepting Accepting 30
30
30
30 30
30 30
30
30
30
30 30
30
in determ in ing invest ment. The first set consists of those markets in which the prices of the units in the stock of capital goods and finan cial assets are determ ined. This is repres en ted in Figure 1a by the Pk func tion. The second set consists of the markets in which the pace of invest ment is determ ined by a combin a tion of finan cing and supply condi tions. This is repres en ted by the I and the Nc func tions in Figure 1b.
In Figure 1b, the I func tion gives the supply price of real invest ment goods as a func tion of the rate of output. The Nc func tion gives the internal finan cing per unit of invest ment as a func tion of the pace of invest ment (Nc is a rectangularhyperbola if the flow of internal busi ness funds is inde
pend ent of invest ment). For any given pace of invest ment the vertical differ
ence between the two curves Nc − PI gives the surplus or deficit of internal funds per unit of invest ment.
The money and capital market determ ine on what terms such defi cits are to be financed and such surpluses are to be util ized. Thus, at a deeper level the Keynesian theory of invest ment has to include a model of the beha vior and evol u tion of money and capital markets [21]. It is by way of the money and capital markets that the finan cial flows set up by the current finan cing of invest
ment feed back to the markets which determ ine the prices of items in the stock of capital. Whatever changes in finan cing terms result as invest ment is financed affect the terms upon which posi tions in the stock of capital goods and finan
cial assets are financed. This in turn affects the price of units in the stock of capital assets: finan cial market connec tions integ rate stock and flow prices.
These rela tions are not neces sar ily simul tan eous and the sequence of reac
tions as well as the initi at ing disequi lib rium are not always the same.3 In partic u lar both the evol u tion and the instabil ity of finan cial sectors can affect the terms upon which invest ment can be financed—and thus its pace—as well as the prices of the stocks of capital goods.
The theory as sketched will abstract from the timeconsuming nature of the invest ment process. Presumably the price of capital encom passes the price of a nuclear power plant. The gest a tion period for a nuclear power plant is at least five years. In the empir ical imple ment a tions of such a model, the invest ment flow relates to the invest ment per period—say per quarter. A decision to pay Pk for a nuclear power plant results in a flow of invest ment It over the next 20 quar ters such that (ignor ing discount ing)
Accepting Accepting Accepting
where each It is determ ined by the tech nical condi tions of produ cing nuclear power plants.
In imple ment ing this theory for an econo met ric model, the invest ment process might well be divided into capital goods order ing and invest ment flow phases. With that dicho tomy in mind it might very well be that the first part of the theory sketched here is more of a theory of invest ment order ing than a theory of invest ment flows. The second part, which emphas izes finan
cing and finan cial reper cus sions, reflects what happens as invest ment flows and finan cing actu ally take place—perhaps long after the initial decision to invest was made.
In summar iz ing research which involved the replic a tion with a consist ent body of data of a number of altern at ive econo met ric formu la tions of invest
ment, Bischoff remarked that accel er ator based models—whether simple or flex ible (the flex ib il ity depend ing upon relat ive prices of inputs)—did better than the cash flow or secur ity price altern at ives. The cash flow or secur ity price models tested by Bischoff are not adequate repres ent a tions of what is here viewed as the Keynesian theory of invest ment. We can take Bischoff’s tests as missing the point about Keynesian models.
However, we have to face up to the fact that we do not expli citly consider the accel er ator in our formu la tion. Long ago (see next essay) I examined inter re la tions between accel er ator effects and monet ary (finan cial) beha vior.
In my formu la tion the accel er ator model gave us exante invest ment. Expost invest ment resul ted from the inter ac tion of monet ary (finan cial) factors and accel er ator consid er a tions. That formu la tion ignored the supply func tion for invest ment goods output and the pricing process for stocks of capital goods, which are central to the present expos i tion.
Investment activ ity is the result of a combin a tion of productiv ity and spec u lat ive factors. The productiv ity and scarcity of capital services result in current and expec ted future cash flows to owners of capital. The accel er ator basic ally is an asser tion that if an output greater than current output is to be produced, an incre ment to the capital stock of a partic u lar size can be expec ted to yield satis fact ory or adequate cash flows. In terms of price theory concepts this means that for a given long run average vari able cost curve—derived from a “produc tion func tion”—there exists a minimum average plan ning total cost curve. This curve defines the sets of product prices and outputs that are expec ted to result in such adequate cash flows. Thus, the cash flows from oper a tions of real assets as used in the text that follows embody the productiv ity of invest ment.
Let us assume that wages are fixed and as a result the minimum supply price of the invest ment good is PI(0). If Pk > PI(0), then invest ment is taking place at a rate so that PI = Pk after allow ing, as in the text, for finan cing discounts. Any entre pren eur purchas ing an invest ment good at Pk = PI > PI(0) expects to “earn” extraordin ary quasirents for a long enough time, so that when PI = PI(0), the extra costs will have been written off. Thus expect a tions with respect to future cash flows are imbed ded in Pk. The higher Pk for a given PI(0), the larger and the longer the expec ted dura tion of premium quasirents. Because the capital stock is increas ing due to invest ment, the higher Pk the greater the capital stock that is assumed to be neces sary for quasirents to be at their normal or adequate level. Thus, the excess of Pk over the minimum normal supply price of invest ment output is a measure of the differ ence between the exist ing and the target capital stock.
Accelerator ideas can thus be read into the pricing of capital assets.
However, the strong Keynesian view is that in an economy with cyclical exper i ence and capit al ist organ iz a tion spec u lat ive elements domin ate productiv ity consid er a tion in determ in ing the price of invest ment. By concen trat ing on how elements other than expec ted cash flows affect the price of real assets in our uncer tain world, the spec u lat ive aspects of the invest ment decision are brought into focus.
To summar ize: In market invest ment rela tions estim ated over periods char ac ter ized by rather steady economic growth vari ables which are inter
preted as embody ing accel er ator concep tions turn out to have a large explan at ory weight. In our present formu la tion productiv ity concepts are embod ied in the cash flows from oper a tions that capital assets earn, espe
cially in the cash flows anti cip ated as invest ment decisions are being made.
However, the spec u lat ive element is intro duced into the pricing of assets by way of the contin gent cash flows by way of “sale.” At times the weight of the two sources of value for assets changes so that spec u la tion domin ates. This is so even though for long periods the regu lar ity of invest ment, output, and cash flows from oper a tions are such that meas ures which can be inter preted as reflect ing accel er ator rela tions seem to domin ate in what happens.
In Bischoff’s research the cash flow model, which is poorly specified from the perspect ives of this paper, does very well in explain ing invest ment over the data period. It does quite poorly in explain ing invest ment in 1969 and 1970. However, as I have emphas ized [23] a run of success such as was enjoyed in the 1960s will trigger a euphoric invest ment boom. During this boom debt finan cing of invest ment will expand rapidly, i.e., the investment
gross profit after taxes rela tion will change. Whereas the 1969–70 period illus trates the weak ness of a narrowly construc ted cash flow formu la tion, it is evid ence that a finan cial theory of invest ment, which allows for liab il ity struc tures—actual and desired—has a large measure of plaus ib il ity.