The Economic Theory of Henry George:
A Comparative Review of His Work and His Era
Phillip J. Bryson
David C. Lincoln Fellow
Brigham Young University
For Presentation to the David C. Lincoln Fellows Symposium
Not for citation
The Economic Theory of Henry George:
A Comparative Review of His Work and His Era
This paper reviews Henry George’s theory of distribution as expressed in several of his writings, although the principal focus is on Progress and Poverty. George’s work is not only good reading; it had an impact on the economic discussion of his time and much more. At the same time, it was not a work that provided him admission to the economics establishment of his day. At the time, economists were less inclined to find redeeming qualities to his theories than much more sophisticated, contemporary economists are inclined to admit.
Because the work reads well, one might be inclined to attribute more merit to Progress and Poverty, than it really deserves. If we are to take the work as a contribution to the development of economic analysis, it cannot be evaluated in a vacuum. It must be measured against the leading works of its time. Yet space constraints make it impossible to make more than the most important comparisons to other works of that age.
The choices for such comparison are obvious. The first will be Alfred Marshall’s writings, which in a sense became the most important competition for George’s popular book. Marshall became the first of the neo-classical economists; his work bridged the classical era and the new era of analysis which he founded at Cambridge. George may have contributed the best classical analysis in history, but he brought economics only to the brink of a new world. Marshall constructed the bridge from the classical world to the contemporary world of economics.1
The second choice for comparison will be the works of Francis Amasa Walker, the leading academic economist in George’s America. His works were not to become a permanent part of economics as Marshall’s did, yet Walker would have been the spokesman for all the academic economists, both scorned and ridiculed by George himself and who remained unshakably determined to keep George in the nether regions of the outsider.
A comparison of these historic figures will help us put George into a real context, but we will also want to sketch a picture of other professional views of George, not by his but by our contemporaries. We wish to know how George is viewed with the benefit of time and by those in command of the best of economics available historically – and most of history’s greatest economists are still active. With so many creative minds researching economics across the globe, the discipline can be expected to make continual improvements, which it does. We don’t expect George to be timelessly powerful or timelessly “correct,” but from the standpoint of improved theory and with the greater objectivity of passing decades we can hope to come to a better understanding of George’s contribution.
The contribution was a powerful one. Aside from the fact that he had more of an impact on an international reading public and even on the political developments of his place and time, Henry George was also an important actor, perhaps even the final one in the field of classical economics. Self-taught in classical economics and gifted in its written expression, he might have been viewed as the American father of classical economics, which discipline featured mostly British and European figures. He was not associated with the American academy, being alienated both by his own choice and through the academy’s refusal of recognition. Part of the motivation of this study is to make clear why George is not the father of American economics when he held such a lofty place among the few gifted, early-American economists.
This review of the George contribution will begin with his proffered theory of distribution before moving to a comparison of Marshall’s basic theory and a discussion of Walker’s work. This implies a comparison of the classical and neo-classical approaches to the theory of distribution, ultimately transformed, mostly by Marshall, into a theory of factor prices. It will conclude with a review of the judgments of some renowned “contemporary” economists who have written about George and his works. They are “contemporary” because they include the great minds (such as Josef Schumpeter, a half-century immortal who rightly claims membership in the group) whose writings and viewpoints make up the basic canon by which contemporary economists are educated.
I. George’s Theory of Distribution
The price of grain grown on land of the highest quality consists of wages and interest. But the price of grain from plots of lesser quality must also cover higher costs. Whenever the market price is more than enough to cover the costs of wages and interest on the highest quality land, costs which Henry George refers to as “the margin of cultivation,” a rent will accrue to the landlord. George refers to the price just covering wages and interest on land of the best quality as “the rent line.”
Wealth produced in every community is divided into two parts by what may be called the rent line, which is fixed by the margin of cultivation, or the return which labor and capital could obtain from such natural opportunities as are free to them without the payment of rent. From the part of the produce below this line wages and interest must be paid. All that is above goes to the owners of land.2
Interest, of course, accrues to capital, which is defined as “all wealth used to produce more wealth.” Labor is all human exertion and its return in distribution is called wages.
George and the Wages Fund.
According to classical theory, the “political economy” of George’s time, wages were seen as fixed by the ratio of laborers to the amount of capital devoted to the employment of labor, the so-called “wages fund.” Classical economists conceived of production as a problem of employing workers before they had produced any output with which they could be paid. Current wages would be drawn from advances of capital accumulated before the production cycle began. The actual wage depended on the size of the fund divided by the number of workers to be employed. Wages were also believed to exhibit a tendency to the “lowest amount on which laborers will consent to live and reproduce.”3
But if wages were a function of the quantity of labor employed and the capital devoted to its employment, the classical mind would infer that high wages, the product of scarce labor, must be accompanied by low interest, the product of abundant capital. Or if abundant labor produced low wages, high interest would arise from the scarcity of capital inherent to that situation. George completely rejected this conclusion and the wages fund theory. He wrote of “a general truth that interest is high where and when wages are high, and low where and when wages are low?”4 (37)
George’s attempt to disprove the classical position and to reject the wages fund theory led him to the conclusion that wages, rather than being derived from a wages fund, i.e., an advance provided by capital, are actually paid from the output which labor itself produces.
On this point, George wrote with very practical simplicity of the producer.
Make an exact inventory of his capital on Monday morning before the beginning of work, and it will consist of his buildings, machinery, raw materials, money on hand, and finished products in stock. Suppose, for the sake of simplicity, that he neither buys nor sells during the week, and after work has stopped and he has paid his hands on Saturday night, take a new inventory of his capital. The item of money will be less, for it has been paid out in wages; there will be less raw material, less coal, etc., and a proper deduction must be made from the value of the buildings and machinery for the week’s wear and tear. But if he is doing a remunerative business, which must on the average be the case, the item of finished products will be so much greater as to compensate for all these deficiencies and show in the summing up an increase of capital. Manifestly, then, the value he paid his hands in wages was not drawn from his capital, or from any one else’s capital. It came, not from capital, but from the value created by the labor itself. There was no more advance of capital than if he had hired his hands to dig clams, and paid them with a part of the clams they dug.5
Some have seen in George's criticisms of the wages fund an important insight that production is a continuous process in opposition to the traditional view of classical economics that it is a point-input, point-output process. Naturally, the inflexible "yearly harvest" notion of the earlier classical economists is not an inalienable requirement of wages fund theory. Some economists have historically appeared more understanding of the fumbling analysis of the earlier wages fund economists than of George’s early insight that production theory should be based upon a continuous production function.6
Some later economists were prepared to supply a beating to George for failing mathematically to prove the inadequacy of the wages fund doctrine. The debate continued for some time after J.S. Mill made an initial recantation of the wages fund and others came to the defense of the theory. Contemporary evaluation is contributed by Samuelson,7 who suggests that modern economists should understand that the wage fund should not be “confused with the totality of “circulating capital” and that it is the malleable result of the equilibrium process and not a causal determinant of the level of the real wage in any meaningful long run, intermediate run, or short run.”8
Marshall seems largely to have avoided the whole issue by moving beyond the wages fund and, rather than attacking the idea, quietly letting it die in his writings. The concept fails to receive mention in two chapters on a “Preliminary Survey of Distribution” and again in two chapters on “Earnings of Labor” in Marshall’s Book VI on Distribution. Marshall emphasizes the labor market’s forces of supply and demand as the basis of wage theory, completely ignoring the wages fund. He does, of course, dedicate a brief appendix to the wages fund theory, paying lip service to the idea that it undergirds contemporary thought on wages, although Marshall’s own analysis demonstrates it does not. His relative kindness to the theory may have been in part a slap at Henry George himself, since Marshall well knew of George’s objections to it. He restates the proposition simply, observing that “when anyone works for hire, his wages are, as a rule, advanced to him out of his employer’s capital – advanced, that is, without waiting till the things which he is engaged in making are ready for use.” He then admits that these “simple statements have been a good deal criticized” (nasty Henry!) “but they have never been denied by anyone who has taken them in the sense in which they were meant.” Whether or not denied, it does make sense simply to ignore it, since production and wages can quite adequately be analyzed with modern tools without building on any wages fund foundations.
In his Appendix J, Marshall claims of the “vulgar” form of the wages-fund theory that “the amount of wages payable in a country is fixed by the capital in it.” According to Marshall, that statement cannot be inferred from the conclusion that in agricultural produce, where there is but a single harvest annually, “if all the wheat raised at one harvest is sure to be eaten before the next, and if none can be imported, then it is true that if anyone’s share of the wheat is increased, there will be just so much less for others to have.” The vulgar form of the theory suggested to “the old economists,” Marshall explains, that the amount of wages was limited by the amount of capital, and this statement cannot be defended…It has suggested to some people the notion that the total amount of wages that could be paid in a country in the course of, say a year, was a fixed sum. If by the threat of a strike, or in any other way, one body of workmen got an increase of wages, they would be told that in consequence other bodies of workmen must lose an amount exactly equal in the aggregate to what they had gained.”9
The issue has also been addressed in terms of fixed and circulating capitals and some are convinced that it should be resolved by mathematics rather than straightforward logic. Paul Samuelson, the patriarch of economic theory, notes that Frank Taussig’s Wages and Capital (1896) was hailed by the famous Jacob Viner, Samuelson’s teacher, as a successful vindication of a qualified wage fund – “and even as a successful refutation of Henry George’s muddled notion that production in a steady state can be validly regarded as being timeless and synchronized.”10 But Samuelson graciously adds in a footnote: “A referee reminds me that such great scholars as John Bates Clark (1907) and Frank Knight (1934) also displayed the Henry George muddle that confused steady-state surface appearances with timeless synchronization of production.”11 Henry George does not fare worse in the Samuelson analysis than do other great classical economists as Samuelson respectfully, although frankly, continues the classical debate on some unfinished theoretical issues that need not detain us further here.
Returns to Labor and Capital.
George lumps labor and capital together as recipients of a single share or proportion in national distribution, the other share accruing to rent. Labor and capital received wages and interest for their production contribution and land received rent. The market mechanism would keep the wage/interest proportion of national income roughly constant and wages and interest would rise and fall together. He expressed the idea simply:
For if wages fall, interest must also fall in proportion, else it becomes more profitable to turn labor into capital than to apply it directly; while, if interest falls, wages must likewise proportionately fall, or else the increment of capital would be checked. (186)
So when wages are low, interest must fall as well, or producers will hire labor not only as a substitute for capital, but also to produce more of the higher-return capital. Conversely, if interest declines, wages must do so likewise, otherwise, capital would cease to be accumulated because of the lower returns it would offer investors relative to those of labor. It would also reduce the proportion of capital’s share in distribution because some capital would be diverted from productive to nonproductive uses because of its lower returns.
According to George, as the “margin of cultivation” declines, i.e., as wages and interest decline, or as the costs of agricultural production on the land decline, the share accruing to rent must increase. In the Georgian model, this would be seen as movement along a distribution transformation (or factor returns) curve from the right to the left, with wage and interest shares Figure 1
A Georgian Distribution Transformation Curve
declining and rent increasing in the process. Speculation also affects this model by pushing the margin of cultivation line from the right to the left. In George’s words “the speculative advance in land values tends to press the margin of cultivation, or production, beyond its normal limit, thus compelling labor and capital to accept of a smaller return, or (and this is the only way they can resist the tendency) to cease production.”12 Thus, in the course of “progress,” wages and interest decline and rent increases beyond what would normally be the case.
George held that social progress entails increasing population and land use. There is a functional relationship between population, the expansion of land use and the growth of rent, the return to the factor land. That can be shown in our simple model as movement along a “Progress” function, which simply denotes that social progress includes both a growing population and increasing land use, the latter both to accommodate larger populations and as a reflection of land-intensive social activities. In sum, as progress occurs (population and land use both increase), rent increases while wages and interest drift south.
George on Land Use and Progress
Social forces automatically push society up the population/land use curve, although more capital-intensive production, promoting greater output through the adoption of new techniques, saves labor. So population/land use is a function of increased population, but also of the adoption of labor-saving improvements.
The growth of capital-intensive production is reflected in Figure 3 as a capital curve which is expressed as a function of growth and development along with increasing population and land use. In George’s words, “The effect of inventions and improvements in the productive arts is to
Capital and Technology Effects in Progress
save labor—that is, to enable the same result to be secured with less labor, or a greater result with the same labor.”13
When development and technical change encourage more capital-intensive production, managers will move in that direction. Greater labor-saving investments can be seen as shifting the investment curve downward, as shown in Figure 4. Movement along the progress Line or left along the population/land use axis is a function of increasing population, but also of the adoption of labor-saving improvements. More capital working with labor causes the returns of all factors to increase, i.e., it causes the factor returns curve to shift out as shown in quadrant I of Figure 4.
Continuing with the quote above on labor-saving innovations, George writes “while the primary effect of labor-saving improvements is to increase the power of labor, the secondary effect is to extend cultivation, and, where this lowers the margin of cultivation, to increase rent. All of the progress moves society to the left on the population/land use axis and causes an increase in the capital stock and labor-saving innovations as land use increases. But the labor-saving capital causes the factor returns or distribution transformation curve to shift out to the right. This permits workers and capital to enjoy greater factor returns as seen in Figure 4. Being better off, the workers adopt a life-style that also uses more land.
Unfortunately, however, in the long run, the margin of cultivation declines again, being pulled back to the left. Wages and interest are not improved, but as we would expect in a Georgian view of the universe, rent increases. This is viewed in Figure 4 as a movement from the right to the left vertical line in quadrant I showing the renewed decrease in the margin of cultivation produced by investment and labor-saving innovations. From the higher point on the factor returns curve, the horizontal line over to the progress curve, and the vertical line extending from that point down to the investment curve all reflect greater investments, increased population and land use, but the same tendency toward declining wages and interest and increasing rent. George contends that the long run tendency is for increasing land use, a tendency which does not exclude land use by successful workers, some of whom apparently become landowners as these developments produce subsequent movement up the progress curve.
George summarizes this process as follows:
But labor cannot reap the benefits which advancing civilization thus brings, because they are intercepted. Land being necessary to labor, and being reduced to private ownership, every increase in the productive power of labor but increases rent—the price that labor must pay for the opportunity to utilize its powers; and thus all the advantages gained by the march of progress go to the owners of land, and wages do not increase.14
For George, speculation was an inherent part of the process of development. His negative view of the process assumed that the share of rent in national income would increase partly as a result of the speculation that accompanies economic development and diminishes the returns to capital and labor. It is the reduction of the earnings of wages and interest that produces the down side of the business cycle. Moreover, as speculators withhold land from productive use, they curtail production. And Marshall was basically in agreement with this point in his Principles, admitting that "antisocial" forms of speculation posed a potential threat to economic progress. At the same time, Marshall did not fail to see the positive, market functions of speculation.15
A Marshallian Evaluation of George’s Theory of Distribution
Marshall saw his own doctrine as an extension of classical theory. He perceived George’s analysis to be an attack on or at least a distortion of classical theory. In fact, both were extending classical theory, but Marshall was extending it into the modern theory that undergirds contemporary economics. The Georgian extension simply did not bridge the classical and contemporary eras. When Marshall “built upon” the classical theory, he did not hesitate to demonstrate its inadequacies, thereby establishing the need for his own innovations. When George wrote of such inadequacies, Marshall instinctively defended classical doctrine from Georgian alterations or extensions.
When Marshall analyzed distribution in his Principles, he began by noting how French and English writers over the past century had “represented value as governed almost wholly by cost of production, demand taking a subordinate place.” 16 Their results would not be far from the mark, Marshall observed, in a stationary state. He intended to demonstrate what corrections would have to be made to bring their results into harmony with the actual conditions of life and work. That would largely be to explicate the implications of the demand for labor.
Marshall was proposing that we see returns to factors of production not as the outcomes of macro processes in which factor shares interact somehow, as though at an aggregate level, to determine the distribution of the national income. In the simple model presented above this sentiment appears in the classical simplification of making bundled rent and interest a function not of the individual markets which actually determine prices and quantities but simply of the aggregate of land rent. In contemporary economics each factor’s share in national income is the aggregate of what happens in individual factor markets. Forces of both supply and demand prevailing in those markets determine prices, quantities and aggregate shares.
Contemporary macro-economics is based upon just such an understanding. The micro-economic foundations of factor markets establish aggregates of the factor returns in national income. In classical economics, individual markets were generally overlooked and factor shares were seen as some simple division of the total product. For George it was simply a question of labor’s wages and capital’s interest simply collecting what was left over after the landlords collected their rent. For Marshall it was all a question about the individual factor markets underlying factor returns.
George perceived rent to be a very special case, since the factor land was non-reproducible and strictly limited in supply. Marshall brought economic understanding to a point where land could be treated as just one of n factors of production. To do so, it was necessary to demonstrate that the returns to all productive factors have in common the same basic principles that determine prices in both commodity and factor markets. Those markets differ in specific ways, of course, but their common characteristics make outcomes analogous and predictable.
It is commonly understood today, first, that the cost of an input will be an opportunity cost, i.e., an input must be paid what it can earn in its most remunerative alternative employment. This opportunity cost is the factor’s transfer earnings. Second, any earnings in excess of a factor’s transfer price constitute rent. If the supply of a productive resource is strictly limited and it can be used for only one productive use, transfer earnings would be zero and the entire return would count as rent. Since in reality no agent is incapable of being reproduced or of being adapted to other productive tasks, we must look at the time frame in which such flexibility is to be achieved. Blaug points out that fixed capital earns quasi-rents rather than interest in the short run, since in that time frame the supply of machines cannot be augmented or adaptable to other productive processes. But it’s clear that in the long run new machines can be employed and old machines modified to perform new tasks, so “quasi-rents are always in the process of being eroded.”17 Thus, other factors of production earn quasi-rents on the same basis that land earns rent.
Contemporary economics no longer recognizes any need for special treatment of the factor land or for a theory of ground rent.18 On this basis, Marshall’s objection to the ‘single tax’ becomes sensible. It is that all productive factors, not simply land, earn short-term ‘rents.’ Even Ricardo’s long-run differential rents are incentive payments which encourage economical use of fertile and increasingly scarcer land. Blaug quite fairly observes that George might have responded to this Marshallian reasoning by asserting “that no quasi-rent has either the persistence or the generality of ground rent, and Marshall would probably have agreed with that.”19
But Marshall developed his analysis of land rents in a world in which individual markets determined prices and quantities, which were then aggregated into factor shares without the cryptic macro relationships of the classical world. Thus, factor returns could be seen as an aggregate of individual market decisions made by managers engaged in productive processes. Marshall makes reference to “the alert business man” striving to find “the most profitable application of his resources, and endeavouring to make use of each several agent of production up to that margin, or limit, at which he would gain by transferring a small part of his expenditure to some other agent; and how he is thus, so far as his influence goes, the medium through which the principle of substitution so adjusts the employment of each agent that, in its marginal application, its cost is proportionate to the additional net product resulting from its use.”20
Marshall felt that George’s theory confused cause and effect. According to George, in the course of economic progress, lower average wages are caused by changes in the value of land. In contrast, Marshall explains wage changes on the basis of the theory of competitive markets. In contemporary parlance, as producers strive to minimize production costs, they will watch factor prices and try to balance the ratios of marginal factor productivity to factor prices across all factors of production. This amounts to saying that for any factor of production, the producer is interested in both how much total costs increase with the purchase of an additional unit of a factor and how much an additional unit of that factor will add to total revenues.
An alternative formulation of this principle is that in making the decision to hire additional units of any productive factor, the firm maximizes the return associated with those additional units. The firm will equate the marginal revenue product (MRP) of factor A to the product of that factor’s marginal productivity and the marginal revenue derived from the sale of that output. The MRPa is thus defined as the product of MPa, and the MRx (MRPa = MPa x MRx). More simply, we can say for the moment that, ceteris paribus, an increase (reduction) in the demand for labor will increase (reduce) wages. By the same token, an increase (reduction) in the supply of labor will reduce (increase) wages. Marshall outlined this theory in early lectures he gave on George’s already popular work, Progress and Poverty, but it was in the Principles of Economics (1890) that he fully developed his theory of competitive markets.21
As we have seen, then, the cost-minimizing employer of productive factors will substitute among his purchases of various factors according to their market prices and their marginal productivities. Marshall teaches that “the medium through which the principle of substitution so adjusts the employment of each agent that, in its marginal application, its cost is proportionate to the additional net product resulting from its use.”22
Marshall presented his own version of factor pricing in Note XIV of the Principles.23 Let α1, α2, α3, . .. αn represent different kinds of labor to be used in constructing a home. β, β,’β”, etc., represent different kinds of rooms for the home. V will represent total outlays for productive factors, so V, β, β’,β”, etc., are all functions of α1, α2, α3. H is basically the housing utility or benefit anticipated from the rooms to be constructed, a function of β, β’, β”, etc., and also of α1, α2, α3. For the sake of simplicity we can let H represent total receipts we would derive from sales of the products factor A will help produce. Marshall seeks to find the marginal investments of each kind of labor for each kind of use with the following expressions.
dV = dH dβ = dH dβ’ = dH dβ” = …
dα1 dβ dα1 dβ’ dα1 dβ” dα1
dV = dH dβ = dH dβ’ = dH dβ” = …
dα2 dβ dα2 dβ’ dα2 dβ” dα2
Marshall instructs us that these equations represent a balance of effort and benefit. The real cost to the producer of some small additional amount of labor employed to cut and process timber will be neatly balanced by the benefit accruing to their completed labors. If the principal here decides to pay a carpenter instead of doing the work himself, V will represent not his personal total effort, but his expenditures for the labor employed. In that instance, the rate of pay the carpenters will receive for their additional effort (the agent’s marginal demand price for their labor), is given by dV/da; while dH/dβ, dH/dβ’ are the monetary value to him of the marginal utilities of extra rooms constructed, or his marginal demand prices for them. dβ/da and dβ’/da are the marginal efficiencies of carpenters’ labor in this project. According to the equations, the demand price for carpenters’ labor tends to be equal to the demand price for extra rooms in the home, being multiplied for each room by the marginal efficiency of the carpenters’ work in providing that extra accommodation.
Generalizing this statement, the marginal demand price for hired labor is the marginal efficiency of the labor times the marginal demand price for the product. In other words, wages here tend to be equal to the value of the output produced, i.e., the marginal efficiency of a unit of the labor times the value of the additional product generated. Marshall referred to this as the “net product” of the labor employed. Marshall declares this proposition to be very important, containing “within itself the kernel of the demand side of the theory of distribution.”24
In more current notation,
Marshall’s β, β1, …βn = X, X1, …Xn, or outputs.
Marshall’s α = f, f1, …fn or factor inputs.
Let V = C (cost), H = R (revenues, receipts or benefits).
TC, and X1, …Xn, are functions of f1, f2 , …fn .
H = g(X1, X2,X3) and H = g(f1, f2 , f3). The equality of marginal returns and costs associated with each input type is expressed thus:
(1) MC/df1 = dC/df1 = dR/dX1 ∙ dX1/df1 = MRx2/df2 = MRx3/df3
Marshall says that this expression is a balance of effort (input cost) and benefit (utility or potential revenue resulting from the use of an additional unit of an input). We would express this today in a form which Marshall would have understood immediately. For the competitive case, the value of the marginal product of input a will tend to equal the wage of input a, or more generally, the value of the marginal product of any input, will be equal to the price (cost) of that input.
As we saw above, the value of the marginal product of input a is VMPa (the price of X times the marginal product of input a, MPa). The firm’s optimization is achieved by setting VMPa = wa. Rewriting, px = wa / MPa, or 1/px = MPa/pa or MPa/wa. For imperfectly competitive industries, the marginal revenue product, MRPa is defined as the marginal revenue of the output, MRx, times the marginal product of input a, so that MRPa=MRx (MPa). The only difference between the competitive and imperfectly competitive cases being the use of px in competition (which is to be equated to marginal cost for an optimization of net revenue) and MRx in imperfect competition (which is lower than px and in this case is equated to marginal cost in the stead of the price).
In Marshall’s equation (1), the expression dR/dX1 ∙ dX1/df1 combines two derivatives, the increase of revenues as output increases at the margin and the increase of output as the use of factor input 1 is increased at the margin. The first expression represents marginal revenue and the second expression represents marginal physical product. So Marshall equates the marginal cost of hiring input 1 to the marginal revenue product of that input. These are very contemporary expressions of economic analysis.
Note that the inputs or factors of production are interchangeable. The market for each one operates rather independently of the markets of other inputs. Land adds to production as do other inputs and is not treated differently from them. The demand for an input is a derived demand, derived from the demand for the product which that input helps produce. In short, the Marshallian theory has merely been tweaked, i.e., it has been clothed in modern notation and presented in more restrictive form to characterize the value to a firm of a factor’s output as the product of its marginal productivity and its market price (or in cases of imperfect competition, its marginal revenue).