An old cartoon by James Thurber shows one of his overpowering women, evidently an actress, sitting down next to one of his mousy men, and saying: “Now let’s talk about you. What did you think of my performance?” With that lesson in mind, my favorite way to pay tribute to an old friend like Eli Schwartz on occasions like this is to reread one of his works and try to hold up my end of an imaginary conversation. What came readily to hand was his 1993 book Theory and Application of the Interest Rate. The tone of this book is distinctive. The mainstream theoretical core is always evident, but the text is full of practical advice and numerical examples. I will try to keep with that way of proceeding. The part of the book where I thought I might have something to add is the final Chapter 10 on “Growth, Profits and the Interest Rate.” (From now on I will refer to the author, rather artificially, as “ES.” The natural “Eli” is too informal for print, and the conventional “Schwartz” is impossibly formal.)

There are a few places where I would be inclined to amend ES’s discussion. To begin with, he says pithily, without much in the way of reference: “Thus, in the pure models, the rate of return on capital is equal to the interest rate; the rate of growth is equal to the rate of savings relative to the capital stock; the amount of savings equal to investment is equal to property income; thus, by definition, the growth rate is equal to the rate of return on the capital stock, or finally the growth rate, assuming neutral technological progress, is equal to the interest rate.” I think this takes too seriously the special case in which all property income is saved and all other (i.e., wage) income is consumed. I don’t think this assumption was ever thought to represent the “pure” or general case. It is an extreme, and possibly illuminating, special case, and it is the so-called “golden rule” that leads to the highest feasible level of steady-state consumption per person in a growing economy. But it is not a good vehicle for more general theorizing. If all wages are consumed and a fraction s of property income is saved, then ES’s final statement is modified to say that the growth rate is s times the interest rate. And, by the way, it is very important to realize that this equation is to be understood as determining the (steady-state equilibrium) interest rate from the growth rate and s, not the other way around. It is only fair to say that in the next paragraph ES remarks that the “savings equals property income” assumption, with possible Marxian overtones, has little to recommend it. In fact, the literature treats it as a pedagogical special case.

Later on in the chapter, in empirical comments that I will come to soon, ES takes pains to point out that actual experience strongly rejects this special case. In fact, property income consistently exceeds investment. It should: the essence of the “golden rule” is the implication that then a higher saving-investment rate would lead eventually to higher sustainable consumption per head. (And, in the opposite inefficient case, if saving exceeded property income, a lower sustained saving rate would increase steady-state consumption per head.)

This is fundamentally a trivial matter, and a digression. ES’s real interest lies elsewhere. Investment is driven by economic profit (or quasi-rent), the excess (or deficiency) of a firm’s net income over (or below) the cost of capital. So how can it be that, in a growing economy, always in equilibrium, the rate of return on capital should just equal the interest rate? The latter can be taken as the cost of capital (with proper adjustment for depreciation, etc.); but then there is no room for any positive profit. What motivates the investment that is needed to keep the economy growing? There is an easy answer to this question in pure theory, and ES had more or less given it just a page earlier in a brief discussion of the classical stationary state. There no net saving is occurring, and one has to wonder why the equilibrium rate of interest should not be zero (except for possible risk premia to safeguard principal). ES goes on at that point: “The only reason for the existence of any net (i.e., positive) interest in such circumstances would rest on the existence of a strong time preference, such that if there were no interest, some of the present stock of capital would be consumed.”

The same sort of reasoning applies in a growing steady state. If investment were to fall short of what is required to keep the return on capital just equal to the interest rate, the capital-labor ratio would fall, the rate of return on capital would rise a little, some positive gap would open up, and additional investment would be stimulated until the pure profit disappeared. If investment were to exceed the steady-state requirement, the same process would occur with all the signs changed. The practical side of ES would notice at once that this is a little too good to be true. Everyday dynamics would see to it that the observed course of events would exhibit at least small fluctuations around the steady state, with the return on capital sometimes above and sometimes below the rate of interest. One force that might keep the deviations small is that some firms would be experiencing positive quasi-rents while other firms are facing negative quasi-rents. This is roughly where ES’s thoughts go, though he doesn’t put them in exactly that way.

A deeper macroeconomic question, far beyond the scope of ES’s book, is whether there are forces in a modern capitalist economy that tend to convert these inevitable small deviations occasionally or perhaps regularly into larger self-reinforcing fluctuations around the steady-state growth path. Schumpeter gave one answer to that question, not much accepted these days. Axel Leijonhufvud sketched a quite different version in his notion of a “corridor,” within which the growth equilibrium is stable but outside of which all bets are off; this idea also failed to attract attention. (One reason may be that a positive theory of macroeconomic policy might be a necessary part of such an investigation.)

ES is interested in profits and interest, not in trends and business cycles. So he proceeds to focus on those temporary deviations around a path of (otherwise) equilibrium growth. The result is interesting and, I think, valid. Growth theorists, including me, working mainly though not exclusively with continuous-time models, have worked out the basic mechanics of an economy whose growth is driven by technological progress. Always, however, it has been “disembodied” technological progress; no innovation-specific investment is required to make new technology effective. My own attempt, 50 years ago, to model growth as requiring the “embodiment” of new technology in tangible investment petered out for lack of either empirical confirmation or good follow-up ideas. ES, without thinking in those terms, arrives at roughly the same place.

By thinking in discrete-time terms and imagining technological progress as a series of discrete innovations, he represents each innovating firm as needing also to make a standard choice-of-capital-intensity investment decision. It will invest to the point where the marginal return on further investment will equal the cost of capital. There is no reason why the usual concavity conditions should not apply, so the supra-marginal bits of investment will earn more than the cost of capital. These quasi-rents will eventually be eroded by competition with newer, more productive, technology; but there would always be a moving float of quasi-rents over and above interest earnings. ES doesn’t think about this whittling away of profits, but taking account of it would not undermine his point. (Schumpeter did, of course, focus famously on the erosion of profits—“creative destruction”—but in his story profits are a temporary monopoly return, whereas ES’s just reflect the normal excess of average over marginal returns to investment.)

This is a useful idea. It could use more working-out than ES can give it in a couple of pages. From the growth-theoretic standpoint it introduces a long-run difference between a model in which technological progress is disembodied and one in which it is embodied in innovation-specific capital investment. Previously there seemed to be only short-run differences, and that is one reason why the disembodied case, though it violates common sense, seemed to be an adequate vehicle for theory. Notice that the point here does not turn on whether technological change is exogenous, but only on whether it has this disembodied or “atmospheric” character. But the more immediate implication, and the one that interests ES, is primarily empirical.

The fact that the aggregate return on capital, as measured by corporate profits, say, is consistently larger than implicit interest on the stock of corporate capital is sometimes taken as at least mildly paradoxical. ES cites James Tobin as fretting that corporations seem to be able to earn 10% a year after tax even while families are prepared to save and acquire assets that pay much less. Doesn’t that suggest that the socially desirable rate of investment must be much larger than the current rate? Tobin apparently attributed a substantial part of the gap to monopolistic distortions. ES first calls attention to the important, though unmeasured, role of quasi-rents in guiding the level and direction of business investment. He also endorses, as anyone would, the Knightian view that (entrepreneurial) profits are the reward for the residual bearing true uncertainty. Tobin would have accepted the general point, but would have pointed out that “social” uncertainty is probably much less than private uncertainty, because much of the latter is zero-sum, losses here being the mirror image of profits there. “Society” is the all-inclusive diversifier.

In considering the relevant empirical magnitudes, ES tends to revert to the national income and product accounts. He insists correctly that it is really the corporate sector that we should be looking at in this connection. One could go further and take the non-financial corporate sector as the appropriate part of the economy. There is a convenient table that appears monthly in Economic Indicators and annually in the Economic Report of the President and decomposes the gross value added of non-financial corporate business (about half of the GDP). In a year like 2000, net interest (and miscellaneous payments) amounted to 3.6% of gross value added. (If all taxes, including corporate income taxes, are deducted from gross value added, net interest amounts to 4.1% of what remains.) Keep in mind that this is net interest as a fraction of value added, not as a ratio to capital. In the same year, profits after tax were 5.8% of value added. (These figures are cyclically unstable: by 2006, after-tax profits had risen to 7.8% of gross value added, while net interest had fallen all the way to 1.9%; clearly long-term bond rates had not fallen by that much, but value added per unit of capital had risen during the business-cycle upswing.)

The fact that these figures are all given per unit of value added is very inconvenient for making and judging comparisons; but at least we know that in a normal year profits after tax can be half again as large as net interest paid by the non-financial corporate sector. That ratio would remain valid if both interest and profits were referred to the same stock of capital. If we identify after-tax profits with ES’s quasi-rents, the interesting question is whether that is a large or a small discrepancy. We do not have to choose among the possible accounts of the origin of economic profits: Schumpeterian monopoly, good old monopoly, Knightian uncertainty-bearing, and the supra-marginal returns on investment emphasized by ES. All could be playing a role, and probably are.

At a purely terminological level, I would be inclined to regard those quasi-rents tied to capital-embodied technological change as equilibrium rather than disequilibrium occurrences. What happens in the aggregate is of course the sum of microeconomic events. If there is a reasonably steady flow of innovations, some of them no doubt more capital-intensive than others, there is likely to be a continuing and perhaps fluctuating flow of net quasi-rents. (I say “net” because there will be some negative quasi-rents associated with obsolescence.) For the individual firms involved, there will be disappointed or unusually favored expectations; and the process may be felt as disequilibrium. For the observing macroeconomist, however, the whole process can be thought of as a sort of stochastic equilibrium.

I mentioned at the beginning that the 1993 book represents a distinctive way of doing economics, with easy transitions among theory, facts, real and made-up examples. It strikes me as a good way to do economics, and it is certainly characteristic of Eli Schwartz. I wish it were characteristic of more economists.