..V^'VERs/TY OF Digitized by the Internet Archive in 2011 with funding from University of Illinois Urbana-Champaign http://www.archive.org/details/reconciliationof76simo Faculty Working Papers A RECONCILIATION OF THE POPULATION-PUSH AND INVENTION-PULL THEORIES OF DEnOGRAPHIC-ECONOinC HISTORY Julian L. Simon #76 College of Commerce and Business Administration University of Illinois at Urbana-Champaign FACULTY WORKING PAPERS College of Commerce and Business Administration University of Illinois at Urbana-Champaign December 13, 1972 A RECONCILIATION OF THE POPULATION-PUSH AND INVENTION-PULL THEORIES OF DEnOGRAPHIC-ECONOraC HISTORY Julian L. Simon #76 ;.,^M^:..:Uvf 1.^ :^M'Vi A RECONCILIATION OF THE POPULATION-PUSH AND INVENTION-PULL THEORIES OF DEMOGRAPHIC-ECONOMIC HISTORY* Julian L. Simon INTRODUCTION Since Malthus the established theory of demographic-economic history, has been what will be called the "invention pull" hypothesis in this paper. This hypothesis suggests that from time to time inventions appear, inde- pendently of population growth, which increase productive capacity and provide subsistence to more people. Population then increases to use this new capacity xmtil all the productive potential has been exhausted. Ac- cording to the Malthus ian hypothesis, then, the history of population growth is only the reflection of the history of autonomous invention. *Referees were particularly helpful in suggesting important changes in a prior draft. / ,;,,:'y- ' .iT»v*>. s In 1965 Boserup published a lengthy statement of another point of view, here called the "population push" hypothesis. This hypothesis asserts that though production- increasing inventions may occur indepen- dently of the prior rate of population growth, the adoption of "new" knowledge depends upon population growth. Hence in the population- push hypothesis , population growth is necessary for there to be a change in productive techniques.* Whereas the Malthusian invention-pull hypothesis is a single influence explanation of history, the population-push hy- pothesis sees two forces at work: 1) independent invention occurring sometime prior to its adoption; and 2) population growth leading to adoption of previously-unused existing knowledge. Neither hypothesis has anything to say about the causes of the inventions themselves. The two hypotheses imply very different judgments about population growth. The invention-pull hypothesis sees nothing good or necessary about population growth, either in past history or (by implication) in the future. But the population-push hjrpothesis views population growth as necessary, though not sufficient, for economic growth. ^'The kernel of this idea may be found in Von Thunen's explanation of dif- ferences in farming methods in Europe in his classic economic analysis of the inter-relationships of agricultural method and location (1826-1966). The economic elements of the analysis , along with a wealth of supporting data for the peasant economy of Russia, are given quite explicitly by Chayanov (1925-1956). The hypothesis was used frequently by Gourou in his geographic analysis of primitive agriculture in the tropics, by Van Bath (1963) in his history of agriculture in Europe from 500 A.D. to 1850 A.D., and probably by other writers of whom I am not aware. It was Boserup 's contribution, however, to have developed the idea at length. She applied it to the long course or pre-industrial history and is most responsible for the idea getting some attention. Clark also has developed this idea broadly, and promulgated it (1967; 1969). !..• .'■ . The first aim of this paper is to explicate the invention-pull and population-push hypotheses and the economic mechanisms that presumably underlie them, to aid in the examination of their claims to explain economic-demographic history. The niain finding is that the apparent conflict between the hypotheses does not exist. It results from failure to distinguish between the types of inventions to which each of the two hjrpotheses does and does not apply. Put crudely, if the invention-pull hypothesis is restricted to inventions that are purely labor-saving relative to established practice, and if the population-push hypothesis is restricted to inventions that are relatively output- increasing and labor-using, both make micro-economic sense. Together they constitute complementary explanations of complementary forces , rather than being in conflict. The second aim of the paper is to examine some historical and anthropological examples of economic-demographic growth in order to determine for each of the cases whether one or the other hjrpothesis seems to fit best. Enough examples are fourid in which each of the hypotheses fits to suggest that both processes are important in history and economic development . This paper is concerned only with those possibilities that eventuate in economic change. Ignored here are all those possible sequences of events beginning with population growth which do not result in economic cJ-iange, at least in various periods of time. This is very different from the policy -oriented question which asks the likelihood that a given set of events will be followed by economic change in a specified time period — and most especially, the likelihood that population growth will be followed by economic change. Boserup considered the issue in the framework of subsistence agri- culture , as did Malthus , and the focus of this paper is therefore limited to subsistence agriculture, though at least Clark (1967) extends the same argument to the full sweep of human history. INVENTION PULL: THE MALTHUSIAN EXPLANATION OF POPULATION GROWTH \ Though this theory is labeled "Malthusian," it is not explicit in Malthus' Essay on Population . It is, however, the explicit line of thought of most writers on this subject, of which Childe (1937) and Cipolla (1952) may serve as representatives. The argument begins with a society that is somewhat above the subsistence level. Population ex- pands until it reaches the subsistence limits of the technology in use. Then sometime thereafter population size becomes stationary, by mortality alone in the narrow Malthusian view which shall be discussed here, or by mortality and/or birth control in the wider Malthusian view which is closer to Malthus' later editions. Someone then makes a discovery which permits more food to be produced on the same land area, and a shift ta the new technolog]^' then takes place. Population expands again until it reaches the si;tbsistence limit of the newer technology, and so on. If a society is observed to be at a position above subsistence , it is assumed in this theory to be in a transitory state on the way to the stationary equilibrium subsistence-level. The process is shown in Figure 1: a direct causal line from an autonomous invention to a change in the food situation, to a de- crease in mortality, to an increase in population, the process continuing until the new food constraint is reached. Technological change is assumed to have social and economic causes apart from the growth of population itself. ■;v< .-: uJ ■ -u- Figure 1 A particularly clear-cut contemporary version of the invention-pull Malthusian theory of progress and growth is that of Schultz (1964-), who emphasizes a) education that en-ables farmers to use existing technology, and b) developmental research that adapts fundamental inventions to par- ticular situations . According to Schultz the cause of a shift in pro- duction Technique is the arrival of knowledge that the farmer can clearly perceive as being profitable to him, including allowance for risk. (An Importan aspect of Schultz 's arg\ament — as will be seen later — is that individual and cultural preferences with respect to work and leisure are not invoked to explain farmers ' choices of agricultural techniques ; 1965 , pp. ?7-28.) Let us first analyze the basic Malthusian process from the standpoint of the individual, assumed to be a single head-of-household provider vrho has rights to the pioduce of a given piece of land." The agriculturalist can choose between different amounts of product and leisure, as shown in his present production-pQSsi.bility ' frontier Q^Q^ in Figure 2. He also has a set of indifference curve.'^ T^ , T]_, I2, and I3, that show the loci of indif- ference between vardous amounts of leisure and output. *The ensuing micro-analysis has much in common with Mellor's interesting paper (1963), though Mellor's analytic objective was different. The description of early-agriculture tenure systems is not accurate, but the abstraction should cause no trouble here. >-7»'!\> "■ -> ' ^'' '^< Autonomous Invention Improvement in Food Situation -) Decreased Mortality One-time Increase \j in Rate of Population Growth Figure 1 Invention-Pull Malthusian Process For the Subsistence-Agriculture Community -5- The Intersection, Z^, of I and Q Q is the point of highest attainable o o a a satisfaction, so output will be Q , with the given land and accompanying production function Q Q . di 3. ' Figure 2 Now consider what happens if at some later time population increases by a factor of 4, and the average plot is now only a fourth of the previous size. Assuming the same technological know-how but the smaller plot of land, the production-possibility function facing the representative farmer is now q q , half the height of Q Q at each amount of labor.* If he has the same tastes as before, he will now work harder--M. hours instead of M — but produce less — output Q, (though production per unit of land arises) • According to the Malthusian hypothesis, as population increases even more, the -production-possibility frontier facing each farmer will continue to shift downwards until each man is working as hard as he can with the same technology as before and producing only a subsistence output, at which point population growth ceases. A key digression: Figure 2 assumes that there is physical capacity for harder work in some or most cases. And there is evidence to believe that this is so among primitive agriculturalists. Among the Hadza in Tanzania "Over the year as a whole probably an average of less than two hours a day is spent obtaining food..." (Woodburn, 1958, p. 5'+). The "Kung Bushmen of Dobe (Botswana), despite their harsh environment, devote from twelve to nineteen hours a week to getting food" (Lee, 1968, p. 37). And the Australian aborgines of Amhem Land average less than four hours a day in getting food (Sahlins , 1968, pp. 86-87, quoting McCarthy and McArthur, 1960). Some observers (e.g., Thomsen, 1969, p. 180) believe that the state of health *This assumes a Cobb-Douglas production function with exponents of .5, ■--^.1 ., t l... , ! >- { I ..( -'W' V-i-'V' t ■^' «4 1.0 .7 .9 SLASH-AND-BURN AGRICULTURE 11 800 12 800 13 6000 14 6000 15 24000 16 24000 4 8 4 6 6 8 160 320 1200 1800 7200 9600 50 82 540 605 1920 2160 .64 >1.0 .9 ^1.0 .8 .9 SETTLED THREE -COURSE AGRICULTURE 17 24000 18 24000 19 48000 20 48000 21 48000 22 72000 23 72000 6000 6000 4 6 4 6 8 6 8 4800 7200 9600 14400 15200 21600 28800 2160 2410 3360 4320 4800 5040 6480 PANEL B SLASH-AND-BURN AGRICULTURE BEFORE INVENTION I 600 ¥ 1200 ' 1 or>n 300 540 700 >1 9 7 .9 1.0 .7 .9 -8- whether the population is close to or far from that level of subsistence at which further population growth must cease. We must now consider the characteristics of an invention whose adoption would begin immediately no matter what the circumstances. Such an invention must "dominate" the old technology. This clearly must be an invention that will yield the same amount of output for less labor than does the established practice , assuming no change in the amount of land or other capital. There would be no point in increasing the total amount of output in most circumstances , because total calorie assumption is reasonably inelastic.'"' There would not be any point in adopting an innovation that would increase the amount of output per unit of land unless one could reduce the amount of land in use and thereby get the same am.ount of output with less labor. One class of examples that fit the invention-pull hypothesis (type 1 in Table 2) includes replacement of the wooden plow with the iron plow, and the sub- stitution of one shape of digging stick, sickle, or scythe with another and faster-operating shape, because their effects are to reduce the amount of labor required to produce a given quantity of food. Another relevant class includes inventions that yield greater output with the same labor per acre. Such inventions make it possible either to enjoy greater output with the same labor or to cut back on capital and labor and have the same output as before. Examples include a better calendar and seed varieties *The invention-pull and population-push theorists generally assume away shifts in food quality, so that will also be done here. This assumption runs some risk of invalidating the entire issue, however. :..r and plants, such as the potato when it came to Ireland. This is type 2 in Table 2.* Table 2 Invention types 1 and 2 may both be regarded as "labor saving" because they produce the same or more output with less or the same amount of labor (assuming a fixed amount of the major capital element, land).''"'' The effect of such inventions in the context of slash-and-bum agriculture may be seen by comparison of lines 101-104 with lines 105-108 in Panel B of Table 1. It is an important part of the intellectual history of the invention- pull hypothesis that the kind of inventions to which it refers has generally not been made clear. THE POPULATION- PUSH HYPOTHESIS Next we consider the "population-push hypothesis." The key idea is that at a given point in time an agricultural people knows methods of , obtaining higher yields from their lands than given by the methods they use. But the higher-yield methods demand more work. An increase of population then pushes people to adopt the new methods de spite that *The difference between types 1 and 2 in this context is that an innovation that increases productivity with the same land and labor can produce more food per capita, and hence increase potential population. An innovation that only reduces labor and does not increase output cannot provide for more people, but only gives more leisure. **Such statements as "labor saving" are comparisons relative to the agricultural technology then in use. To be labor-saving, capital using, etc., is not an inherent property of an invention, but rather refers to the difference between a given invention and existing practice. Lack of recognition of this relativity has sometimes flawed discussions of technological advance. Table 2 A Classification of Inventions Type Labor per acre Output per acre Examples : 1 Less Same Better sickles and scythes 2 Same or Less More Better calendar; some new seed and plant varieties , e.g., potato 3 More More Irrigation; shorter fallow systems ; some new seed varieties. e.g. 5 new rice •* Less Less (usually) Mechanization Note: The comparative statements "more," "same," and "less" are relative to the agricultural technology previously in use. -10- they require more work. For comparison with Figure 1, the population- push hypothesis may be schematized as in Figure 3. Figure 3 There are two population-push mechanisms that can bring about the shift to the nev; methods. In the first mechanism, there comes to be less land available to each family because of the increased number of families . It is assumed that two technological methods are then known , represented in Figure 2 by Q Q and Q.Q, for population-density state 0, or by q q andq,q, in population-density state 1. Consider a representative head of a house- hold in population-density state 0. Method Q Q gives higher yields at lower labor levels, whereas Qt,Q^^ gives higher yields at higher labor levels. In state 0, the family will choose to be at Zq with production method a because it hits a higher indifference curve, I . than can be achieved with production method b. But vj-hen population increases and the land available to each family shrinks to the point that it is in state 1, it chooses productive process b with which q, touches indifference curve I,, at Z_. This requires the family to work M„ rather than M^ hours , but yields output Qj instead of output Q, . The second population-push mechsinism refers to an increase in the size of the representative family, assuming the same size plot of land. If there are, say, six rather than four and a half consumer equivalents in the family, it is reasonable to assume that the subsistence level will be higher and that there will be a shift in people's indifference curves, perhaps from the I set of indifference curves to the f family. With the n. .-I O :.■- ;pf Fertility / Food Situation V Changes in Technology 1 Increase in Rata of Population Growth / Population-Push Hypothesis For a Sxibsistance Agriculture Ccnmunity In the Presence of Unused Technological Knowledge Figure 3 -11- original plot of land, the optimization point for the I indifference set is at Zq using technological method a. But the optimization point for the I in- difference set is at Zg, a point on the productive-possibility frontier at which method b is used and which produces output Q,, which is more than Qq. The population-push hypothesis implicitly suggests that population growth puts both of the above mechanisms into operation. And both mechanisms work in the same direction to induce a shift to previously-known but previously- unused production methods. To discuss the population-push hjrpothesis in a more aggregated manner, let us start with a hunting and/or food-collecting group of ten people on 800 square kilometers of land. This is ample to support them with a minimum of labor, say tv;o hours work by each person on 250 days a year, as shown in row 1 in Table 1. (In arctic and desert regions food gathering is much more laborious, of course.) With a group as small as ten people the workings of chance necessarily make for some variation in population size over time, leading either to the disappearance of the small family group or to its increase. If population rises toward forty people the group will find it harder to make a living because game does not increase with the population size — probably decreasing, instead, as hunting is intensified. The workers now need to work four to eight hours a day instead of only two hours a day, to kill or collect the same amount of food per person as before (rows 2-^). And if population then grows even more, it will be difficult or impossible to make a living by hunting and collecting, no matter how hard the people work. According to the population-push hypothesis , what now happens is that the group begins to farm a bit. The data shown in rows 5-10 are ,„ Kv "■;-' -12- intended to represent a very primitive kind of agriculture , perhaps that which was done at Jericho or Jarmo where little or nc land clearing needed to be done. Because of the shapes of the idealized production functions shown in the table , the forty people could get more food from four hours fanTiing per person than from four hours hunting axid collecting. (Compare rows 3 and 6. But note that hunting and collecting would still provide more food for the' forty people with a two hou r work day than would farming, rows 2 and 5). Please note that the population-push hypothesis assumes that the group already has the technological knowledge necessary for this primitive agriculture long before it shifts to the new technique. The shift from one food-getting method to another need not be sudden or all at once. Rather, the shift may be gradual, as is consistent with the basic economic logic implicit in the production function. The forty people could minimize the time required to get the basic food ration by employing both methods at once — at least until population grew much larger. This is consistent with the observation that in many primitive communities the women farm while the men hunt (Clark and Haswell, 1969, p. 27), The idealized scheme in Table 2 traces the process as population on this 800 square kilometer tract expands to 200 people, according to the population-push hypothesis. At that point all work would be primitive agri- culture and none would be hunting (rows 7 and 8). Then population expands to 800 people , which nears the limit for this type of simple farming (even if animals have been tamed and kept in pasture.) The popula- tion-push analysis suggests that in the course of the expansion to 800 people agricultural technique shifts again, this time to slash-and-burn agriculture. V/ith short work days, slash-and-burn agriculture produces less food than the more primitive agriculture (row 9 versus row 11). -13- But if the 800 people work somewhat harder, they can get a full ration of food with less effort using slash-and-burn long-fallow agriculture than using a more primitive system (row 10 versus rovj 12). So — assuming as the population-push writers do , that people already know about the method — people will now gradually shift to slash-and-burn agriculture. And the shift to slash-and-burn will be completed as population expands to, say, 6000 people on the 800 kilometer tract. Then, as population grows even more, there will be a shift to three-course agriculture. These successive shifts from one t3rpe of agricultural system to the next , each system demanding more labor per worker and a shorter fallow period than the one before, and each shift caused by growth in population, are the heart of the population-push hypothesis. A variant hypothesis, closer to population-push than invention-pull, is that technological development in agriculture is induced by changes in demand elsewhere in the economy. For example, a rise in the degree of urban- ization, ceteris paribus, increases the demand and hence the price for agricultural products . Changes in crops and methods of farming may be caused thereby (Dovring, 1966). This hypothesis is clearly more appli-ca- ble in stages later than subsistence agriculture , so the matter will not be pursued here. Now let us consider the characteristics of the kinds of inventions that might not be adopted until some time after their discovery and as a result of a population-push. These will not be labor-saving inventions, because there is every reason for the adoption of labor-saving inventions to begin immediately after discovery. Instead they will be inventions that meet the need of producing more output as population grows and which require more labor--the latter being the reason that their adoption did x. -14- not begin immediately. The most important examples of such inventions (type 3 in Table 2) are short-fallow and multi-cropping systems , all of which require more careful farming and more hours of labor per unit of output than do long-fallow systems. Some of the new seed varieties, e.g. , rice that requires the laborious Japanese method of wet-farming, are related examples. ANALYTIC EVALUATION OF THE TWO HYPOTHESES ■ To recapitulate , the invention-pull hypothesis is a reasonable description of the process that begins with exogenous labor-saving inventions. These Malthusian-type inventions are immediately "profit- able" and dominate previous technology in the sense that they are better in all relevant ways. Therefore, it is reasonable to suppose that they will be adopted immediately, and will later lead to new population growth. And the population-push hypothesis is a good description of the place of labor-using inventions in economic-demographic history. In the population-push hypothesis , the key technological discoveries do not dominate the old technology at the time of invention, and in fact are less desirable at the time of invention because of their labor intensivity. The new invention becomes desirable later on, after population density has increased and total food "needs" have increased, according to the popula- tion-push hypothesis. This brief analysis of the difference between the hypotheses focuses on a) the properties of the key inventions, whether labor-using or not; and b) people's "needs" for food and leisure at the time of the inventions. Now, we know that some inventions are almost purely labor-saving (e.g., implements such as the bronze and iron sickles and scythes; see Curwen and 7.1-1 ■ '■■ ■• ■ai; :■■!■- "J -15- Hatt, 1953, especially p. 9^). We also knov*/ that other inventions are heavily labor-using (e.g., wet-farming and double-cropping). The fact that both types of inventions have been important in history implies that both invention-pull and the popluation-push mechanisms have operated in history. The two hypotheses do not conflict with each other. Hence, they should not be seen as competitive explanations of pre-modern agri- cultural development, but rather are complementary. As to the relative importance of the two mechanisms in history, there is at least one fact indicating that both vrere important: Even after much change in technology and after much population growth , the amount of work done on yearly crops by individual farmers in densely- populated places is not higher by a different order of magnitude than in sparsely-populated places . In both situations farmers work most of the daylight- hours during the harvest season and much less dur5.ng most of the year. This suggests that invention-pull was not the whole story, because if it were, all inventions would be labor-saving and the work done would have become less as agriculture advanced. On the other hand, the fact that the amount of work per farmer has also not increased enormously suggests that the population-push mechanism is not the whole story either, because the technological changes it describes all require more labor than before. This demonstrates that both mechanisms must have been at work in history to an important degree. The amount of work done on yearly crops is not, however, a very good measxire for comparison of the two hypotheses in situations of more advanced agriculture. This is because long-run investment becomes progressively more important as agriculture advances. The investment in slashing and burning in slash-and-burn agriculture is all used up after r- ;,. :;>7' f! •- ' >(\ jT-'V -15- two or three years. But the stone clearing, stumppulling , digging of irrigation wells and ditches , polder-making , placing drainage tiles , and so on that are necessary for more advanced agriculture are very long- lived investments — and they also require enormous amounts of back- breaking labor. Very little of this labor (which tends to support the population -push theory) shows up in the yearly time-budgets in settled agricultural regions. This means that data gathered about the labor spent on current farm operations in the context of advanced agricultiire does not help discriminate between the population-push and invention-pull hypotheses. And another point about Investment: its absence from Boserup's theorizing about the population-push hypothesis is an important lacuna in her scheme. A word about more advanced agriculture : Neither of the two hypotheses describes the development of modern annual-crop and multi-crop agriculture in which output per farmer rises .* The invention-pull point of view on this obviously is that tY^e rate of exogenous inventions sped up for exogenous reasons. This comprehends the development of the internal combustion engine and many of the other components of farm machinery. - But an increased rate of exogenous inventions has no logical connection with the invention-pull hypothesis. On their side, population-push theorists *Part of the explanation clearly is that the number of farmers does not rise anywhere near as fast as total population, which is tied up with the development of the industrial sector. But this is because of the increase in agricultural productivity as well as because of the increase in industrial opportunities. There certainly have been increases in output per farmer holding the amount of land constant, which is the phenomenon for which an explanation is sought. -17- point to social changes that occur as people progress to more advanced agriculture. More economical social organizations with more division of labor evolve as people live closer together. Roads are built. Most important to Boserup and Clark, however, is that changes in individual psychology take place, people becoming more willing to work hard over sustained periods . Together with the other social changes , this is the engine that eventually brings about a "take-off" far beyond subsistence and near-subsistence living, according to Boserup and Clark. But these changes are not inventions , or if they are inventions , there is no lag between invention and utilization. These changes are induced just as new seed varieties are induced by U.S. market forces and by purposeful foundation or government research stimulated by foreseen needs. Of course there is a lag between the invention of the tractor and its use in India, but that is of the same nature as the delay between the invention of the fork-lift truck and its use in warehouses in India. The point is that after the passage from mostly-subsistence to raostly-market agri- culture, neither the invention--pull hypothesis nor the population-push hypothesis has much to say that is distinctive and useful. The e.xplanation of the development of business farming is like the explanation of the development of other business enterprises — with regard to the effects of both the increase in demand and the increase in technical knowledge. , The analytic discussion up to this point might suffice to make the argument put forth by this paper. Nevertheless, it seems worthwhile to examine some historical cases to provide some relevant, concrete evidence, too. This will be done in the next section. ('^v I :' ' ■■•--", ■■>.,'.! -18- SOME HISTORICAL AND ANTHROPOLOGICAL EVIDENCE 1. Early Biblical Period Genesis makes very clear that the shift from food-collecting to tilling the soil was perceived as an increase in labor- intensivity. "Behold, I have given you every herb yielding seed, which is upon the face of all the earth, and every tree, in which is the fruit of a tree yielding seed — to you it shall be for food." (Genesis 1:29). But then: "...cursed is the ground for thy sake; in toil shalt thou eat of it all the days of thy life. Thorns also and thistles shalt it bring to thee; and thou shalt eat the herb of the field. In the sweat of thy face shalt thou eat bread, till thou return until the ground;" (Genesis 3:17 - 3:19). According to the Bible, the technological shift in the Garden of Eden was caused not by population growth >ut by sin. Aside from that, however, the description is that of Von Thunen , Gourou , van Bath and Boserup. The rest of the story in Genesis may be prototypical, at least for people in hot , dry , coiontries . If the tribe flourishes rather than being overcome by mortality , there come to be a few more members of the family and tribe than before, and the people then use a somewhat larger area o'f land for pasture, hunting and gathering, all without a major decrease in product per person, probably. And there is piK>bably some gain in efficiency from their cooperative efforts in hunting and defense. But if there come to be enough people in the tribe so that at the edges of their area the land is less bountiful, "personal income" will fall. When this happens the tribe will probably split as with Abraham and Lot ; one group will move to different grounds . The group that moves is likely to have poorer prospects than the group which remains , and the movers will certainly incur a cost in moving. Hence the overall level of -19- per capita income apparently falls over the last part of this population expansion, or at least people v;ill have to work harder on poorer hxinting and collecting areas to sustain the same standard of living. Overa ll judgment: The Biblical period is consistent with the population-push hypothesis . 2. Other Hunting And Gathering Groups A crucial element of the population-push hypothesis is that there is pre-knowledge of techniques that will increase aggregate output from the given land area. But there are at least some situations where such techniques are clearly not known to hunters and gatherers. One such situation was that of the Netsilik Eskimos in the 1920 's and 1930 's, whose situation "did not leave the Netsilik with much leisure" (Balikci, 1968, p. 82). Life was very hard for the Netsilik by any material test. This does not fit the population-push hypothesis . And before agriculture was invented no one knew about it. Of course it is conceivable that in all societies agriculture was discovered before it was needed, but this hardly seems likely. The control of plant reproduction is nowhere near so obvious that anyone who could benefit from it would automatically know about it. All cultivated plants have been derived from, wild varieties, and the first step in cultivating them consists in taking the seeds of the wild plants , soviing them in suitable soil, and by care and. attention promoting their growth... It may seem incredible to us that for countless thousands of years it never occurred to man to take this simple first step, but the very fact that he did not do so for so long emphasizes the magnitude of the discovery that eventually led him to take it. (Curwen and Hatt, 1953, p. 15) But the knowledge of agriculture was certainly available to some per~agri- cultural groups before their population was pressing on its limits, and the -20- knowledge came to be used as population grew. The knowledge of agriculture has been showri to have diffused from one or a few discovery points in Eurasia — though also independently in the Americas, etc. Hence, some groups are likely to have learned of agriculture while they still had plenty of land on which to hunt and gather. It should not surprise us, therefore, that some or most hunting-and-collecting tribes live a comparatively easy life, as suggested in the examples adduced earlier. Overall judgment : Some instances of hunting-and-gathering groups illustrate the invention-pull hypothesis , and others illustrate the population-push hypothesis. 3. Slash-And-Bum Agriculture Slash-and-burn agriculture uses land more intensively than does collecting or hunting, but still not very i.ntensively by standards of developed agricultural sectors today. As done today in the tropics, "The first trees are felled with axes , and when dry the vegetation is burnt; after the crop harvest the parch lies fallow and the forest regains control until it is once more burnt" (Gourou, 1966, p. 31). The plot is then farmed for one, two or three years in succession, and then the land lies fallow for up to 30 years but sometimes as little as two or three years, "...between eight and tv;elve years are necessary to get a good cover of v;oody vegetation" (Gourou, 1966, p. 38). Meantime, the ■ agriculturalists work other land. The important point here is that a large area is needed to supply food for a group of people — up to 30 times as much as is cultivated in a single year. Vfhether in earlier times practitioners of slash-and-burn agricul- ture generally knew of more "advanced" methods vre cannot know for sure. cl'ir) \[l'. «•' 'Ow . i'- . '.. -u- -^ tnl' -21- But there is no reason to doubt that present-day slash-and-bum agricul- turalists know about other techniques , as shown by the fact that the two systems coexist in some places (Gourou, 1966, pp. 107-108). The reason agriculturalists still prefer to continue using the slash-and-bum technique, according to Gourou (1966, p. 52), Clark and Haswell (1957, Chapter VII), and Boserup (1965, pp. '+H-'+8), is that, for the quantities of food produced on the available land, slash-and-bum agriculturalists work fewer hours than if they were to use more intensive methods (at least, up to machine methods) to produce the same amount of food. And there is also much evidence that when population density increases, people do shift from slash-and-bum to more intensive shorter- fallow agriculture. One of the most persuasive pieces of evidence is that when population density decreases - -perhaps due to an increase in tribal safety on low lands , an event exogenous to population growth — the process reverses and people go "back" to less -intensive longer-fallow methods (Gourou, 1966, p. 107; Boserup, 1955, pp. 62-63). Other persuasive evidence comes from anthropological restudies. When one observes a society at only a single point in time , or over a very long sweep of history, one cannot be sure that it is not the change in agricultural technique that is the leading force , with population growth only following. But if one observes a society at a point in time, notes an increase in population taking place , and then observes a generation later that a shift in agricultural technology occurred, this argues strongly that it was population growth that was the cause of the change in technique rather than the reverse — especially when the farmers tell the observer exactly these motivations for the change in method. Chan Kom, a village in Yucatan, Mexico, was studied by Redfield and Rojas in 1931, and then by *u;'' „.'■:■'■: -22- Redfield in 191+8. Population increased from about 250 people to about, 445 during those seventeen years. In 1931, there was unappropriated land available for anyone to farm (1934, p. 42). But by 1948 the situation had changed: "So long as there remained to the south an unpopulated territory, it vjas possible to take care of the increasing num- bers of people by the making of new settlements on that fron- tier. The people of a village would go farther and farther from their home community to find good land... the lands available to the village yield less than they did twenty-five years ago... as land becomes scarcer, it is planted again in a shorter interval after its last abaDdonment to bush... The average size of railpas [maize plots] is surely smaller now... about one-half as large as they used to plant... Four or five of the men have begun the development of such small tracts by planting fruit trees , tom.atoes , beans , and other small crops , by building cattle corrals , establishing poultry on the tract, and, in a few cases, by digging wells .. . • The people see the population press upon the resources. So far as they speculate as to remedies , they turn to the possi- bility of increasing resources. Don Eus sees a hope in diver- sified agriculture." (Redfield, 1950, pp. 54, 55, 57, 171, 172) The changes in Chan Kom fit the population-push hypothesis very well. Another restudied village is Tepoztlan , in the Mexican State of Morelos. The picture one gets from the study in 1926-1927 (Redfield, 1930) and restudy in 1943-48 (Lewis, 1951) is somewhat different bur essentially consistent with that of Chan Kom. Unlike Chan Kom in which only slash-and-burn agriculture was practiced, in Tepoztlan as of 1930 both slash-and-bum agriculture and no- fallow plow agriculture were practiced. In the former technique a hole is made with a dibble or hoe for each seed; in the latter technique the ground is broken with a plow pulled by oxen, and the seeds are dropped into the furrow and then covered with the foot. -23- Slash-and-bum requires much more land, but it also requires much more labor per unit of output--at least in Tepoztlan in the 1940's. Plow culture requires more capital in the form of oxen and plow. Plow agriculture began in Mexico only after the Spanish came; before that the plow was un- known. After the Spanish came there v;as also a massive depopulation of Mexico generally, and of Tepoztlan specifically (Lewis, 1951, pp. 26-30). Therefore farmers must have adopted plow agriculture not because of increasing pressure of population on land, but simply because with the innovation one could get more output with less labor, even though using much less land. This conclusion is consistent with the existence of much \infarmed land in Mexico even until the recent very fast population growth. Hence, the shift after the Spanish came to a system of no-fallow plow agriculture is not consistent with the population-push view of history, but rather fits the invention-pull hypothesis quite well. In the years between the Redf ield and Lewis studies of Tepoztlan , population grew rapidly (Lewis, 1951, p. m8). People responded with an increase in s lash -and-b urn agriculture in fields far away from the town, which contradicts the specific assertions of Gourou and Boserup about the sequence of shifts in technique. But it is also true that the daily trip to the milpa requires many hours back and forth ; this trip increases the labor input so that it must be much greater per unit of output than in plow agricultiire , and in this respect the shift fits the population-push hypothesis. Tepoztlan is also consistent with the population-push hy- pothesis in that knowledge of several techniques is available, and the agriculturalists choose that technique which maximizes their utility from production and leisure, given the land available to them and the indif- ference curves that reflect the number of children that they have. -2H- Overall judgment : Most slash-and-bum agricultural situations fit the population-push hypothesis, but some fit the invention-pull hypothesis. M-. Polynesian Short -Fallow Agriculture Taro is the staple crop on the isolated Polynesian island of Tikopia, which Firth studied in 1928-1929 and again in 1952. As of 1952 the Tikopians were worried about the press of population .upon food resources because of population growth. "The taro resources of the people might possibly be increased by adopting a technique of irrigation and conservation. . -But this would require. . .specific instruction by an external agency" (Firth, 1965, p. 50). Firth implies that the Tikopians really do not know of a technological alternative that, even with increased labor, would expand their food supply greatly. Instead they look to birth control and suicide-migration by canoe to adjust population and food resources. The physical isolation of this island is obviously a key element in its lack of alternatives , both for technological change and for migra- tion. Overall judgment : Tikopia 's situation fits the Malthusian invention- pull hypothesis better than it fits the population-push hypothesis. 5. Shifts in Fallow Periods and Crop Rotations in Europe The highlights of the history of European farming from 500 A.D. to 1850 aTQ as follows. Population hit its nadir, 19.3 million, near the beginning of the period (Clark, 1967, p. 64-). From then on the secular population growth was continuous , though not without major setbacks from time to time . For the most part , then , we may associate changes in farm- ing methods in the Middle Ages with increasing population. •to - ?arf i:, hi: ; -■' 1. -25- A key change was from two-course to three-field rotation methods. In the former, "land was tilled or left fallow in alternate years." In the latter, "...winter com (wheat or rye) was sown the first year, spring com (barley or oats) the second, and in the third year the ground lay fallow" (Van Bath, 1963, p. 59). The three-course method increases total output per \mit of land because more land is in production each year. Van Bath asserts that "It was nearly always at a time of increasing population that there was a changeover from two- to three-course rotation, for instance, in England in the thirteenth century... (1963, p. 60). And in some places where it fitted farmers' needs, "both systems were mmning concurrently, or being alternated according to the crop: wheat in the three-course rotation and rye in the two-course." (Van Bath, 1963, p. 60). Boserup generalizes from this that "virtually all of the methods introduced in this period has been known before-hand" and awaited adoption until population density grew sufficiently (1965, p. 38), .Chayanov (1925-1966) explored in great analytic depth and factual detail the workings of the piish of population and consumer wants upon the behavior of Russian family- farmers at the turn of the century.* Chayanov investigated the "annual labor expenditure" of all the workers taken together on each farm, and found that the chief determinant, given any set of "production conditions", was "the pressure of family consumer demands on the workers," (p. 76) i.e., the number of consumers in the family. "This forcing up of labor intensity, buying increased annual *These farms averaged 20-50% of their income in money and the rest in subsistence in the various areas Chayanov studied (1925-1966, p. 121). This probably makes them as much "subsistence" farms as most of those in Asia today. '.'r. 'I Jtftr(v. :;r. ilTnu noi^q'jlif ■..'jj.Wi, i «;i? nM;-t; iZTi-fi^'l •Vr.f'rti^.nv t."> 1! ■' .\'i. ,r ,-7^ '-T" ■*■ . vnl -:■ ", ■■'■■■■.I :,.;:. o ^-'^ . ..(: •'j' *' '(.' ) i,\. 1 -26- agricultural income at the price of reducing labor unit payment , is achieved either by an intensification of work methods or by using more labor-inten- sive crops and jobs" (p. 113). It cannot be emphasized too strongly that Chayanov's conclusions are based, not on casual observation, but on the extremely detailed careful surveys of Russian agriculture that were made around the turn of the century by skilled, dedicated statisticians- surveys which still constitute perhaps the best available source of data on matters such as this one. And though Chayanov's data, with a few notable exceptions , cover a single point in time , it is reasonable to assume that the cross-sectional differences recapitulate the changes over time as population density increased on the Russian peasant farm.* Von Thunen described with wonderful precision the Belgian and Mecklenburg systems of cultivation , and clearly showed how the difference in techniques used was related to population density (1863-1966, p. 85 f f ) . But details will be omitted because this segment of history belongs more to market than to subsistence agriculture, which is the focus of this paper. But population-pushed changes in crop rotation and related work methods are far from the whole of the story of European agricultural history from the Middle Ages on."" The development of the heavy plow and of horsepower to replace oxenpower were, according to some historians, equally important with the three- course rotation in the agricultural revolution. The heavy plow required iron for its blade, and the horse *I am grateful to James Millar for bringing Chayanov's book to my atten- tion. **This following section is based on V/hite (1962, Chapter 2). I am grate- ful to Larry Neal for bringing it to my attention. \^' I . .'■•■''•? .■.(•"' ', -27- required both iron for horseshoes and the new chest-harnesses that were being perfected at that time. General use of both the heavy plov? and the horse had to wait upon the increased supplies of iron that occurred in Europe at the beginning of the Middle Ages. As iron came to be available, these innovations — together with the new harnessing and the three-course rotation -- enabled European farmers to work profitably with the heavy soils of more nqrthern lands than had previously been cultivated. According to White (1952, Chapter 2), the inventive knowledge under- lying the abailability of iron and the adoption of the heavy plow, new harness methods , and horsepower were not available before the Middle Ages , in direct opposition to Boserup's assumption. But further. White asserts that the three-field system was also invented in the Middle Ages. He cites the fact the "Charlemagne himself thought of the new pattern... as something so new and significant that he felt impelled to rename the months in terms of it" (1962, p. 69). White's account is very much one of new inventions pulling upward the size of the population, especially in Germany and Scandinavia (1962, p. 51), rather than of population pushing adoption of previously -known methods. New crops are still other exogenous technological advances which did not have to wait on further population increase in Europe for adoption. "The earliest type of technical change that came to modify medieval agriculture in Europe was in the introduction of new crops. Some of these changes were the pre-condition for changes of other kinds" (Dovring, 1965, p. 631). rne potato had the most dramatic impact. "...[0]n a pathetically smcLLl patch of ground one could grow in potatoes from two to four times as much food as one could in terms of wheat or other grains , enough indeed to feed a family of more than average size... It was introduced [into Ireland] about the year 1600 and before the end of the seven- teenth century had been generally adopted by the peasantry. ■'.."j .'/:/ '. .■>; .1 .r;n -28- By the end of the eighteenth century the common man was eating little else... The un- speakable poverty of the country should, it would seem, have militated against any considerable population increase. Yet the population did increase from 3,200,000 in ITSM- to 8,175,000 in 184-6, not counting some 1,750,000 who emigrated before the great potato famine of 18H5-1847. It was perfectly obvious to contemporaries , as it is < to modern scholars , that this Irish population could exist only because of the potato. Poverty-stricken though it might be, the Irish peasantry was noteworthy for its fine physique. Clearly people were doing very well physiologically on their potato fare. Young people rented an acre or less for a potato patch. On the strength of this they married young and had large families . " (Langer, 1968, pp. 11 and 15) ' Though Lai'iger emphasizes the effect of potato cultivation in increasing fertility, other v/riters (e.g., McKeown and Brown, 1955) emphasize its effect in reducing mortality, especially infant mortality due to malnutri- tion. Whichever is correct, there is little dispute that the "invention" of the potato and its diffusion to Ireland pulled population to a larger size than it would have reached otherwise. • Ov6i''3-ll judgement : There were some innovations and events during the Middle Ages in Europe that apparently support the invention-pull hypothesis , and others that apparently support the population-push hy- pothesis. The relative importance of these innovations and the dates of their inventions are subject to scholarly dispute, however. Therefore, an overall judgment is difficult to form — perhaps because more is known about this situation than about other situations discussed here. 6. Traditional Agriculture in Ch ina" The population of China increased over the almost six hundred years from 1368 to 1957 from 65-80 million to 647 million. But per person *This section is drawn from Perkins (1968, particularly chapters 1 and 9). -29- consumption did not decline secularly over this period, and perhaps rose. No fundamental shifts in mechanical technology took place. Some new crops were introduced, but none had a revolutionary effect. Opening up new lands accomodated the largest part of the population growth. Extension of water-control systems and double cropping contributed less , and both of these methods had been well-known before 1358. These facts are more consis- tent with the population-push hypothesis than with the invention-pull hypo- thesis. This Chinese history raises a puzzle. If the Chinese farmers were apparently able to increase total production under the pressiire of more people, why did they not raise production even faster? One possible answer is that the parallel trends of production and population are just a coin- cidence. But this coincidence seems unlikely because the methods of raising production v/e re known. Nor can one satisfactorily explain the relationship by the increase in labor alone, as Perkins shows (1968, pp. 79-- St) ; such an explanation would require that there are no ceteris paribus dirai.nishing returns in agriculture, which there clearly are. The most likely explanation why per-person consumption did not rise even faster, I judge, is that the Chinese farTners' preferred not to trade more work for more output. This choice may (ov may not) have been heavily influenced by the absence of consumer goods which would whet their desires. In "any case, however, the outcome fits the population-push hypothesis. The extent of change in the individual Chinese farmer's behavior as a result of "population pressure" may seem from the statistics greater than it actually was, because it was surely the young families just starting out who supplied much of the energy. If Chinese life was like Irish life in the 18th and early 19th centuries (Connell, 1965, pp. i+28-i+29) i}'J-' -Ki •,. 7V ■i-.i .1 , -30- it was the young just-marr'ied men cuid women who cleared uncultivated lands and settled there. The increase in output required of a mature father was much less than proportional to the number of persons in his nucleus family, especially after the children grew up. Overall judgment; The last six centuries of Chinese history are well-explained by the population-push hypothesis. SUMMARY AND CONCLUSION Two apparently-conflicting hypotheses --the invention-pull hypothesis and the population-push hypothesis — have been offered as explanations of demographic-economic growth in near-subsistence agricultural situations. Tlie invention-pull hypothesis asserts that the diffusion of new methods begins immediately after invention occurs , and lihat the diffusion makes possible additional population growth which then takes place. The popula- tion-push hypothesis asserts that a pool of unused agricultural knowledge is available at each point in time , but that each more -productive method requires more labor per worker. An increase in population is therefore necessary "co force the adoption of the more- productive methods, accordip'2; to the population-push hypothesis. This paper first demonstrates analytically that the invention-pull hypothesis refers onl\r to inventions that are labor-saving relative to tne u^athods in use . And the population-push hypothesis refers only to inven- tions that are output-increasing and labor-using relative to the methods in use. Once this distinction is made, the two hypotheses are seen to be complementary rather than mutually exclusive. The second part of the paper reviews a set of historical and c-nthropological cases of econom.ic-demographic change as they are relevant -31- to the two hypotheses. Some cases are found which are well described by the invention-pull hi'pothesis, while others are well described by the population-push hypothesis. That is, some important inventions — such as the potato — have been labor-saving, and therefore adoption began immediatal after invention. Others — such us slasi.-and-burn agriculture — have been mostly output-increasing but labor-using, and their adoption has awaited further population growth. This review confirms that both the invention- pull and population-push hypotheses have an important place in explaining economic-demographic history. Neither hypothesis, however, explains original inventions themselves, and neither applies very well to the develop- ment of modern business farming. References Asen Balikci, "The Netsilik Eskimos: Adixptive Processes", in Riciiard IJ. Lee and Irven DeVore (eds.), Man The TTunter (Chicago: Aldine, 1968). Ester Boserup, The Conditions of Agricultural (irrovth (Chicago: Ala A.V, Chayanov, The Theory of Peasant Economy (Komewood:Irwin, IV. Edited by Thorner et al . '«'■. Gordon Childo, Man Hakes Himself (London: Vatts , 1937). ourlo M. Cipolla, The Economic History of World Population (BaltxiriOi i. Pelican, 1962). Colin Clark and Margaret Hftsvell, The Economies _of Subsistence Ag^-j ? •' (New York: St. Martins, 2nd edition, ,1966). ""olin Clark, Population Growth and Land Use (New York: St, Martin:.. Colin v''lark, "Misconceptions on Hunger, Population", Chicapo Sun. Oct. 3, 1969, p. 28. «.nsley J, Coale and Ed^ar M. Hoover, Population Growth and Eco noir... Devolcpment in Low-Incomo Countries (Princeton: Princeton U^ 1958). K.H. Connell, "Land and Population in Ireland, 1780-1845", in ]).V, and D.E.C, Eversley, Popula-tion in History (Chicago: Aldina, 423-433. i. Cecil Curwen and Gudraund Hatt, Plough and Pasture (New York: Co.. 1953). •jincoln H. and Alice Taylor Day, Too Many Americans (New York: Hou^; Mifflin, 1964). , < 'I' Harold F. Dorn, "Mortality", in Philip M. Hauser and Otis Dudley Duncan (eds,)» Tlie Study of PoTMilation (Chicafijo: U. of C. Press, 1957). Telke Dovring, "Ths Transforaation of European Agriculture", Chap. VI in H. J. Habakkuk and U. Postan (eda), Th<9 Caabridge EconoKiic Hiatory of Europe , Vol. Vi, Part 2 (Cis«bridge» CUP, 1965), Err- Polke Dovring, review of Boserup, J. Econ, History , December, 1966, 380-381. Raymond i'irth, Primitive Polynesian Kconomy (London: fioutledge & Kegan, 1965). Pierre Gourou, Tl.e Tropical World, Its Social and Economic Conditions and Its Future Status , Translated by R.D, Laborde (New Y^ork: Wiley, 1966). Alfred L. Kroeber, Anthropology , new rev. ed. (New York; Harcourt Brace, 1948) Richard 3, Lee, "What Hunters Do For A Living, or, How To Make Out Pn Scarce Resources", in fl.B, Lee and Irven De Vore (eds.), Man The Hunter (Chicago: Aldine, 1968), pp. 30-48. William L. Lander, "Europe's Initial Population Explosion", American Historical K.eviev , 69, Oct. 1963, pp. 1-17. Reprinted in David M. Heer, Iteadintrs on Population (Englewood Cliffs: Prentice-Hall, 1968). Oscar Lewis, Life in a Mexican V"illa.q,e (Urbana: U. of Illinois Press, 1951). T.R. Mai thus, Principle of Population . (London: T. Densley, 1803). Jay R. Mandle, "The Decline in Mortality in British Guiana, 1911-1960", Demography . 7, Aug. 1970, 30L-316, Frederick D. ticCarthy and Margaret Mc Arthur, "The Food Quest and the Time }''actor iu Aboriginal Economic Life", in Charles P. Mountford (ed.), Records of the American-Australian Scientific JNXpedition to Arnhem Land, Vol. 2 ; (Melbourne: Melbourne U. Press, I960). Thomas McKeown and d.G, Brown, "Medieval Evidence Related to English Population Changes iu the Eighteenth Century", Population Stiidies , 9, 1955, pp. 119-41. Repriiited in David M. Heer (ed.), Iteadinx^s on Population (Englewood Cliff: Prentice-Hall, 1968) John U. Mellor, "The Use and Productivity of Farm Family Labor in Early Sta.^es of Agricultural Growtii", J. Farm Econ . 45, Aug., 1963, pp. 517-3 Peter Newraan, "Malaria Control and i\ipulation Growth", The Journal of Development Studios . Vol. 6, Jan. 1970, 133-158. Sigisraund Pelier, "Births and Deaths Araon;^ Europe's Ruling Families Since 1500", D.U. Glass and D.fi.L, Eversley (eds.), Population in History (Chicago: Aldine, 1965), pp» 87-100. Dwight Perkins, Agricultural Development in China. 1368-1968 (Chicago: Aldine, 1969). Robert Redfield, Tei^oztldn - a Mexican Villarre (Chicago: U. of Chicago Press, 1930). Uobert Redfield, A Village Tjtat Qiose Progress (Chicago: U. of Chicago Press, 1957) . Robert Redfield and Villa Rojas, Chan Kan; A Maya Village (Washington: Carnegie, 1934). Theodore W. Schultz, Transforminrr Traditional Agriculture (New Haven: Yale, 1964). Moritz Thomsen, Living Poor (New York: Ballantine , 1969). B.H, Slicher Van Hath, The Agrarian ilistory of V/estern Europe - 500 -185O (London: Edward Arnold, 1963). Johann H. Von ThUnen , The Isolated State (New York: Pergamon, 1966). Lynn White, Jr. , Medieval Technology and Social Change (New York: Oxford, 1952). James Woodhurn, "An Introduction to Hadza Ecology", in Richard B. Lee and Irven De Vore, Man the Hunter (Chicago: Aldine, 1968). '.• r ' . V > / I