My focus will be upon Boulding's effort to construct a "classical" type of macroeconomic distribution theory. I will force all else through this funnel of vision. In his rejection of Walrasian equilibrium, Boulding distinguishes between the exchange process on the one hand, by which existing assets, including money, are circulated among various owners, and the processes of production, consumption, income and outgo on the other, by which assets are created, destroyed, and accumulated. As he says, marginalism treats the firm as if it has nothing but an income account. It has no balance sheet, no funding problems, and no dynamics. The firm maximizes a strange variable called "net revenue," of which no accountant has ever heard.

Boulding, therefore, uses the balance sheet as the central analytical concept of the theory of the firm and introduces asset preferences, which hitherto had been confined almost exclusively to banking theory. Boulding's unique contribution is to show how the personal income distribution, effected by a potentially volatile financial transfers item (T), is a determinant of output (and, by extension, employment). Indeed, Boulding's T is a potential source of financial fragility. I proceed to integrate Boulding's income distribution theory with Kalecki's quantitative notion of economic power. Finally, I draw a connection between Minsky's financial fragility hypothesis and Boulding's T via Kalecki. As I will contend, this cluster of ideas-which brings together production and money-sheds considerable light on the leveraged nature of a "casino economy."[1]

The Macroeconomic Income Distribution Identities

**Summary**

The balance sheet is central. Profit-making is related to the gross growth in the value of net worth. The separate summations of business and household balance sheets provide identities for gross profits (non-labor income) and for the wage bill, the component parts of which are in some degree parameters of behavior and which can therefore be used to construct macroeconomic models.

The distribution of national income between wages and gross profits is determined by several related sets of decisions. Gross profits are greater the more business adds to its stocks of capital (investment decisions), the more it finances investment out of profits, the more it distributes interest, rent, and dividends, the more consumer credit it extends (finance decisions), and the more the money stock shifts out of household into business accounts (liquidity decisions). The more the money stock shifts out of business into household accounts, the less the increases in consumer credit and sales of securities to households, the less will be the distributions to capitalists in rent, interest, and dividends.

The conclusion is heretical. Investment, consumption, finance, and liquidity decisions decide the distribution of income rather than wage bargaining (except in so far as such bargains react on these decisions). Whereas Walrasian theory sees payments to capitalists as one side of an equal exchange, Boulding has financial transfers by the firm returning through its front door. Being one-way transfers does not mean that they are grants, as defined later in Boulding's "grants economy." In Reconstruction, the financial transfers by business firms to households boomerang to the advantage of the bottom line.

**Business and Household Net Worth**

The aggregate net worth of businesses is equal to their money holdings (Mb), plus their value of goods holdings (Qb), plus the net household indebtedness to business (Kh - Kh'), where Kh is total debts due to businesses from households and Kh' is the total debts due from businesses to households. The latter consist mostly of the bonds of businesses held by households. From these variables, Boulding derives the business-net-worth identity, which can be expressed as the differential equation

dGb = dMb + dQb + dKh - dKh' = Sb (1)

in which the d's are accounting period time derivatives.

The quantity dGb is that part of profits not distributed in dividends and can be called business saving if positive-dissavings if negative. Business saving (Sb), therefore, is generally not determined by the decisions of businesses to save, but rather result from other decisions that determine business savings such as the effort of individual firms to increase their net worth-through an increase in money holdings (dMb), an increase in the value of durable goods (dQb), an increase in business credit extended to households (dKh), or a decrease in bonds issued to households (dKh'). As we will see, business decisions to save, reflected in dividends, determine not business savings but the level of profits.

Thus, if V is the aggregate profits of business, it must equal business savings plus business distributions in interest and dividends. Total profit is the gross increase in net worth, out of which interest and dividends are paid. Thus,

V = dGb + D, (2)

where D represents aggregate distributions in interest and dividends. Business savings is

dGb = V - D (3)

Imagine that dGb has already been decided by the variables of Equation 1-the increase in money and goods holdings plus net credit extended to households. The aggregate distributions are mostly governed by dividend declarations since interest payments are based mostly on past contractual obligations. If dGb is independent of such distributions (as stated in Equation 1), business payments of interest and dividends comprise a "widow's cruse" whereby the greater the business interest and dividend distributions (Equation 2), the more will be returned to business in profits to distribute. Profits leap from one pocket into another on the same pair of pants, and business wears the pants. An increase in dividends by one firm will swell the profits of other firms.

There is a second way of explaining this apparent paradox. Once saving (dGb) is decided by elements unrelated to the saving decision, the savings part of aggregate profits has been predetermined. If, in turn, the business sector decides to reduce its savings, it can only do so by increasing its dividends. This increase in dividends-which does not alter the original savings level at all-is added to saving and enlarges gross aggregate profits.

Substituting the RHS of Equation 1 for dGb in Equation 2, gross aggregate business projects are

V = dQb + (dMb + dKh- dKh' + D) (4)

All items on the RHS, except for dQb, are "transfer items" or T. I will have more to say about T later.

Since the K's are assets in one balance sheet and liabilities in the other, and since all the net worth of businesses must ultimately be allocated to households, there is a comparable household identity for wages. In turn, household saving can be found by differentiating that sector's net worth identity, and

dGh = dMh + dMb + dQh + dQb - (dKh - dKh') = Sh. (5)

In turn, according to Boulding [1950, 248], "the values of all titles to property cancel out when we add up all balance sheets, and we are left with the identity that the net worth of households is equal to the 'value' of all 'real' property, including money." Thus, the change in household net worth is reduced to,

dGh = dMh + dMb + dQh + dQb = Sh, (6)

so that household saving is equal to the changes in the total quantity of money in the system plus the changes in the total value of the stock of real assets (goods) in the system.

When Boulding assumes a given national money supply, any increase in money holdings of business (dMb) is a decrease in money holdings of households (-dMh), and Equation 6 is simplified to

dGh = dQb + dQh. (7)

Aggregate saving, all of which ultimately resides in households, equals the increase in aggregate real capital. Later, he suggests that "relaxing this assumption does not make much difference to the argument" [Boulding 1950, 253]. This I doubt. Since there is a net accumulation of goods during the accounting period, Boulding must be assuming (though he is silent on this) that the net increase in output is financed by expanding credit (with no increase in endogenous money) or by an increase in the income velocity of money.

Obviously, household saving also equals household income less household consumption where income is the gross addition to households' net worth. It consists of wages earned (W), dividends paid (D), and business savings (dGb) that accrue ultimately to households. Given household consumption (Ch), we can write:

dGh = (W + D + dGb) - Ch. (8)

Taking the value of dGb from the RHS of Equation 1 and of dGh or dQh + dQb (the total accumulation of durables and materials) from Equation 7, we have

W = Ch + dQh - (dMb + dKh - dKh' + D). (9)

The same financial transfers item (dMb + dKh - dKh' + D) = T appears in aggregate profits (Equation 4) and aggregate wages. Thus, we can write

W = Ch + dQh - T, and (10)

V = dQb + T. (11)

We can think of the "transfer" as being from wages to profits with no necessary quid pro quo in goods.

**Net Domestic Product and the Transfer Item**

Aggregate wages and profits exhaust aggregate output so that net domestic product or income is

Y = W + V = Ch + dQh + dQb. (12)

GDP, of course, would include depreciation, which is not distributed and is not a part of national income (Y).

Paradoxes abound and enliven Reconstruction. In particular, paradoxes populate Boulding's unique contribution, the "transfer item," or transfer of wages to profit (T). The first term in this transfer (dMb) is the net transfer of money from households to businesses that reflects the relative liquidity preferences of households and businesses (measured by the velocity of money).

The second transfer term (dKh) is the change in consumer credit. Such credit, which would include credit cards and installment credit, can shift the income distribution from labor to favor profits. Whereas an increased use of credit by a particular household can enhance its standard of living, all households doing the same thing will reduce the income/product share of labor. Could this perhaps help to explain why-at the end of the consumer-credit-fueled economic expansion of the 1980s-the income distribution had shifted in favor of the rich?

The third transfer is the increase in business indebtedness to households (-dKh'), mostly comprised of corporate bonds. It is given a negative sign so that the greater the volume of securities sold to households, the more income is redistributed toward wages and away from profits. This generates still another paradox: Since the sale of bonds reduces the profits share in expectation of which the securities are offered, the purchase of bonds (usually by the upper class) benefits a particular investor to the detriment of his own class. This attraction to bonds is not fatal, because undistributed profits can be used to increase the level of business investment.

The fourth transfer is business dividends and interest payments (D). It is the source of still another paradox, the aforementioned "widow's cruse." Since business saving has been determined by other forces, each dollar increase in D swells aggregate profits a dollar. Boulding concludes that dividend policy does not affect consumption patterns, except as these are affected by the liquidity of the shareholder [Boulding 1950, 257].

Thus far, Boulding has assumed that the central bank and private banks are not creating any new money nor is the federal government issuing new debt. Later, he relaxes these assumptions. When one does [Canterbery 1993, 168], the RHS of Equation 12 becomes

Y = Ch + dQh + dQb + (dMh + dMb + dBh + dBb), (13)

where dBb comprise changes in the value of government bond holdings. As Canterbery [1993, 168] concludes, once these dynamic elements are introduced, some of the components of T reappear in the nominal national income. Now the nominal national income can increase as a direct result of credit and money creation. An increase in real national income is by no means guaranteed because real investment may not increase.

So far, so good. The income distribution results are similar to those of Michal Kalecki [see, e.g., Davidson 1960, chap. 5]. In particular, the "widow's cruse" is essential Kalecki. Even though these equations provide remarkable insights, an economist could easily raise the objection that they nonetheless are mathematical identities and not functions. However, Boulding does go on to develop two distributional models with functional relations.

Income Distribution Models with the Transfer Item

**Distributional Model I**

Boulding shows how his identities provide a foundation for a large number of macroeconomic models of comparative state equilibrium and even of dynamics. He presents two models. Model I is driven by the functions in which the more common symbols C, I, and Y are used. Household consumption (or C), however, is the sum (Ch + dQh). Consumption is a function of the wage share of output (W/Y), and business investment is a function of the profits share (V/Y). Boulding's T is viewed in two ways: One T divides profits from wages with T independent of the distributional shares and its components given; and a second T is a function of the profits share. If T is given, we have seven independent equations and seven unknowns: If T is variable, the model is overdetermined. In the latter case, we can write the consumption, investment, and transfer functions as:

C = C(W/Y) (14)

I = I(V/Y)

T = T(V/Y)

Then, the identities of composition are:

Y = C + I (15)

Y = W + V

W = C - T

V = I + T

Consumption is a positive function of the wage share, and investment is a positive function of the profits share. Profits and wages also are determined by the levels of consumption and investment adjusted by transfers.

Boulding [1950, 259] provides only a geometrical "solution" for this model. A simplified version of his geometry appears as Figure 1: it reveals a remarkable intuition. Along the base O[sub w]O[sub v] is measured the proportionate distribution between wages and gross profits; at O[sub w] all the product goes to wages, and at O[sub v] all of the product goes to gross profits. Consumption, or CC, falls with the wages share, and investment, or II, rises with the profits share. Total output, or YY, initially rises as the profits share grows but peaks at Y[sub x], the distributional equilibrium (which may or may not be at full employment).

Vertical, parallel lines divide the distribution of total output into deciles. The loci of T's (at given values of the financial components of T) comprise what I call a factor-payments transfer curve, a curve that divides the total output into wages (the upper portion) and profit (the lower portion) at each level of distribution. At any distribution, T = V - I = C - W. For example, distribution 0[sub 2] represents 20 percent profits and 80 percent wages; I[sub 2]T[sub 2] then is total profits and T[sub 2]C[sub 2] total wages, and O[sub 2]T[sub 2] is the equilibrium factor-payments transfer.

Now, we introduce a variable financial transfer function or T(V/Y) = dMb + dKh - dKh' + D, wherein all components vary with V/Y. In order to identify it separately from the aforementioned factor-payments transfer curve of T's, the financial transfer item curve is RR. The changes in money holdings of business and the changes in debts due business from households is assumed not to be much affected by the wage-profits distribution. The third item, the change in debts due from businesses to households, will be closely related to the volume of investment: At the extreme, where all investment is financed by the sale of securities to households, dKh' = I. The fourth item, D, is likely to be related to profits because the higher go profits, the higher will distributions be. If all profits are distributed, V = D. At the extreme, where all investment is debt-financed, not only does V = D, but dKh' = I and dMb + dKh = 0, the equation of the R-curve becomes identical to the T-curve (i.e., R = V - I = T), and the solution is indeterminate. Once gross profits exceed contractual distributions, both D and dKh' are likely to rise. With T positive, gross profits are greater than business investment, and D is likely to rise faster than dKh'; then, the R-curve is likely to slope upwards. (The reverse is the case, however, if there is a strong tendency to finance investment by issuing new securities, as with the corporate debt explosion during the 1980s.)

We arrive now at Boulding's incredibly prophetic intuitive insight. First, he notes, "if the slopes of the T-curve and the R-curve are similar the equilibrium, if it exists, will be highly insecure and 'shiftable'; i.e. a slight change in the underlying functions will cause a large shift in the equilibrium position or will even destroy the possibilities of equilibrium altogether" [1950, 262]. If the R-curve lies wholly above the T-curve, there will be a constant tendency to increase the profit share until the economy breaks down. Or, if the R-curve lies wholly below the T-curve, there will be a constant tendency for the wages share to increase. "The possibility," he concludes, "of a shift from a stable equilibrium to an apparently chaotic state of perpetual disequilibrium . . . has some basis in experience" [1950, 362]. He goes on to cite 1929 and the distributional shift toward wages and away from profits and the accompanying unemployment. This, he suggests, could have been caused by a shift in the investment function, say, to IT in Figure 1 (reducing investment) and the consequent shift in the T-curve to I'C (increasing the wage share).

Although Boulding no doubt did not realize it, his analysis of both the relative slopes of the T- and R-curves and their separation precluding their intersection is a classic example of the use of chaos theory! The curves are S-curves that, at best, intersect more than once (multiple equilibria). In fact, in his geometry, there is only one stable equilibrium point; it rests at T[sub r]. There exist unstable extrema at I and C, whereas T[sub x] is an unstable equilibrium.

**Distributional Model II**

In Boulding's distributional model II, variables are expressed in absolute terms. These functions can be expressed as

C = C(Y) (consumption function) (14)

I = I(V) (investment function).

The identities are

Y = C + I = W + V (income identity), where, again, (15)

W = C - T (wage identity)

V = I + T (profits identity).

The wage and profit levels will be different at each level of output. In turn, those differences will result in different allocations of the output between households and businesses. The level of investment (I) is a positive function of profits (V), and consumption (C) is a positive function of national income (Y). Income, output, and expenditures are simultaneously determined.

In Figure 2, output is measured along the horizontal axis, whereas the various components of output are measured vertically. The consumption function is given by CC, showing how much will be consumed at each level of output (income). If the transfer factor is independent of output, equal to CW, the wage curve, WW, lies below the CC-curve by the amount of the transfer factor. This is the absolute wage bill at each level of output, when output is at its equilibrium level. At any output 0Y[sub 1], Y[sub 1]C[sub 1] is the amount consumed, and Y[sub 1]W[sub 1] is the wage bill. If the 45 degrees line OE[sub 2] intersects Y[sub 1]C[sub 1] at L[sub 1], then actual output is Y[sub 1]L[sub 1](= OY[sub 1]), and profits will be W[sub 1]L[sub 1], actual investment being C[sub 1]L[sub 1]. (In this case, there is disinvestment.) The investment function is a positive function of profits (as shown in the smaller figure). At a profits level of L[sub 1]W[sub 1] or 0V[sub 1] in the smaller figure, the planned investment will be C[sub 1]E[sub 1] (= OI[sub 1], in the smaller figure).

This model has less intuitive appeal than model I. Suppose, Boulding suggests, transfers change so that the income distribution shifts in favor of profits relative to wages. Aggregate planned expenditures rise since consumption now is assumed independent of the wage share. In Boulding's diagrammatic example [1950, 266] and in Figure 2, the fall in wages is somewhat less than the change in the transfers, while the rise in profits is somewhat greater.

Boulding separates the effect of the shift from profits to wages into a "transfer effect" plus an "income effect." The transfer effect simply is the amount of the transfer being added to total profits and subtracted from total wages. The income effect is the rise in national income due to the shift being divided between wages and profits, softening the fall in wages while hardening the rise in profits. The share of this division depends on the relative slopes of the consumption and investment functions. The steeper is the consumption function, the greater the increase in income that will accrue to wages. If consumption is reduced by the shift of wages to profits, as well it might be, this counters the output-increasing effects of a shift from wages to profits. If this effect is sufficiently large, there may be net output-decreasing effects, as in model I (to the right of Y[sub x]).

Suppose, says Boulding, consumption rises exogenously. With no change in transfers or the relation of investment to profits, the wage curve will rise by the same amount as the consumption curve. At each output level, profits will suffer relative to wages; this is because the composition of output determines its distribution as in the identity W = C - T. Kalecki is clearer on this in his more complex model: A rise in the wage bill in the consumption sector has no effect on absolute profits, but it will reduce the share of a larger national income going to profits. Only if the wage bill in the investment sector rises would profits and the profit share rise. (Although he does not say it, Boulding implicitly may be assuming that the bulk of wages are paid in the consumer products sector.)

Even so, it is difficult to understand how this model generates an equilibrium output and income level. If transfers vary with business investment and with dividends that, in turn, vary with profits, where would output and income settle? Boulding says as much when he concludes that "the dynamics of all these models is likely to be complicated by the dynamic instability of the transfer factor, and even of the consumption and investment functions themselves" [1950, 269]. The modeling of this instability goes beyond the mathematics used by Boulding.[2]

Supra-surplus Capitalism and Kaleckian Power

**Going Beyond Boulding**

Boulding appears correct in arguing that transfers and thus the financial system can alter the income distribution and the level of national income. Increases in consumer credit, decreases in corporate bond sales to households, and an increase in dividends because of a more liquid business sector will reduce the absolute level of wages. The use of profits as the growth in net worth and the introduction of financial variables at the firm level are a fruitful means to model this interrelationship between household and business expenditures and the distribution of income between profits and wages. It has the merit of taking distribution theory out of the world of marginalism whereby factors, rather than people, receive income. At even the microeconomic level, the refinement of classical economics threw out the Ricardian baby with J. B. Say's bathwater. I would suggest nonetheless that the notion of subsistence household expenditure be resurrected as the complement of subsistence wages, leaving a discretionary wage increment that determines discretionary spending.

Classical subsistence income and neoclassical discretionary income coexist in surplus capitalism. Production surpluses are created when net output exceeds biological necessities and depreciation on replacement capital. In what I [1984, 1987a, 1987b, 1995] have dubbed "supra-surplus capitalism," production surpluses are so great that producers and governments must spend great resources in order to massage the interests and demand of the middle class (and beyond) to absorb supra-net-out-put. When private demands nonetheless falter, federal budget deficits financed by the Central Bank and depreciation of the international value of the currency prop up supra-surplus capitalism.

An excess of income above physiological subsistence leaves a demand wedge that provides breathing space for producers. Imperfect competition abhors wedges and sets into motion forces to fill the space. The price markup tool accomplishes the task. Within the wedge the price elasticity of demand, in part, can be manipulated by producers.[3] At this juncture we can return to Kalecki.

Kaleckian Power and the Lerner Index

Earlier I mentioned the similarity of Boulding's distributional identities and those of Kalecki. In the simplest version of Kalecki, the general income and product identities are the same as Boulding's. However, Kalecki's definition of profits, in Boulding's terms, is

V = dQb = I (16)

Kalecki does not have a transfer item that includes the other forms in which business savings are held. However, Kalecki's idea of the degree of monopoly is something with which the later Boulding probably could agree-the Boulding of the integrative system [1978]. With only slight amendments or liberties taken, we can fit Kalecki's notion of power into the income distribution model. Moreover, this formulation provides an initial value of profits (V).

Kalecki also begins at the firm level in which price exceeds marginal cost by the unit values of capitalist income and overhead. By a standard formula of price theory, the difference between price and marginal cost can be expressed in elasticity terms so that:

(p - mc)/p = l/n = u. (17)

The small letters refer to the firm in which n is price elasticity of demand expressed as positive, and u is the degree of monopoly power or Lerner's index of monopoly power [1934]. Since a markup on average cost can be used to generate the capitalist's income and overhead profits, the equivalent markup version of Equation 17 is

p - ac = p/n = pu = unit markup, (18)

where ac or average cost is substituted for marginal cost.

We can redefine pu as Boulding unit profits since the transfer item includes forms in which profits are distributed or held. Consider a perfectly competitive firm facing an infinite price elasticity: The ratio p/n becomes zero, p = ac = mc, and the firm has zero power to raise price. A firm facing a very low elasticity would have a high degree of such power. In a dynamic model, the producer would use marketing tools to differentiate a product or service-making it seem unique-in order to reduce the price elasticity of demand and increase its power over price.

At the macroeconomics level, let U be a unit markup vector for the entire economy [a column vector of all (p - ac)'s], and let Q' be a row vector of the related outputs. Then,

V = Q'U, (19)

and,

Q'UY'[sup -1] = VY[sup -1], (20)

where, given Q', the larger the elements in U, the greater the powder of firms generally to increase the aggregate profits share or VY[sup -1]. It is easy to show how Q can be determined m an input-output model [Canterbery 1988; 1994] in which average cost and demand also are determined.

Profits as a share of the national product or income is a macroeconomic version of the Kalecki-Lerner measure of the degree of monopoly power, all of which reverts back to the price elasticities of demand. The Q'UY[sup -1] = VY[sup -1] holds because the percentage markup (from the degree of monopoly power) determines the initial profit level. Given the initial level of profits, Boulding's T would further influence the distribution between profits and wages. In this way, the role of market power can be brought under Boulding's umbrella.

Kalecki has business investment as a function of profits and profits as a function of investment (the causality running both ways). But ultimately because firms can decide how much to spend but not how much to earn, investment determines profits. In Boulding's model I, the causality also appears to run both ways. American data for the post-World War II era do show that profits and investments are closely related, but investment lags two years behind profits; that is, the real world relationship between profits and investment does not appear to be simultaneous.

The Transfer Item, Minsky, and Financial Fragility

Surprisingly, there is an unbroken lineage among Kalecki, Boulding, and Hyman Minsky. This genealogy can tell us more about Boulding's T. Minsky [1982] begins with Kalecki's accounting and accepts the Kaleckian markup. In Minsky's elaboration, however, the internal funds so generated by the firm are levered by debt to finance the acquisition of additional capital assets. In the aggregate, investment requires external financing. Thus, investment requires the issuance of bonds, which do appear in Boulding's T. Investment is financed by both internal (Kalecki) and external (Minsky-Boulding) funds.

The income distribution still plays a dominant role because Minsky sees household consumption depending upon household incomes, which includes, for the higher-income households, interest and dividend income from lending to business. The income constraint for the firm is provided by retained earnings and the amount financed by debt (loans from households). If, in the aggregate, the external funding needs of business exceed the household saving made available to finance investment, the shortfall will have to be met by some combination of an increase in the money supply and a decrease in households' money holdings (that is, by an increase in the velocity of money). Since the money supply is endogenous, private banks respond by providing finance and altering the money supply. This fits Boulding to a T!

As I suggest above, the components of T are unstable and lead to a dynamic indeterminacy in Boulding's model that may reflect the real world and Minsky's idea of financial fragility. Rising wage costs during an economic expansion at a constant markup elevate production costs. Since the amount of markup is not unlimited (price elasticity of demand for products at the firm level is not zero), only a generalized inflation can assure full employment. In this process, a rising share of investment is financed by debt. Bankers and businessmen go along with the rising ratio of debt to internal financing so long as they are reasonably convinced that inflation will continue.

Any slowdown in wages does not alter contractual debt commitments so that the burden of debt rises during disinflation or deflation. Debt-financed investment decreases, and purchases of investment goods financed by money supply increments decline. Business firms will begin to pay off debt instead of buying new plant and equipment. Employment falls with the decline in use of the existing capital stock.

A liquidity crisis ensues if the central bank attempts to reduce outstanding debt. Financial panics, such as stock market crashes, may occur as finance becomes more and more fragile. I would note that during the fourth quarter of 1989, the ratio of debt servicing by U.S. corporations to declining profits was about 60 percent. In Boulding's T, this heavy debt burden would have to be reflected as an increased sale of bonds, as indeed it was. An economic recession in this debt environment could lead to financial panic with devastating consequences for changes in net worth of U.S. corporations.

Recently there have been many opportunities for panic. The stock market crashed in 1987. The savings and loan industry began its monumental collapse in 1988. Many giant banks have failed since the 1970s. Giant insurance companies have been failing. Thus far, why haven't we seen a debacle like 1929-19337 Minsky [1985, 17] has provided the answer: the Federal Reserve has served as lender of last resort, as it did when it declared that liabilities of giant banks have a special, protected position in the United States (the Continental Illinois case). Of course, in the unique case of the failed S&Ls, we had a taxpayer bailout. Moreover, a chaos-like collapse of that industry and of the stock market did happen. A model of the private sector that provides a vision of dynamic instability is hardly beside the point. We still do not know what would happen if taxpayers refuse to bail-out banks "too large to save" or what would happen if the Federal Reserve and the Treasury end up with all the Treasury securities now outstanding.

Concluding Remarks

Why has the mainstream ignored Boulding's macroeconomic contributions? First, I offer what might be called "excuses." Boulding's theory is very complex. Moreover, his notation and his mathematics are not always clear. He does not help his cause when he uses phrases such as "the value of real capital," and he makes it difficult to make some distinctions such as that between what I have called the factor-payments transfer function and the financial transfer function. Boulding himself seemed to throw up his hands in his later Economic Analysis, a book that fails to mention either the effects of the distribution of income on the macro-economy or the transfer item. Finally, Boulding's notation and awkward mathematics were faced with a new standard for the use of mathematics in economics set by Samuelson [1947], who made the mathematics of neoclassical optimization virtually unassailable.

Why are these "excuses"? Neoclassical economics has been polished by nitpickers for more than two centuries. Mainstream economists have not rejected a single intuitive neoclassical idea befitting the paradigm, going back to the earliest marginalist genes, no matter how inelegant the mathematics. No less than a John Maynard Keynes rejected only Say's law (which has experienced a revival worthy of "olde tyme" tent meetings) while committing mathematical hari-kari with his own theory of insufficient aggregate demand. Mainstream economists have not bothered to polish Boulding for the more fundamental reasons rooted in their commitment to the neoclassical paradigm.

In this perspective, the neoclassicals can find much to ignore. The mainstream does not like models that collapse, because of the dangerous inference that private market capitalism invites collapse in the absence of reforms. The mainstream does not like indeterminacy because it violates the fundamental premises of maximization (Samuelson's [1947] powerful influence again). The mainstream does not like the idea of a "microeconomic" phenomenon such as the distribution of personal income affecting the macroeconomy. If the income distribution is related to macro-economics, the marginal revenue product explanation is lost and so too the closing of the marginalist system of markets. Moreover, a personal income distribution suggests that there may be income classes: the economist even hinting at such unmentionables can be branded a "Marxist." Finally, the implication that financial transfers can actually alter the income distribution means that, in part, the income distribution is not driven entirely by "pure" economic forces, and Michael Milken was not receiving simply his marginal revenue product.

The complete modeling of a merger of Boulding's T, Kalecki's degree of monopoly power, and Minsky's hypothesis is difficult but no doubt possible. I will stop at this point, a juncture at which I can imagine, at least, that I am ahead.[4]

Boulding's warning about the instability of T perhaps was 40 years too soon. As John Kenneth Galbraith has advised younger economists, for the sake of reputation it is best not to jump to the front of the parade until it is in front of your door. But, I suspect, if Ken Boulding had had 1950 to do again, he would still have preferred to be in the vanguard. Boulding was a genius beyond and above economics. He also was the reliable Quaker among those heading the parade for world peace; that he seemed always to be at peace with himself helped immensely in all these endeavors against the stream.

Notes

1. Canterbery [1987a, 1995] defined the "casino economy," although other writers either before or after have coined the term "casino capitalism."

2. In a later publication, Boulding [1985] seems aware of some of these problems and makes some corrections.

3. Canterbery [1984, 1987a, 1987b, 1995] has described this process.

4. Actually, I did not stop but only paused at this point. Since writing the first draft of this paper for a conference in honor of Boulding, I was inspired by his T and by Minsky's hypothesis to attempt more exact modeling [1993]. A still more mathematical version of the "casino multiplier" [1995] awaits publication. The great difficulty of modeling a dynamic T, even with today's tools, causes one to admire Boulding's pioneering effort even more.

References

Boulding, Kenneth E. "Puzzles Over Distribution." Challenge 28, no. 5 (November/December 1985).

-----. Economic Analysis, 4th ed., vols. 1 and 2. New York: Harper & Row, 1966.

-----. Economics: A New Theory of Societal Evolution. Beverly Hills, Calif.: Sage Publications, 1978.

-----. A Reconstruction of Economics. New York: John Wiley & Sons, Inc., 1950.

Canterbery, E. Ray. "Galbraith, Sraffa, Kalecki and Supra-Surplus Capitalism." Journal of Post Keynesian Economics 7 (Fall 1984).

-----. "The Theory of Supra-surplus Capitalism." Eastern Economic Journal 13 (October/November 1987a).

-----. The Making of Economics, 3rd ed. Belmont: Wadsworth, 1987b.

-----. "Dynamics of Economic Growth." Paper given at Kaldor Conference, Hungarian Academy of Sciences, Budapest, September 3, 1988.

-----. "Reaganomics, Saving, and the Casino Effect." The Economics of Saving, edited by James Gapinski. Boston: Kluwer Acadamic Publishers, 1993.

-----. "A General Theory of International Trade and Domestic Employment Adjustments." International Trade: Regional and Global Issues, edited by M. Landeck. New York: St. Martin's Press, 1994.

-----. The Literate Economist. New York: HarperCollins, 1995.

-----. "Chaos Theory and the Casino Economy Multiplier." Journal of Post Keynesian Economics (Forthcoming, 1995).

Davidson, Paul. Theories of Aggregate Income Distribution. New Brunswick: Rutgers University Press, 1960.

Lerner, Abba P. "The Concept of Monopoly and Measurement of Monopoly Power." Review of Economic Studies (June 1934).

Minsky, Hyman P. John Maynard Keynes. New York: Columbia University Press, 1976.

-----. "Money and the Lender of Last Resort." Challenge (March/April 1985).