Since the publication of Enrique Florescano’s study of Mexico City maize prices more than a decade ago, scholars have wrestled with the question of rising prices and their impact on economic growth and social stability.1 Through careful analysis of the records of Mexico City’s alhóndiga, Florescano documents a rise in maize prices by as much as 100 percent between the beginning of the eighteenth century and the end of the colonial period. He does not believe that this rise was the result of a long trend toward higher maize prices, but rather of inflationary pressures on maize prices from the 1780s into the 1800s. Before the 1780s, he points out, maize prices fluctuated frequently and widely, but the average price from cycle to cycle remained relatively constant. Under these conditions the price curve would be flat rather than sloped in one direction or another. Although he discounts the importance of long-term trends, he concedes that the eighteenth-century maize curve can be divided into three periods in which trends can be recognized. From 1721 to 1754 maize prices appeared to rise gradually. This was followed by a “tendency” toward lower prices from 1755 until 1778. Finally, from 1779 to 1814, “rising prices dominated once again.” Of greater significance to Florescano, however, is that until the last quarter of the eighteenth century prices were held in check mostly because sharp increases were counterbalanced by sharp decreases. In the last quarter, however, the typical oscillations were accompanied by a steady rise in prices.2 If, as some believe, wages and salaries remained virtually unchanged throughout the eighteenth century, then a late eighteenth-century upward swing in prices could lead to widespread price inflation and financial hardship.3 All these conditions combined to make the colony ripe for rebellion in 1810.4
Economic historians have long studied prices to learn more about the economy in general. “Price history,” writes Lennart Jörberg, “can become the gateway to further research in ‘general’ economic history.”5 In light of the importance of Florescano’s work and the research that it has spawned, it may be an appropriate time to reexamine the gateway with particular reference to questions relating to cycles and trends in New Spain. I have approached this reexamination in three steps. First, I have brought together price series for various cities and regions found in recently published studies in order to compare their specific characteristics and general patterns. These include mostly maize prices, but also prices for wheat, flour, and frijoles. With prices for several commodities and from several regions we can take a second step; to begin the task of constructing a colony-wide series for each of these commodities. At this stage the only colony-wide price series of significant length is a maize series, which I describe and analyze in this essay. The existence of price indices for maize and other commodities would provide scholars with another tool for analyzing inflationary or deflationary trends in the public and private sector of the colonial economy. A third step, which can be undertaken without such price indices, is to compare and evaluate growth rates in prices with those for silver registrations and agricultural tithes. The extent to which mineral or agricultural production ran ahead or lagged behind maize-price trends might offer a preliminary view of the growth or the lack of growth in the eighteenth-century Mexican economy.
To sharpen the focus on trends rather than cycles will give a different slant to the description and analysis of eighteenth-century prices. The secular trend for the eighteenth century shows that maize prices rose modestly but irregularly. By stressing the intermediate- and short-term trends within the secular trend, we can highlight periods of concentrated growth when prices climb at much faster rates than they do in the secular trend. Only rarely, though, do downward trends emerge, for after famines and other agricultural disturbances, prices normally plummeted to low levels from which they started to rise once again. In the decade after the 1786 famine, however, downward trends of from five to ten years’ duration have been tentatively identified in several series. When the upturn came in the late 1790s or early 1800s, it began at a higher level than was the case after earlier famines. Upward trends of the intermediate- or short-term variety are much more frequent and visible because the general tendency in the eighteenth century was for prices to rise. Since other demographic and economic indicators were on the rise in the eighteenth century, prices could be expected to follow suit. By comparing growth rates for prices with those for population, mining, or agriculture in the long, intermediate, or short term, we can draw a bead on the important question of how much real economic growth the colony posted in the eighteenth century. The prospects for such growth probably diminished in the second half of the eighteenth century because prices made a strong surge, not only in response to short-term events like famines and epidemics, but also to long-term demographic and economic changes.
Sources and Methods
All price historians search for data that cover more than a few years, are generally reliable, and include more than a single region. Historians working in colonial Mexico have neither the best nor the worst of worlds in this regard. We now have several long runs of maize prices, which, when merged or combined, provide us with a fairly extensive series from the 1520s through the independence period. In addition, there are numerous hacienda and urban studies in which prices of other commodities such as wheat, flour, vegetables, and meats are reported and analyzed. The commodity series are not yet as long or complete as the various maize series, but each represents another step in the creation of a data base for eighteenth-century prices. The studies from which I have drawn price data are well known to colonial scholars. They include the works of Woodrow Borah, Sherburne Cook, Charles Gibson, and Enrique Florescano for Mexico City and the Valley and David Brading, Flor de María Hurtado López, and Claude Morin for the Bajío and the Michoacán region.6
The principal tool that I have used to identify and verify trends is regression analysis (ordinary least-squares variety). This involves finding, if possible, the line that best fits and describes the changes of a dependent variable in terms of the changes of an independent variable. In the real world, of course, year-to-year changes in a variable like prices can be highly volatile and yet, despite that volatility, a trend can emerge. To fit a line is to highlight a trend that may not be immediately obvious even after the data are plotted on a graph. Although our ultimate goal is to try to explain which factors cause the data to behave as they do, our immediate task is to try to describe that behavior. Although we can probably never know the true relationship between prices and time or prices and any other variable, we can at least estimate what that relationship might be. Regression analysis is one way to estimate that relationship.7 But we are interested in finding more than the best fitting line. We also want to try to determine how much the trend may explain the behavior of the data over time. Not only can we show the difference between the estimated value (the trend line) and the observed value (the actual price), but we can also narrow that difference. To narrow the difference is to reduce the amount of unexplained variation between the estimated and the observed values. For example, when the secular trend is broken into segments according to the worst famines and the highest prices, r-squared values, which show how much of the trend can be explained, improve significantly, It is important, then, to pay attention to the impact that famines, especially the major ones, had on eighteenth-century prices. Other evaluations and tests of the relationships among the various data sets are possible, although they must be pursued with care because these data, like other historical statistics, are not always fully trustworthy.
To calculate the rate of growth along a trend line we have converted the data, shown on the vertical axis, to logarithms. In simplest terms, we have used logarithms to transform non-linear data, such as highly volatile maize prices, into a linear form from which we can then estimate the rate of change.8 It should be noted for readers unfamiliar with compound annual growth rates that a yearly increase of one percent in any variable over a century will mean that the variable grew 2.7 times. To attain the same growth in a shorter period, say a quarter of a century, would require rate of increase in the range of 4 to 5 percent each year.
Maize Price Trends
The colony-wide maize series that I have constructed contains about 750 individual prices from various towns and regions, and encompasses nearly the whole colonial period from 1525 to 1819. Before I turn to a description and analysis of eighteenth-century maize-price trends, however, I want to discuss possible trends for the whole colonial period. To identify these trends, I have used two different techniques. The first was to examine all the data (ca. 750 items) for trends and the second was to compute an average for each year (218 items) and to analyze the averages for trends. When all the individual prices and the annual averages are plotted (Graphs 1 and 2), they both follow a modestly upward course. Along the trend line the rate of growth is 0.5 percent per year when all the prices are included and 0.4 percent when just the annual averages are included. For all the data the average price is 9.5 reales per fanega of maize (hereinafter r/f), and for the annual average prices the average is 11. On the basis of these two approaches it is reasonable to conclude that during the whole colonial period the price of maize on average was 10 to 11 r/f and the price rose about 0.5 percent each year.
Graphs 1 and 2 also reveal that maize prices did not move along a straight, upward line. From the end of the first quarter of the sixteenth century to the middle of the seventeenth century, maize prices advanced rather energetically at a rate of 2.2 percent per year; but from the middle of the seventeenth century (1661) to the end of the colonial period, they climbed much more slowly, at a rate of o.5 percent annually. When plotted, the data for the first half of the colonial period follow a sharp, upward course, but the data for the second half of the colonial period exhibit only a slight, upward tilt. Eighteenth-century prices have a stronger and clearer upward tilt than do prices from the middle of the seventeenth century through the early eighteenth century and prices in the second half of the eighteenth century (to 1809) move up more aggressively than prices in the first half. (See Graph 3.) This suggests that maize supplies grew scarce in the first third (an approximate periodization) of the colonial period, became more plentiful in the second third (although shortages were frequent) and grew scarce once again in the final third.
In traditional, agricultural societies, like colonial Mexico, population and price trends can have a close relationship over the long term.9 Describing the individual trends is often much easier than trying to relate trends or to analyze their separate or common characteristics. In the case of Mexico, the trends for the second half of the colonial period make somewhat more sense than those for the first half One might assume that during a protracted period of depopulation, such as occurred in the first half of the colonial period, contraction in demand would cause prices to fall and eventually to stagnate. But, to judge from the available data, maize prices rose sharply in the sixteenth century and into the seventeenth century. To explain the rise. Cook and Borah suggest that new silver discoveries along with increasing wage rates in mining camps and urban areas pushed up prices because fewer rural laborers were available to work the farms and to produce the staples. By the middle of the seventeenth century, when depopulation and the boom in mining had abated, so, too, had the price spiral.10 The population began to recover in the middle seventeenth century, and the recovery began to accelerate in the early or middle eighteenth century. In their late work, Cook and Borah developed a series of growth rates to describe population changes in an area designated as “West-Central Mexico.”11 In this region the rate of depopulation ran as high as 5 to 6 percent per year in the middle of the sixteenth century, although it had dropped to zero by the middle of the seventeenth century. As the demographic recovery took hold in the second half of the seventeenth century, the population grew at just under 1 percent from 1650 to 1750 and at just over 2 percent during the second half of the eighteenth century. Growth rates for the whole colony cannot be generalized from these regional figures; but if the rates for Mexico were at all similar to those for West-Central Mexico, then its population growth could be estimated at between 1 and 2 percent from the middle of the seventeenth century to the middle of the eighteenth century and from 2 to 3 percent from the middle of the eighteenth century to the end of the colonial period.
From an economic standpoint, population growth, especially after 1750, led to more and more marginal land being brought under cultivation in order to accommodate the increase in demand. Such lands cost more to develop and to use than existing acreage; the higher costs meant higher prices for the products of these lands. Over time these higher prices would pull up all prices. In this regard, Eric Van Young has asserted that population growth could have directly stimulated late colonial technological change in the form of a “more intensive application of the available technology and recombination of the major factors of production— land, labor and capital,” but failed to do so. Eighteenth-century Mexican agriculture remained primarily an extensive system as the rise in production of food, fueled by population growth, depended heavily on addition of land, not intensification of production. Although improvements in agricultural techniques led to larger harvests and higher yields, they did not assure higher profits or lower prices.12 Often after 1750 food production played “catch up” with population growth. Other factors besides popúlation growth can affect supply and demand, but they are not easy to untangle and to document. For one thing, real wages probably declined as the population grew. With less purchasing power, to the extent that it existed in such an undeveloped economy, consumers might seek maize substitutes and, if they succeeded, their collective actions would reduce the demand for maize. Even under normal conditions, however, maize prices, like those of many staples, tended to be inelastic in that the demand for maize had an upper limit. Farmers had little incentive to increase their maize output inasmuch as an increase in output or in yield did not mean an increase in revenue, even if maize prices were lowered, because of a limit to the demand. Indeed, as some no doubt discovered, they had more to gain by reducing or manipulating maize supplies in order to keep the price high. Higher production costs (even with higher yields) combined with unattractive market conditions made maize an unappealing investment even though the price was rising.
Eighteenth-Century Maize-Price Trends
The eighteenth century represents the strength of the colony-wide maize series because it incorporates a half-dozen different series, some of which cover between a half century and a full century. Table 3 lists the annual average price by place and three colony-wide averages. The first colony-wide average (AV9) is an average based on prices from as many as nine different prices and the second (AV6) includes prices from the six different places—Valley of Mexico, Mexico City, Léon, Silao, Zacatecas, and Hidalgo—that have the longest and strongest series of the complement of nine. Let me underscore, as a glance at Table 3 quickly reveals, that nine or six different prices do not exist for every year in either colony-wide maize average. Finally, I have calculated a five-year-moving average, which may include for any single year an average from as many as nine different prices. Deciding where to start and to end the series for the eighteenth century requires, as usual, some arbitrary decisions. Maize prices fell between 1696 and 1700, although the price was slightly higher in 1700 than 1699. A price of nine r/f in 1700 was low, but not the lowest price that maize sold for during the eighteenth century. At the other end, I have generally used 1809 as the terminal point in order to minimize the influence of the revolt by Miguel Hidalgo and the insurgency movement that it spawned. Prices continued to rise after 1809 and probably peaked in 1811; but they were pushed up in part by war-related shortages and disruptions. The only obvious differences between the AV9 and the AV6 show up for the period from 1788 to 1809, when the addition of prices from Michoacán and Oaxaca yields a lower colony-wide maize average. The moving average, when plotted, has a different appearance from AV9 and AV6 because it smooths out the peaks and valleys, although following a definite upward course. Unless noted to the contrary, trend lines and growth rates are based on calculations from the AV9.
Annual average maize prices (AVg) between 1700 and 1809 from Table 3 are plotted on Graph 4. Although widely dispersed around the mean of the annual averages, these averages follow an upward course at a rate of 0.7 percent per year. Trends are less apparent and less verifiable in the first half of the eighteenth century than in the second half, in part because data for certain crucial years between 1700 and 1720 are missing. The colony-wide series includes only prices from León and Silao for 1702-7 and for 1715-18 and only prices from the Valley of Mexico and the capital for 1708-13. During the first half of the eighteenth century, maize prices in León and Silao were generally lower than Valley prices in years of normal or large harvests; but by the second half of the eighteenth century, the maize-price differentials were neither as great nor as frequent. Using a single set of local or regional prices may provide too narrow a base from which to compute a colony-wide average for a given period. Only in 1714 can we compute an average from prices for both regions. More data would probably not change the obvious—that prices were higher in 1714 than they were in 1700 or in 1720—but they would help to clarify intermediate- and short-term trends in the early eighteenth century. The probability (91 percent) of an upward trend (with an annual growth of 0.7 percent) in the first half of the eighteenth century is less than the probability of an upward trend (with a rate of 2.2 percent per year) in the second half Even when the 1786 price, the highest ever recorded, is dropped, the annual growth rate in the second half is still high (2.0 percent per year). This is also evident in the rise in the average price of maize from 10.9 r/f between 1700 and 1750 to 13.0 r/f from 1751 to 1800.
On Graph 4 we can observe that maize prices reached very high levels on three different occasions during the eighteenth century. In these three instances maize harvests for one or more years were almost totally lost, and surpluses, if they existed at all, were inadequate to meet the demand and to prevent a famine. By employing “dummy variables,” built around these three major famines, we can determine how much the famines influenced the trends. R-squared values, which indicate how much of the variation in the data can be explained by the trend, improve dramatically from 20.4 to 45.3 percent. At the height of each famine the colony-wide price of maize was higher than before: 1714 = 24.0 r/f, 1750 = 31.2 r/f and 1786 = 42.0 r/f Individual characteristics in all these famines are different, however. In the first famine, a series of bad harvests, perhaps three or four, caused prices both to oscillate and to rise through 1714. By contrast, the 1750 famine was confined more or less to a single year, although in the decade before this famine periodic minor shortages, accompanied by price hikes, gave an upward slant to the price curve. After these two famines, prices dropped sharply once the crisis eased. In the final case, 1786, when the price of 42.0 r/f shattered all other records, prices had been rising for two decades; and even after supplies were replenished, prices remained at higher levels than after earlier famines. All of this suggests that any adjustments—a fall in demand from depopulation or a growth in supply because of new, big harvests—were more far-reaching after famines in 1714 and 1750 than after the one in 1786.
Acute famines arose out of unexpected and extraordinary circumstances, like a summer frost or a spring drought. These famines may also represent the culmination of a process in which changes in supply and demand were being reflected in slowly and often erratically rising prices. In the first period (1700-1714) prices advanced at an annual rate of 8.3 percent, among the steepest for any period in the eighteenth century. Scanty though the evidence may be, it nonetheless suggests an irregular, but upward, swing between the early years of the first decade and the middle years of the next decade. As high as prices were from 1711 through 1714, they did not surpass the levels that prices reached in the middle 1690s or the early 1640s. For the second period, 1715-50, prices grew by 2.6 percent annually. This period is difficult to analyze because it had more of an undulating pattern than the others. In the initial years, 1715-17, prices were extremely low, at two to four r/f; and then for the next two decades they swung between lows of six to eight r/f and highs of 13 to 15 r/f The swings did not necessarily negate the trends. From the early 1730s through the early 1740s a moderate upward trend can be observed, and then, after a brief interlude, in the middle 1740s that trend reemerges. Even when regression analysis is applied to the moving average for the periods 1716-49 and 1716-52, it yields yearly growth rates of 1.3 percent and 1.4 percent, respectively. It is probable that some major famines, perhaps like the 1750 famine, had more of a tug on maize-price trends than others. In this case the price of maize in 1750 was twice as high as in any year since the last major famine. Still, prices had a tendency over the short or intermediate term—for example, from the early 1730s through the early 1740s—to edge higher rather than to decline or to drift. A large harvest might pull them back temporarily, but under the pressure of underlying demographic changes, in combination with weather-related disturbances, prices began to rise once again.
In the third period (1751-86), even though maize prices rose at a comparable annual rate of 2.5 percent, the trend is clearer and stronger than in the previous period. After the crisis of 1750, maize prices fell to their lowest levels in nearly thirty years. By the late 1750s, prices had almost doubled; then in the early 1760s they retreated again. From the mid-1760s to the mid-1780s, prices climbed steadfastly higher. The advance was not without occasional relief, as, for example, in the late 1760s or late 1770s, when harvests were large enough to cause prices to drop. From 1763 through 1786, maize rose from an average 5.8 r/f to 42.0 r/f for an annual growth of 3.8 percent. Even if the periodization is changed from 1763-86 to 1763-88 or 1763-91 in order to moderate the influence of the shortages of the 1780s on prices, the annual growth rates remain high, at 3 to 4 percent each year. These two decades, in which prices rose so steadily, must have helped to lay the groundwork for the devastating famine of 1786 as well as for an agricultural crisis that would linger for the next two decades. Population growth, which translated into increasing demand at a time of possibly shrinking supplies, was beginning to have a direct impact on the agricultural system and in the market place.
Steadily and rapidly rising maize prices in the two decades between the crises of the 1780s and the insurgency movements would certainly offer support to those interpretations that stress the connection between economic instability and social unrest. But maize prices, it appears, did not follow that course exactly, for they did not resume their ascent again until the late 1790s or early 1800s.13 What may have happened after the 1780 famines is that maize prices dropped, but not as far or as fast as they had after other major famines. For example, the drop in price from 1750 to 1751 and 1750 to 1752 was 55.1 percent and 87.2 percent, respectively. By contrast, the decline from 1786 to 1787 and from 1786 to 1788 was 41.5 and 62.2 percent, respectively. As a result, maize prices were lower than they had been during the crises of the 1780s; but they were also higher than they had been after such crises in earlier periods. Famines, tragic though they were in human terms, acted as a temporary economic corrective in that the higher the death toll, the greater the release of pressure on food supplies and prices. If high enough, the death toll might allow for a period in which the supply of food could accommodate the level of population. The period immediately following the 1780 famines is different in one crucial respect from other post-famine periods. Even though prices were dropping, they stopped falling in a moderately high range, after which they started climbing again. This raises two interesting points. First, the 1786 famine, although severe, may not have had a high enough death toll, given the past population growth, to bring about a slackening in demand to the extent necessary to drive prices down farther; second, the postfamine harvests may have been too small to create the surpluses that in the past had caused prices to plummet before demographic and economic pressures combined to set them on a new, upward course. The behavior of maize prices in the quarter century before the Hidalgo rebellion is more complicated than previously assumed. Instead of suffering a precipitous drop, prices moved downward in a zig-zag fashion for the half-dozen years from the late 1780s into the 1790s. When they turned up again, they did so from a level that was high compared to earlier post-famine maize prices. There is no doubt that maize was costly and was becoming more costly in the ten to fifteen years before the independence movement broke out. What is now more evident than before is that the rise in prices in the late third and early fourth quarters pushed the price of maize to a new and higher plateau from which the next upward trend was to be launched.14
We can broaden the perspective of colony-wide maize-price trends if we relate them to movements in silver registrations and tithe collections. Silver registrations may be used to measure the output of Mexico’s silver mines. From 1700 until 1810, output grew at a compound rate of 1.4 percent annually, while maize prices rose by less than 1 percent per year. Silver production rose just slightly faster in the first half of the eighteenth century, at 1.7 percent per year, than in the second half, at 1.3 percent. Although small, the differences in rates suggest that the growth in output of silver slowed down in the half century before the Hidalgo rebellion, in part, we assume, because the jump in maize-price rates points to higher costs and lower profits in the mining business.15 Tithes provide a rough measure of the colony’s agricultural output. Landowners paid the tithes in kind, on the basis of assessments against their operations, to special collectors who sold the produce for cash. Brading spells out for León and Silao how the tithes were paid: 75 to 80 percent in maize, 12 percent in wheat, 7 percent (and sometimes higher) in animals, and the remainder in other crops.16 Although many technical problems are associated with the preparation of tithe series, they can often be minimized or solved. Two such series now exist for Michoacán and Oaxaca; when they are combined, they represent a first step in developing a colony-wide series for agricultural output. In the period 1701—1800, the combined tithes rose by 1.3 percent per year. In the first half of the eighteenth century, the combined tithes increased at an annual rate of 1.6 percent and in the second half at 1.4 percent. Over the long term, maize prices climbed much more slowly, at a rate of 0.7 percent per year from 1700 to 1809 and at a rate of 0.6 percent per year from 1701 to 1800 (the latter dates correspond to the period for the Oaxacan tithe series and the combined series). Thus, one might argue that because of more slowly rising prices (maize representing all commodities), Mexico’s late colonial economy enjoyed continued real economic growth. When, however, the maize-price trend is divided at 1750, a different picture emerges. In the first half of the eighteenth century, maize prices rose at less than 1 percent per year, but in the second half (to either 1800 or 1809) they rose at more than 2 percent per year. This points to a pattern in which growth slowed, and then vanished, as costs, reflected in rising prices of some products, mounted. Changes in growth rates for prices correspond to estimates of changes in population growth, that is, a rise of about 1 percent in the first half to between 2 and 3 percent in the second half of the eighteenth century. To expand production, however, to accommodate the rise in demand may have been complicated by the scarcity of land for cultivation, high costs for land improvements, the shortage of capital for investment, and competing economic obligations or opportunities.
Local Maize-Price Trends
Analysis of local maize series, from which the colony-wide series have been constructed, may shed additional light on the eighteenth-century maize economy. The local series include those for the Valley of Mexico, Mexico City, León and Silao, and Hildago. I have also constructed two regional series based on data from the Mexico City area and the Upper Bajío. Wherever possible, I will compare trends and growth rates for individual towns and cities and for the two regions.
The maize-price series for the capital is better than the series for the Valley of Mexico, although the latter, which runs for nearly three centuries, is considerably longer. Before 1660 the colony-wide series is based solely on maize prices collected by Cook, Borah, and Gibson from the Valley and nearby areas. The weakest link in the Valley series is from 1645 to 1750 because of long gaps without any prices.17 From 1751 through 1809, however, with far fewer gaps. Valley prices increased at a rate of 1.6 percent annually. In the capital the granary series starts in 1708, although it has a long gap from 1712 through 1720. The granary price rose at an annual rate of 0.5 percent between 1708 and 1809 and at an annual rate of 0.6 percent between 1721 and 1809. (See Graph 5.) For the period 1721-50 the granary price increased at an annual rate of 1.5 percent. Oscillation characterizes granary prices in the 1720s and 1730s, until the 1737 epidemic, when they began to climb. By 1742 they had reached their highest level since 1710. Granary prices dropped sharply in 1743, remained low in the middle 1740s, rose steadily in the late 1740s, and set a new high in the 1750 famine. Alhóndiga prices, in contrast to Valley prices, increased more slowly, at a rate of 1.4 percent yearly from 1751 through 1809.18 Although granary prices in the post-1750 famine declined to levels comparable to earlier recoveries from maize shortages, the long-term trend was for the price of maize to rise. Between 1751 and the next major famine in 1786, prices rose 1.8 percent yearly. As usual, the rise was erratic and unpredictable, as much for the producer as for the consumer. Prices rose modestly in the late 1750s and early 1760s and then fell rather sharply to their lowest point in the eighteenth century. From the middle 1760s through the middle 1780s, they followed an irregular, but upward, path at a rate of 3.7 percent per year. Several possible trends rather than a single trend emerge for the period from the 1786 famine through the 1810 revolt. From a high of 40 r/f in 1786, granary prices fell into the teens, where they bobbed around for nearly a decade. By eighteenth-century standards these postfamine prices were high, with a range of 10 to 20 r/f for the next decade or so. Only in the middle (around 1804) of the first decade of the nineteenth century do they appear to resume their upward march, largely in response to small harvests because of the eruption of a minor epidemic in the Valley. As a result, just before Hidalgo’s revolt, maize was selling at a price several times higher than it had fifty or a hundred years before. Escalation was as much a characteristic of Mexico City’s maize prices as was oscillation in the eighteenth century, certainly from the middle of the century until Hidalgo’s revolt.
Mexico City granary prices exhibit two interesting characteristics, which can probably be attributed to the fact that they were regulated. First, over the course of the eighteenth century they rose at a slower rate than unregulated prices in other places, and, second, they showed somewhat less volatility. The two characteristics were related. Alhóndiga prices were, in effect, average prices of market transactions. They did not fall as low or rise as high as the alternating conditions of abundance and scarcity might dictate for unregulated prices. Thus, despite the drawbacks, regulation imparted a degree of stability to maize prices.19
The strong upward trend from the mid-1760s through the mid-1780s may reflect a growing imbalance between the demand for maize and its supply in Mexico City and the environs. In the 1760s maize was plentiful and cheap in and near Chalco. Twenty years later, because of population growth in the capital and nearby towns, demand for and the price of maize had risen appreciably. Other influences were at work in the eighteenth-century maize economy. Maize farmers faced a variety of problems—large surpluses, depressed prices, increasing costs, low profits—which they tried, not always successfully, to solve. Further, good weather and productivity gains made their problems worse. Large surpluses and low prices caused farmers to reduce their plantings in maize in order to minimize their losses. In an area where tillable land was limited, the acres transferred from maize to other crops or activities might never be restored to maize, even when the market favored such restoration. Although the evidence is marginal and incomplete, it suggests that maize hacendados consolidated their holdings, quit business altogether, or rented arable lands for other purposes. In Chalco, the region that sold Mexico City most of its maize, the haciendas specializing in maize went from 57 to 1741 to 36 in 1773. Those 57 haciendas produced about 2,000 fanegas per hacienda, for a total of just under 114,000 fanegas, while the 36 haciendas thirty years later produced about 2,500 fanegas each, for a total of 90,000 fanegas (plus another 30,000 fanegas from rented hacienda lands). In the middle of the century, with 46 haciendas, the average output was more than 5,000 fanegas each, or a total of 250,000 fanegas. This amount was harvested for sale at a time when Mexico City needed about 160,000 fanegas. The excess in supply led to lower prices in the short run, but a loss of maize-producing haciendas led to higher prices in the long run. By the end of the century, Mexico City was consuming more than 200,000 fanegas, an amount exceeding Chalco’s output. The rise in demand, along with higher costs to cover maize purchases from other regions, helped to boost prices. At the same time maize producers may have deliberately sought to restrict the supply (by planting less and by storing more) even though the market might have justified an increase in maize plantings. The fear, based on recent past experience, was that to expand the production of maize would lower its price, but would not raise the producer's income. Moreover, since many producers were deeply in debt as a result of land acquisition and improvement during the previous several decades, they used high maize prices to generate income to maintain their operations and to pay their debts. The full dynamics of these ever-changing economic conditions remain to be documented but an outline of how they were unfolding is beginning to emerge more clearly now.20
Even Florescano’s calculations, which are designed to accentuate cycles, show that granary prices nearly doubled between the middle 1760s and the middle 1770s and then more than tripled between the late 1770s and the middle 1780s, with but a brief respite of two years of low prices. The average price of between 9.0 and 9.5 r/f in the late 1770s compares favorably with the average price of between 8.5 and 9.5 in the middle 1760s but the length of time—twelve to fifteen months of prices below 10 r/f in the 1770s and thirty-six to forty-eight months of similar prices in the 1760s—dilutes the value of the comparison.21 A “light harvest” in 1784 followed by almost no harvest in 1785 caused granary and market prices to skyrocket. “A state of near starvation” is how conditions in and around the capital during the height of the famine have been described.22 Maize prices in the capital and around the Valley doubled between 1784 and 1785 and doubled again between 1785 and 1786. In 1786 the annual average price at the granary reached an all-time high of 40.1 r/f, although during the summer of 1786 monthly average prices were as high as 46-48 r/f The high prices underscore how scarce maize was, but changes in prices say nothing at all directly about how much the population suffered. If the price of maize can be used as an index of misery, comparatively speaking, then the Valley of Mexico, including the capital, did not suffer as much as the Upper Bajío (including Zacatecas), where the 1786 average was 44.0 r/f That the price of maize was higher because the famine was worse in the Upper Bajío would indicate how the intensity of colony-wide famine varied from region to region.23 Was the famine of 1786, to use the language of Van Young, a “Malthusian demographic disaster?”24 Although Van Young is primarily concerned with Guadalajara and cites evidence for the lack of such a disaster from León, he raises a question that applies to the whole colony. What constitutes a Malthusian disaster is a matter of interpretation and a source of disagreement; putting aside those arguments, we can observe that since postfamine prices remain high by comparison with earlier postfamine price levels, this may be an indication of an acute famine with a moderate death toll. In Mexico City maize prices ranged from lo to 30 r/f between 1787 and 1810 with an average of 19.5 r/f The release of pressure on the food supply from depopulation was brief The Valley and Mexico City had to endure high prices and minor shortages until the end of the colonial period but without experiencing, it would appear, the relief that a Malthusian catastrophe would have imposed.
Analysis of maize prices in the Upper Bajío offers both confirmation of and contrast to the findings reported for the Valley of Mexico. The longest series of more than a century are for León and Silao, followed by series of a half century for Dolores Hidalgo. The other components are too limited to be analyzed individually but can be used to expand the data base into a regional series for the Bajío. Maize prices in the Upper Bajío, like those in the Valley of Mexico, will rise over the long term, although at different—but often faster—rates. (See Graphs 6, 7.) In general, demographic and agricultural changes similar to those just discussed in the Valley of Mexico played a role in the Upper Bajío. Unlike the Mexico City granary series, the León and Silao series is constructed from tithe transactions.25 The León series runs from 1661 to 1768 and the Silao series from 1661 to 1789; when they are combined, they cover the period 1661-1789. No prices exist at all for perhaps a third of the 130 years. To overcome the gaps in the individual series, I have used the combined series from 1700 through 1789, in which approximately a dozen years are missing. During this period the average price for the combined maize series is 8.8 r/f. When Brading analyzed these series, he concluded that “maize prices in León and Silao followed no fixed pattern.” This is partially correct in that trends before the eighteenth century are hard to verify.26 Several trends can be identified and analyzed in the eighteenth century. For example, from 1715 through 1789, prices rose at a constant rate of 1.1 percent each year. Where comparisons can be made between León-Silao and Mexico City, the former's prices generally rose more rapidly than the latter’s. From 1715 to 1750 the León-Silao series rose by 3.2 percent per year, whereas from 1721 to 1750 the Mexico City series rose only by 1.5 percent per year. Similarly, from 1751 to 1789 maize prices in León and Silao climbed more swiftly, at 2.8 percent per year, than Mexico City’s prices (at 1.8 percent per year) from 1751 to 1786. Even from the mid-176os through the mid-1780s, León-Silao prices rose at a faster annual rate (4.0 percent vs. 3.7 percent) than Mexico City prices. The steeper trends in León and Silao resulted from prices being lower in periods of abundance and higher in periods of scarcity.
Useful price data have also been collected from the tithe records of Dolores Hidalgo, not far from León and Silao, by Flor de María Hurtado López for a period of almost a half century.27 A trend line with an annual growth rate of 1.7 percent (98 percent probability) can be fit to the data for the whole period, 1744-89. When we analyze the data for the period between the 1750 famine and the famines of the 1780s, we find a clear upward trend, at a rate of 3.5 percent per year. And from 1763 through 1789, the trend line has a rate of 6.0 percent each year. Both are higher than the León-Silao rates for similar periods.
A larger and richer picture of maize-price trends in the Bajío can be drawn by combining León-Silao data with those from Dolores Hidalgo and several towns in Michoacán.28 The average price in this combined Bajío series from 1744 through 1799 was 10.3 r/f In the brief period before the 1750 shortage, prices for the combined Bajío series rose rapidly, at a rate of 28.1 percent per year. As an indication of how rapidly prices could rise over a six-year period, the average price rose from 5.0 r/f in 1744 to 12.5 r/f in 1747, fell back to 6.6 r/f in 1748, and then jumped to 14.0 r/f in 1749 and to 32.9 r/f in 1750. Prices remained high (15.8 r/f) through 1751, before abundant harvests pushed them down below 10 r/f, where they remained until the early 1760s. From lows in the range of three to six r/f in the middle 1750s to highs above 40 r/f in the middle 1780s, the Bajío series rose at a rate of 3.0 percent per year. From the mid-1760s to the mid-1780s, prices climbed at 5.5 percent annually. The rise was irregular but nonetheless persistent. Bajío residents were paying more for maize in the 1770s than they had paid in the 1760s and more in the 1780s than in the 1770s. After the 1786 famine, Bajío prices may have turned downward at an annual rate of between 6 and 7 percent.29 Lacking Bajío prices after 1800, we cannot say what course they took in the decade before Hidalgo’s revolt.
Demographic pressures—population growth added to urban migration—probably accounted for as much of the maize-price rise in the Bajío as elsewhere. In contrast to the Valley of Mexico, however, where maize production may not have grown as fast as demand or may not have grown at all, maize cultivation was being expanded. Brading reports that the initial emphasis on stockraising changed to a system of mixed farming, which in turn was replaced, where feasible, by a concentration on the cultivation of cereals."30 In León the maize crop, based on tithe receipts multiplied by ten, increased from an annual average of about 27,000 fanegas in 1661-65 to 70,000 fanegas in the late 1760s. In Silao the increase was greater: from an annual average of near 12,000 fanegas in the 1670s to 70,000 fanegas in the 1770s to 130,000 fanegas in the 1780s. The maize production curve did not move along a straight, upward line, for tithe figures indicate maize output could vary from decade to decade. In the long term, however, the trend underscores a growth in output. Wheat production was increasing, too; it grew threefold in León from the 1660s to the 1760s and twentyfold in Silao from the 1670s to the 1780s.31 In short, there appears to have been an explosion in Bajío crop farming during the eighteenth century. While total production of maize was growing, however, it may not have been the region’s most profitable crop. By the nineteenth century, according to Brading, maize was still the region’s chief crop, hut the return on capital invested in maize was only marginal.32 At times in the mid-1750s and the mid-1760s, presumably because of the expanding output, Bajío maize was extremely cheap, even cheaper than maize in the Valley of Mexico and the capital, where surpluses had also depressed prices. With harvests so large and prices so low, how Bajío farmers maintained profitable operations or avoided bankruptcy is not fully explained. Even if periods of low prices caused some bankruptcies and mergers among hacendados, as well as shifts to different crops, the maize crop, it would appear, continued to grow. Many different factors may account for the continuing expansion of the maize economy but the combination of two of them, a rising population and a surging economy, may have predominated. One indicator of a booming economy was that silver registrations at Guanajuato, which was probably the region’s fastest growing city, increased by 3.5 percent per year between 1763 and 1786. In Zacatecas, to the north of Guanajuato and outside the Bajío, silver output rose at a rate of between 6 and 7 percent a year during the same period. Since overall Mexican registrations grew between 3 and 4 percent per year in the late third and early fourth quarters, much of that growth was concentrated along a line from the Guanajuato district into the Zacatecas district.33 The revival of these mining camps had an important demographic impact in that people from the countryside of the Bajío as well as nearby regions migrated to these camps and to other cities in search of employment. The sharp rise in Bajío maize prices between the mid-1760s and the mid-1780s, which culminated in an extraordinarily acute famine lasting from 1784 to 1786, may be a reflection of the unusually heavy demand for maize resulting from urban growth, especially in the mining camps. By the 1780s maize prices, which, except for famine years, had been lower in the Bajío than in the Valley of Mexico, were comparable to those at Mexico City’s public granary.
Late Eighteenth-Century Wheat and Bean Prices
Colonial scholars have assumed that the prices of many commodities like foods and grains followed trends and cycles similar to those for maize prices. It is difficult to test this hypothesis because so little research has been done on other commodities. From the few studies we have, we can compare the movement of maize, wheat, and bean prices in the second half of the eighteenth century. In general, commodity prices rose from 1750 to 1800 or to 1810. Some commodity prices rose slowly while others climbed rapidly, even more so than maize.34 Commodity prices, especially for staples, could be volatile and erratic because, like maize, they could be directly affected by climatological and epidemiological events over which the colony had little control. In the long term, however, what pushed up commodity prices along with maize prices was growth in population and in demand.
One commodity, wheat or wheat flour, bears close scrutiny. The demand for wheat and flour grew along with the urban population and probably more strongly in central and northern Mexico than in the south. Large-scale wheat production made its appearance in the Valley of Mexico at the end of the sixteenth century and was predominantly associated with Spanish rather than Indian agriculture.35 Brading found that wheat production remained fairly steady between the 166os and the 1750s, after which it increased markedly in León and Silao.36 Similarly, Morin found more wheat being produced in the second half of the eighteenth century than in the first half.37 An extensive treatment of wheat-flour production in the Guadalajara market area appears in Van Young’s study. The pattern was similar to what Gibson and Brading found, although described in fuller detail. Because of a growing population and prosperity in late colonial Guadalajara, landowners were encouraged to make investments for irrigation works and storage facilities in order to increase their production of wheat and flour specifically for sale in the city.38 Shipments of wheat and flour to Guadalajara rose from approximately 2,000 cargas around 1750 to nearly 15,000 cargas in 1809, almost an eightfold increase. Spaniards and Europeans had always consumed large quantities of white bread but so, too, according to Van Young, did the eighteenth-century urban poor—the Indians and the castas.39 Van Young does not present a series of wheat-flour prices, although he maintains that the city’s bakers bought wheat for prices in the range of 5 to 20 pesos per carga (hereinafter p/c). The supply of wheat, writes Van Young, was “more elastic” than maize; and wheat farming, which was potentially more profitable, was displacing maize farming. Although the demand for wheat in cities like Guadalajara was apparently growing faster than that for maize—this is hard to prove conclusively—the price of wheat relative to the price of maize, according to Van Young, was not.40 In Hidalgo maize prices did rise more rapidly than wheat prices, 1.7 percent per year from 1744 to 1790 compared to 0.4 percent per year, although there is only a 51 percent probability that a trend existed for wheat prices. The reverse was true, however, from the early 1770s through the mid-1780 famines when wheat prices rose 12.1 percent annually; but maize prices increased by 8.2 percent per year (98 percent probability). A detailed picture of how wheat displaced maize remains to be completed.41
Hurtado López has also culled wheat prices from Hidalgo’s tithe records. His price and production figures indicate that wheat supplies were in excess of demand until the late third quarter of the eighteenth century. It appears, although it cannot be confirmed, that wheat prices were in decline from 1742 through 1772, after which they leaped ahead at a rate of 12.1 percent per year through 1789.42 A straight-line trend cannot be fit to Hidalgo’s wheat series. In general, Hidalgo’s wheat prices have a highly fluctuating character, more so, perhaps, than maize prices. Prices range from eight r/f to 116 r/f or from two p/c to 30 p/c. Still, the findings are instructive because they suggest that if farmers, at least those around Hidalgo, shifted from maize to wheat to realize higher profits, they made little headway until the fourth quarter of the century. There the slump in maize prices during the early third quarter seemed to be less severe than the slump in wheat prices. Wheat production probably had been growing since the late seventeenth century, but, as with maize and other cereals, wheat surpluses could periodically depress prices and erode profits. This would help to explain why prices tended to drift during the early third quarter. Some time in the late 1760s or early 1770s the price of wheat took off. This rise in price remained strong until the late 1780s. This upward trend may help to explain why farmers and landowners were expanding wheat cultivation, although it does not shed any further light on the exact relationship between wheat and maize cultivation, the acres allocated to each and the profits earned from each. Wheat may have been a sound investment for the large hacendados, who could afford the necessary capital improvements, but not for the small-scale operations of tenants, renters, or farmers. Unfortunately, since the Hidalgo wheat series ends in 1790, we cannot say what happened to wheat prices and profits after the famine and depopulation of the 1780s. Wheat may have undergone a more severe contraction in demand than maize after the crises of the 1780s, with the result that wheat prices were lower than they had been in two decades.
A short series of flour prices, based on granary sales in Zacatecas from 1788 to 1819, can be used to analyze their trend between 1790 and 1800. Flour was sold there both on the open market and through the public granary, although not at fixed price. The Zacatecas series has prices for only nine years; but seven of the nine fall between 1788 and 1799. What happened to flour prices in Zacatecas and presumably wheat prices in the nearby wheat-producing areas was that they had stopped rising and had begun dropping in the early 1790s. By the late 1790s, however, in part because of a new shortage, wheat prices shot up again. During the years for which sales exist, prices ranged from a low of 16 r/f or four p/c to a high of 100 r/f or 25 p/c. The average flour price in 1788-99 was 46.7 r/f or 11.6 p/c. Normally, the price of flour ought to be higher than the price of wheat in order to account for the expense of milling the wheat and transporting the flour. In Hidalgo from 1773 through 1789, when prices rose so rapidly, the average wheat price was 54.8 r/f or 13.7 p/c. Flour prices in the range of 9 to 15 p/c in Zacatecas, which imported the product from the Upper Bajío during the 1790s, would probably indicate Hidalgo wheat prices of less than 40 r/f or 10 p/c.43 The extent to which (or the rate at which) wheat prices fell in the Upper Bajío or in the northern agricultural regions cannot yet be measured. In Michoacán, according to Morin, wheat prices ranged from 10 to 20 r/f or 2.5 to 5 p/c during the 1790s.44 If representative of prevailing wheat prices, they point to a sharp decline from prices before the 1786 famine. The impact of such a decline is not yet clearly understood. Were wheat producers so lacking cash to pay their bills that they had to increase their borrowings or sell their operations? Were they less inclined to expand wheat production in particular and grain production in general to bring supply into balance with demand?
If wheat prices were in decline during the late 1780s and early 1790s, then this should be reflected in flour supplies. The flour-supply figures existing for two cities—Zacatecas and Guadalajara—are not easy to unravel and explain. Since the volume of flour entering both cities in the 1790s did not grow, the demand for flour was probably not growing either. In both Zacatecas and Guadalajara migration was the chief stimulus to growth in population and while the heavy influx strained the food supply during the 1786 famine, the situation had improved by the 1790s. The volume of flour entering Zacatecas between 1787 and 1810 ranged from 6,000 to 10,000 cargas per year, for an average of 7,500 cargas per year. In Guadalajara, on the other hand, the range for a similar period was from 4,500 to 17,000 cargas per year, for an average of 10,000 cargas per year. Shipments to Guadalajara rose faster and higher in the first decade of the nineteenth century than shipments to Zacatecas. For the 1790s (actually 1788–1801 with 11 years of data in Guadalajara and 12 years in Zacatecas) the averages were remarkably close: 7,000 cargas per year in Guadalajara versus 7,800 cargas per year in Zacatecas. Beginning in the 1800s supplies rose first in Guadalajara and later in Zacatecas in response, presumably, to rising demand from urban growth. In both cities, as the population began to climb, largely as a result of migration, so, too, did wheat shipments.45 In the case of maize, prices dropped after the 1786 famine (but not as far as they had after earlier famines). In the case of wheat or flour, the evidence is more ambiguous. Immediately after the 1786 famine, Hidalgo wheat prices tumbled from 116 r/f or 29 p/c (in 1786-87) to 48 r/f or 12 p/c (in 1788-89). In Zacatecas the price of flour went down from 48 r/f or 12 p/c in 1788 to 36 r/f or 9 p/c in 1792. Zacatecas’s flour prices during the 1790s had a range which suggested that wheat prices may have fallen as low as they were before the steady rise in the early fourth quarter.
Beans can be examined only briefly. The extent to which beans were used in the diets of Spaniards, Indians, or castas is not well documented, although, according to Van Young, they were a part of the diet “of urban as well as rural dwellers.”46 Both Gibson and Van Young state that the crisis of the 1780s pushed up bean prices as fast as maize prices.47 In late colonial Zacatecas plans were discussed, but never enacted, to place the sale of beans, like maize, under the control of the public granary in order to ensure a steady supply at a reasonable price.48 For three different localities—Hidalgo, Zacatecas, and Oaxaca—we possess published bean-price series. The longest series (1741-89) is for Hidalgo but it does not yield a valid or reliable trend. Of the seven cycles identified by Hurtado López, successive declines in the average price during the first four cycles were followed by increases during the final three.49 Average annual prices in Hidalgo were 14.3 r/f from 1741 to 1790 and 12.8 r/f from 1751 to 1790, with a range of 3 or 4 to 42 r/f A trend line can be fitted to the Hidalgo data from 1752, the end of a decade of fairly high prices, through the 1780s and the agricultural crises. In that period bean prices rose at a rate of 3.0 percent per year. For a comparable period average maize prices in the Hidalgo-León-Silao region rose 2.9 percent per year. These figures indicate, as some have argued, that these two products shared common characteristics in production cycles and price trends. Data are unavailable for Hidalgo after 1790. In Zacatecas, however, bean prices, although highly variable, declined from 1788 to 1808, when a new crisis pushed prices up again. Zacatecan consumers paid an annual average price of 16.1 r/f from 1788 to 1808 with a range of eight to 20 r/f Average prices could have been higher before 1788 and lower after 1807, although the data are much too scattered to support this statement with certainty. In Oaxaca after 1790 bean prices were two to five times higher than they were in Zacatecas. There is no doubt that beans did occasionally fetch such high prices; but it is unlikely that they did so over the long term. Still, beans rose in price during the second half of the eighteenth century at rates not unlike those calculated for maize, in part because of increases in urban populations and in part because of changes in urban diets.
Prices and Agricultural Production
Prices almost always fell when harvests were good and rose when they were bad. What is unknown is how agriculturalists reacted to these events and to what extent they could anticipate them. Surely, over the long term, they could see that they should adjust their harvests to changes in prices; but, in the short term, they did not always have the information or the skill to make such adjustments quickly enough. Thus, agriculturalists, especially the large, commercial operators, had to make some educated guesses about production, price, and profit from year to year. The extent to which they understood the slow but erratic long-term upward movement in maize prices was probably less than the extent to which they not only comprehended but dealt with the short-term trends. Those who have studied the late colonial hacienda system have often argued that hacendados became increasingly adept at exploiting the system to their advantage. Even so, hacendados faced a bewildering array of options and decisions about what to plant, how much to plant, how much to invest in or borrow for capital improvement and how best to protect their profits in a business that was so unpredictable and unstable. To study production trends in conjunction with price trends may offer a few answers to these questions.
The longest tithe series (Graph 8) has been published by Claude Morin for the Diocese of Michoacán, an area that included the modern-day states of San Luis Potosí in the north, Guanajuato in the center, and Michoacán in the south, and that measured 600 kilometers from top to bottom.50 What Morin’s data show is that from the 1680s to the 1800s tithe revenue grew six- to sevenfold. An increase in the total tithe without a corresponding increase in the rate of the tithe could result from some combination of rigorous collection, and demographic or economic growth. No doubt, more efficient administration contributed to larger tithe revenues; but it can only account for a part of the increase. Growth in the economy as well as in the population helped to generate a major part of the revenue increase. When Morin’s data are plotted and analyzed, they show a rate of growth of 1.4 percent per year from 1680 through 1810 with an average for the period of 213, 153 pesos. In the same period colony-wide maize prices rose at about one-third that rate. When the Michoacán tithe series is restricted to the periods 1700 through 1809 or 1715 through 1809, growth rates rise slightly from 1.4 percent to 1.5 percent per year. Compared to maize-price averages for the same periods, the tithe was growing two to three times faster. If we compare Michoacán tithe collections and colonial maize prices for two of the three previously discussed periods, 1715-50 and 1751-86, we find that tithes and prices rose by the same 2.6 percent per year in the first period, and tithes by 1.4 percent and prices by 2.5 percent per year in the second. In the third period, 1787-1809, tithes grew by 1.6 percent per year, but prices first fell and then rose without establishing a clear trend. This raises an important but unresolved matter. At this point, colony-wide, maize-price trends, which have the best chance of being validated, suggest that prices decreased by 10.9 percent per year (98 percent probability) from 1787 to 1792 and then increased by 2.2 percent per year (97 percent probability) from 1793 to 1809. If a trend line for the whole period, 1787-1809, were plotted, it would show a growth rate of only 0.8 percent per year (76 percent probability). Given these growth rates, one might conclude that the agricultural sector not only had recovered from the famines of the 1780s but had also improved upon its performance of the third quarter. When, however, the period is divided into two segments, 1787-92 and 1793-1810, there is no reliable trend in the first; but there is a trend in the second with production falling slightly behind prices: 2,0 percent growth per year (99 percent probability) against 2.2 percent per year (97 percent probability), respectively. If we choose an arbitrary demarcation of 1750 for the eighteenth century for both tithes and prices, we find that tithes rose by 2.0 and 1.6 percent per year and prices by 0.7 and 2.2 percent per year, respectively. What we can underscore now is that for part of the second half of the eighteenth century prices began to climb because the output of food began to lag behind demand.
It may be more appropriate to compare Michoacán tithe collections with Bajío maize prices even though the maize series is not as complete as the tithe series. The best period for comparison is 1751 through 1786, a period for which we have maize prices from León, Silao, Hidalgo, and other towns. Prices rose 3.0 percent per year, but tithes only rose 1.4 percent. In other words, prices were advancing twice as fast as tithes and, if this is an accurate reflection of the relationship between price and output, then this is further confirmation that the rising demand for food and other agricultural products was having the effect of pushing up maize prices and perhaps other commodity prices two decades before the agricultural crises of the middle 1780s. Comparison of tithes and prices after those crises in the Bajío is not possible because of the absence of data. Tithes continued to grow at a rate of 1.6 percent between 1787 and 1810, a rate slightly above that for the period 1751–86 and comparable to the other rates for the eighteenth century.51 Although Morin does not stress the late third and early fourth quarters to the extent that I have, he does acknowledge that despite the growth in the tithe in the second half of the eighteenth century, it was losing ground to the growth in inflation. The increase in agricultural output, as measured by the tithe, may have doubled during the first half of the century but not during the second half argues Morin, because of inflationary pressures.52 Nominal output in agriculture rose throughout the century but real output (comparison of growth rates in tithe collections versus growth rates in maize prices) may have begun to lag in the second half.
A second set of tithe-collection figures (Graph 9) exists for the Diocese of Oaxaca.53 After considering all the problems of collecting and analyzing tithe accounts, the investigators conclude that their calculations point to “un incremento substancial” in the agricultural production of the Oaxacan region during the eighteenth century.54 For the Diocese of Oaxaca tithes rose at a rate of 0.7 percent per year between 1701 and 1800, with an average annual tithe of 60,979 pesos. We do not yet have a series of prices for the Oaxaca region and, therefore, we cannot relate directly trends in Oaxacan tithes and Oaxacan prices. Using the colonial average maize price, we see that prices rose at the same rate of 0.7 percent as calculated for Oaxacan tithes. When compared, Michoacán’s tithes rose nearly twice as fast as Oaxaca's tithes. The correlation (.86) between the two series was highly positive. When the trends are analyzed with respect to shorter periods—1701-51 and 1752—1800—they followed different courses in each of the regions. In Oaxaca, tithes rose at a rate of 0.7 percent per year during the first period and at a somewhat higher rate of 1.0 percent during the second period. In Michoacán, by contrast, the rate of 2.0 percent per year during the first period was faster than the rate of 1.6 percent per year during the second period. Oaxacan tithes sped up after 1750, but Michoacán's tithes slowed down. These figures indicate the extent to which growth rates could differ from region to region, although the regional differences should not be overstated. It is important to note here that the growth in Oaxacan tithes lagged behind the rise in the colony-wide maize series both in the first and the second halves of the eighteenth century, whereas Michoacán’s tithes were ahead of the series in the first half but not in the second. That Oaxacan agricultural output grew faster in the second half of the eighteenth century than in the first half may be explained by a less dense urban population and more tillable crop land than existed in Michoacán, certainly by the latter half of the eighteenth century.55 At the same time, we must point out that growth in Oaxaca’s agriculture did not ever match or exceed the rise in the colony-wide maize series. This may signify that in addition to the successful efforts by the Indian communities to protect their lands from Spaniards, agricultural prices remained too low to justify heavy investments in developing cropland, at least until late in the century. The high maize and bean prices, which we have for the last decade of the eighteenth century, might also signify that demographic changes came more quickly than Oaxaca’s agricultural economy, or especially its maize economy, could adjust to meet the demand.
Recent research has portrayed Mexico’s eighteenth-century economy in contradictory terms: on the one hand, there is expansion, reform, and change; on the other, there is inflation, scarcity, and crisis. At times Mexico seemed poised to take off” and at other times it seemed stymied by an archaic, cumbersome, and unresponsive system.56 From the standpoint of price-trend analysis, economic crosscurrents did exist in eighteenth-century Mexico. On the basis of the secular trend, the rise in prices appears reasonable and manageable within the confines of growth in population and output. When the secular trend is broken down into shorter periods, however, both for the first half and the second half of the eighteenth century, prices show a fairly strong tendency to rise, often faster than other economic series. Rising prices in the short or intermediate term underscore the trouble that the economy had in coping with demographic and economic changes. Real economic growth could not be easily achieved or sustained under conditions of rising prices, although the chances for growth were perhaps better in the first half of the eighteenth century than in the second. From the middle of the third quarter to the middle of the fourth quarter, a period often identified with economic growth and reform, the steady and rapid rise in prices may be an indication of how limited the productive capacity of the colonial economy was. Feeding the population, a long-standing problem, may have grown more severe in the course of the eighteenth century, even as agricultural output was expanding. A probable increase in the flow of people from the countryside to the city only served to exacerbate the problem of feeding the population. Price trends indicate that expansion in supply was losing ground to growth in demand, especially in urban areas.57 Famines offered periodic but only temporary adjustments to a potentially widening gulf between supply and demand. How much economic development, particularly in the agricultural sector, the colony could underwrite and absorb is, of course, an unknown quantity. Whether it was reached in the late eighteenth century may be overshadowed by a new twist to an old policy. In search of revenue to pay for a mounting defense bill, the crown launched a plan for the confiscation of as much hard coinage from private sources as it could find. The result was to limit further the capital resources, even those based on credit transactions, available for specifically agricultural improvement and generally for economic development at a crucial period of the demographic cycle.58
Enrique Florescano, Precios del maíz y crisis agrícolas en México (1708-1810): Ensayo sobre el movimiento de los precios y sus consecuencias económicas y sociales (Mexico City, 1969).
Florescano, Precios del maíz, pp. 115-117, 124, 180-182. Cuadro 7 (pp. 115–117) and Cuadro 19 (p. 181) present the ten cycles that Florescano identifies between 1721 and 1814. He computes several different price averages: an annual average, based on monthly averages for each year, and a cyclical average, based on the annual average for each year within each cycle. He also converts the annual averages and the cyclical averages to index numbers in order to show relative changes in these respective averages against selected base periods. Gráfica 16 illustrates the movement of annual average prices in the form of index numbers with 1726-42 as the base period (equal to 100), Cuadro 19, on the other hand, contains annual averages and cyclical averages for all cycles and it also contains index numbers (base period = first cycle, or 1721-27) for both averages. Although the index numbers differ because the base periods differ, the underlying cyclical pattern, which Florescano stresses, can be observed in the several different indices that he has constructed. Florescano’s analysis, like mine, is influenced by the periodization used. Determining periods for cycles or for trends involves decisions that may be logical but arbitrary. It should also be noted here that other scholars, such as David Brading and Claude Morin, share Florescano’s view in that price series for the localities and regions they have studied are more influenced by cycles than by trends.
The question of wages and salaries has been less thoroughly investigated than prices, although it is the subject of several studies now under way. Complications arise because wages and salaries were paid both in cash and in kind. Thus, the money component is only part of an individual’s total income. At this time, however, there is no compelling evidence that wages rose in step with prices during the eighteenth century.
Florescano, Precios del maíz, pp. 194-195. Recently, John Coatsworth and John TePaske have considered the impact of rising prices in studies concerned with economic growth and public policy. Coatsworth, “The Limits of Colonial Absolutism: The State in Eighteenth-Century Mexico,” in Karen Spalding, ed., Essays in the Political, Economic and Social History of Colonial Latin America, University of Delaware, Latin American Studies Program, Occasional Papers and Monographs, No. 3 (Newark, Del., 1982), pp. 32-34. TePaske, “Economic Cycles in New Spain in the Eighteenth Century: The View from the Public Sector,” in Richard L. Garner and William B. Taylor, eds., Colonial Latin American History: Essays in Honor of Charles Gibson, forthcoming.
Lennart Jörberg, A History of Prices in Sweden, 1732-1914, 2 vols. (CWK Gleerup Lund Sweden, 1972), II, 3.
The list for maize is long.
Woodrow Borah and Sherburne F. Cook. Price Trends of Some Basic Commodities in Central Mexico, 1531-1570 (Berkeley, 1958). Using tribute records, testimonies by contemporaries and various accounts of wholesale and retail sales they constructed a rather impressive series for the middle of the sixteenth century. The series is uneven, however. For some years (1569, for example) they found several dozen prices; for others they found only a few prices or none at all.
Charles Gibson, The Aztecs under Spanish Rule, A History of the Indians of the Valley of Mexico, 1519-1810 (Stanford, 1964), pp. 452-459. His sources were Borah and Cook and documents similar to those used by Borah and Cook. Some refinements had to be made in Gibson's series before it could be used. To avoid duplication prices quoted from Borah and Cook and from the eighteenth-century records of Mexico City’s albóndiga, were eliminated from the Gibson series. Gibson's series includes both ranges of prices and specific prices but I have only used the latter in constructing the colony-wide maize series.
Florescano, Precios del maíz. This remains the most meticulous and extensive investigation yet of late colonial (Mexico City) maize prices. Relying on albóndiga records Florescano calculated yearly average prices based on monthly average prices, which in turn were based on daily fixed prices. Florescano’s study also provides the first thorough examination of seasonal fluctuations and patterns.
D. A. Brading, Haciendas and Ranchos in the Mexican Bajío: León, 1700-1860 (Cambridge, 1978), pp. 180-183. This study contains long series of maize prices for León and Silao between 1660 and 1790. Brading explains how the series originated: “Calculated on the basis of the tithe-collectors’ reports, the prices represent annual averages. . . . The collectors were paid on commission, so it is probable that they sold the maize in the summer months when the supplies were most scarce. Their prices . . . may be somewhat higher than the overall market price for any given year” (p. 180).
Flor de María Hurtado López, Dolores Hidalgo: Estudio económico, 1740-1790 (Mexico City, 1974). Like Brading’s, this series is derived from tithe ledgers. (See pp. 29-30 for a brief discussion of the sources.) The data are presented in the form of annual average prices. The runs cover only about 50 years but they are remarkably complete. Hurtado López is primarily interested in cycles and even though the analysis undertaken for determining the cycles could have been rather easily extended to include the plotting of trend lines and the computation of growth rates, it was not used for these purposes (see Chapter two). A similar study for another parish, by Silvia Galicia Morales, did not come to my attention soon enough to be included. Data from Galicia Morales’s study and from others, as they are published, will be added from time to time to the existing data base.
Claude Morin’s Michoacán en la Nueva España del siglo xviii: Crecimiento y desigualdad en una economía colonial (Mexico City, 1979), pp. 188-200, contains a small series of maize prices for five towns—Guanajuato, Valladolid, Paracho, Huetamo, and Coalcomán—in the 1790s. A list showing the price of maize in June and December for each town was supposed to be prepared annually by the intendant. In the period 1792-99 two such prices were recorded for 1792, 1794, 1795, and 1797. For 1793, 1798, and 1799 only one price was recorded. For some years no prices were listed. The seasonal factor accounts for higher prices in the middle of the year than at the end of the year. Because of the paucity of data for each town, I have computed an annual average for all five towns. The result is a regional rather than a town average for each year.
Richard L. Garner, “Problèmes d’une ville minière mexicaine à la fin de l’époque coloniale: prix et salaires à Zacatecas (1760-1821),” Cahiers des Amériques Latines, 6 (1972). These were annual average prices based on what the vendors at the alhóndiga declared to be their selling prices. These would qualify as estimated prices since there was no way to check the declared price with the actual transaction. The Zacatecas article also contains prices for flour and twenty other commodities.
Hermes Tovar Pinzón, “Elementos constitutivos de la empresa agraria jesuita en la segunda mitad del siglo xviii en México,” in Enrique Florescano, Haciendas, latifundios y plantaciones en América Latina (Mexico City, 1975), pp. 219-220. Tovar Pinzón collected prices from accounts of Jesuit haciendas in different regions between 1767 and 1771. I calculated an annual average for the whole group rather than by single hacienda because no hacienda appears more than once on the list. As a result, these are more nearly like colonywide averages than local or regional averages.
Brian Hamnett. “Dye Production, Food Supply, and the Laboring Population of Oaxaca, 1750-1820,” HAHR, 51 (Feb. 1971), 76-77. The.se data have been collected from intendancy and tithe accounts. Some of these data appeared as biannual prices from which an annual average was computed.
For wheat/flour, the sources are Hurtado López, Garner, and Tovar Pinzón. The same sources, and Hamnett, have been used for prices oí frijoles.
Peter Kennedy, A Guide to Econometrics (Cambridge, Mass., 1979), pp. 2-6, 10-12. Discussions on the advantages or disadvantages of regression analysis are found in standard textbooks on business statistics or econometrics. The econometric program I have used is called Econpak, and was developed by Dr. Milton Hallberg, Professor of Agricultural Economics in the Department of Agricultural Economics and Rural Sociology at the Pennsylvania State University. The program has many different options, including a “dummy variable” option, which I have used to analyze segments of straight-line trends. The printout shows (among other things) regression coefficients, r-squared values and t distributions, all of which can be used to test the validity of the analysis. My cutoff for the t statistic was a significance level of .01, that is, a 99 percent probability that a trend was not simply the result of chance, unless so indicated in the footnotes.
For a discussion of price and population movements and demand and supply elasticities. see Clyde George Reed, Price Data and European Economic History: England 1300-1600 (New York, 1981), pp. 1-25.
Cook and Borah, Price Trends, pp. 48-49.
Sherburne Cook and Woodrow Borah, Essays in Population History. Vols. 1 and 2: Mexico and the Caribbean (Berkeley, 1971), I, 310-312. 354-355.
Eric Van Young, Hacienda and Market in Eighteenth-Century Mexico: The Rural Economy of the Guadalajara Region, 1675-1820 (Berkeley, 1981). pp. 220-224, 348-349.
I have tried various combinations for 1787-1809 with these results:
1787-1809 0.8% per year 76% probability
1787-1792 −10.9% per year 98% probability
1793-1809 2.2% per year 97% probability
1787-1794 −5.0% per year 88% probability
1795-1809 2.5% per year 94% probability
Space does not permit a discussion of the recent price index constructed for Santiago, Chile, except to remark that the Chilean index showed a rise of under one-half of one percent per year from 1749 to 1808 while the colony-wide maize series showed an annual increase of slightly more than two percent from 1751 through 1809. It is also evident that peaks and valleys in the price curves for Chile and Mexico correspond closely, although not exactly. José Manuel Larraín Melo, “Movimiento de precios en Santiago de Chile, 1749-1808, una interpretación metodológica,”Jahrbuch für Geschichte von Staat, Wirtschaft und Gesellschaft Lateinainerikas, 17 (1980), 198-259, but esp. 220-224.
Richard L. Garner, “Silver Production and Entrepreneurial Structure in 18th-Century Mexico,” Jahrbuch für Geschichte von Staat, Wirtschaft und Gesellschaft Lateinamerikas, 17 (1980), 158. Some of the growth rates for the silver data have been calculated specifically for this paper. Using a difíerent approach, John Coatsworth has prepared a set of tables that show the impact of price changes on silver registrations. What all these tables suggest is that by the late eighteenth century the real value of silver fell as maize prices rose.
Brading, Haciendas and Ranchos, pp. 68-69.
To distinguish between prices in the Valley and prices in the capital, I removed those that Gibson specified as Mexico City granary prices.
When the dummy-variable option (two periods, 1721-50 and 1751-1809, instead of one. 1721-1809) is applied to the granary series, the r-squared values, corrected for degrees of freedom, rise from 21.7 percent to 37.1 percent.
By using a simple test, coefficient of variation (the square of the standard deviation divided by the mean), we can establish that León-Silao prices were slightly more volatile than Hidalgos (6.5 vs 5.1) and both were significantly more volatile than Mexico City's (3.0) for the period 1753-87. This period was chosen because the data for all three cities are fairly complete. It is possible to calculate coefficients of variation for a longer period (1724–87) for the Mexico City and León-Silao series. These calculations, 2.1 for Mexico City and 5.7 for León-Silao, yield a slightly greater magnitude of variation than with the above figures. For a critique of the use of variation coefficients in trend analyses, see Milton Hallberg, “Cyclical Instability in the U.S. Dairy Industry without Government Regulations,” Agricultural Economics Research (U.S. Dept. of Agriculture), 34:1 (1982), 11.
Gibson, Aztecs under Spanish Rule, p. 329. The calculations of output per unit are mine. Florescano, Precios del maíz, pp. 182-193, but esp. pp. 183-186.
Florescano, Precios del maíz, pp. 181, 233–234.
Donald B. Cooper. Epidemic Disease in Mexico City, 1761-1813: An Administrative, Social, and Medical Study (Austin, 1965), p. 70.
See Linda L. Greenow’s discussion of R. Spillman’s mapping of the 1786 famine in Credit and Socioeconomic Change in Colonial Mexico: Loans and Mortgages in Guadalajara, 1720-1820 (Boulder, 1983), pp. 170-173, and Figure 5.5. p. 175.
Van Young, Hacienda and Market, p. 103.
Brading, Haciendas and Ranchos, pp. 180–183.
Ibid., p. 183.
Hurtado López, Dolores Hidalgo, pp. 31-49.
Morin lists maize prices for Guanajuato, Valladolid, Paracho, Huetamo, and Coal-coman between 1792 and 1799 in Michoacán en la Nueva España, p. 196.
For the period 1787-99 the rate was -6.99 percent with a probability of 94 percent.
Brading, Haciendas and Ranchos, p. 20.
Brading, Haciendas and Ranchos, pp. 69-71. Some slightly different maize figures are reported for Silao on p. 20.
Brading, Haciendas and Ranchos, pp. 58-60, 69-73, 113-114.
These calculations have been done in connection with an article in which I have compared silver-production rates in Mexico and Peru.
Garner, “Problèmes d’une ville minière mexicaine," Cahiers des Amériques Latines, 6 (1972), 110-111. Hurtado López, Dolores Hidalgo, p. 93. Morin, Michoacán en la Nueva España, pp. 180-200.
Gibson, Aztecs under Spanish Rule, pp. 323-324.
Brading, Haciendas and Ranchos, pp. 70-71.
Morin, Michoacán en la Nueva España, p. 117.
Van Young, Hacienda and Market, p. 59.
Van Young, Hacienda and Market, pp. 61-63.
Van Young, Hacienda and Market, pp. 70-71.
Van Young offers the evidence of one estate. Toluquilla, in 1796-97, when wheat accounted for 35 percent of its income. Hacienda and Market, pp. 219-220, 226-228, 234, 349-350.
Most of Hurtado López’s prices are rounded off to whole numbers. Also Hurtado López calculates three fanegas equalled one carga while Van Young estimates two fanegas per carga. I am inclined to agree with Van Young; therefore, I have used his ratio in converting fanegas to cargas. Where I have applied regression analysis to the Hidalgo series without regard to other series, I have used the prices in reales per fanega rather than pesos per carga.
Garner, “Problèmes d'une ville minière mexicaine,” Cahiers des Amériques Latines, 6 (1972), 6, 8.
Morin, Michoacán en la Nueva España, p. 198.
Garner, “Problèmes d'une ville minière mexicaine,” Cahiers des Amériques Latines, 6 (1972), 6, 8. For Guadalajara, Van Young has estimated how much wheat and flour entered the city while for Zacatecas I have used how much flour was reported to have entered the city by the real aduana. Hacienda and Market, Table 5, p. 60, and “Problèmes d une ville minière mexicaine,” p. 88. Zacatecas’s population trends are discussed in Richard Garner, “Zacatecas, 1750-1821: The Study of a Late Colonial Mexican City” (Ph.D. Diss., University of Michigan, 1970), chap. 2, and Guadalajara’s trends are treated in Chapter 2 of Van Young’s Hacienda and Market.
Van Young, Hacienda and Market, p. 98.
Gibson, Aztecs Under Spanish Rule, p. 316, and Van Young, Hacienda and Market, p. 98.
Treasury Records and Correspondence, 1790-1799, Box A, Zacatecas Collection, Clements Library, University of Michigan.
Hurtado López, Dolores Hidalgo, cuadro 27, pp. 58-59.
Numerous problems have to be overcome in the construction of a tithe series. Morin first discusses the methodological problems of constructing the series and then the analytical problems once the series was constructed. In addition to misrepresentation and outright fraud in reporting the tithe, by both the producer and the commissioner who handled the transactions for the dioceses, there were matters relating to exemptions for Indian villages and losses of tithe districts to other dioceses. Michoacán en la Nueva España, pp. 102-107 for the methodology, and pp. 107-121 for the analysis.
If we could demonstrate that tithes represented the value of output in Michoacán and prices represented the relative change in that value, we might then have a way to make some rough calculations on growth or lack of growth in the late colonial economy. Indeed, through multilinear regression analysis we can measure the rate of increase in the tithe against changes in both time and price. The figures for specified time periods and price trends are as follows: an increase of 1.5 percent per year in the tithes against two independent variables of a colony-wide maize average and time, 1700-1809, compared to 1.5 percent per year against a single independent variable of time; an increase of 1.5 percent per year in the tithes against the aforementioned two independent variables from 1751 to 1809, compared to 1.7 percent per year against the single independent variable of time; an increase of 1.1 percent per year in the tithes against the two independent variables of the Bajío maize average and time, 1751-86, compared to 1.4 percent per year against the single independent variable of time. In effect, we can say that the rise in prices (as measured by maize) over time reduced the value of production in the agricultural sector. If these "deflated” tithe figures were compared with population growth rates, we might have a clearer understanding of whether a serious agricultural crisis was in fact unfolding. Tentatively, at least, it can be suggested that if the population was growing at between 2 and 3 percent a year in the second half of the eighteenth century, then such growth was beginning to outstrip the productive capacity of the agricultural sector to keep up.
Morin, Michoacán en la Nueva España, p. 116.
Rodolfo Pastor, L. Adleson, E. Berra, F. Hurtado, J. MacGregor, G. Zermeño, and E. Trabulse, eds. Fluctuaciones económicas en Oaxaca durante el siglo xviii (Mexico City, 1979). See the extensive tables beginning on p. 66. In this essay I have used only the figures for the Diocese of Oaxaca although the authors present separate figures for Mixteca (in two columns) and for the Diocese of Oaxaca. See the map of the Diocese of Oaxaca, p. 24.
Fluctuaciones económicas en Oaxaca, pp. 33, 35.
William B. Taylor, Landlord and Peasant in Colonial Oaxaca (Stanford, 1972), pp. 107, 110, 138, 140–143, 162-163, 198, 201, and Morin, Michoacán en la Nueva España, pp. 101-120.
Coatsworth, "Obstacles to Economic Growth in Nineteenth-Century Mexico,” American Historical Review, 83:1 (1978), 91-94.
A simple calculation—summing the population figures for the intendancy capitals—shows that these major urban centers comprised fewer than 250,000 or from 4 to 5 percent of the total population. Alexander von Humboldt, Political Essay on the Kingdom of New Spain, 4 vols., trans. by John Black (New York, 1966), I, 97.
Garner, “Exportaciones de circulante en el siglo xviii (1750-1810),” Historia Mexicana, 124 (1982), 544-598. See also TePaske, “Economic Cycles in New Spain in the Eighteenth Century.” A cruciai question to be investigated and answered if we are ever to chart Mexican economic development between 1785 and 1810 concerns the availability of capital. On the basis of numerous hacienda studies and now Greenow’s Credit and Socioeconomic Change, we know that the business of lending money and extending credit by ecclesiastical organizations and private individuals grew substantially in the late colonial period. In the aftermath of the famines of 1784-86 and new imperial policies concerning forced loans and donations the amount of money available for investment may actually have shrunk in the quarter century before the Hidalgo rebellion. Greenow’s figures indicate that lending activity may have begun to tail off during the 1790s and 1800s, although her analysis does not address this issue directly. Figure 2.1 (p. 27) and Figure 5.1 (p. 154) show a decline in total annual borrowings and in borrowings by select groups from about 1790 to 1820. She characterizes the period from 1780 through 1797 as one of “unprecedented levels of borrowings, perhaps as a result of a series of economic crises in 1785 and 1786.. . .” (p. 26). In fact, total borrowings began to decline shortly after the end of the crises and continued in a downward fashion until the end of the colonial period except among the merchants, whose borrowings continue to climb until the middle of the 1790s before turning downward. What accounts for this decline in borrowings remains obscure, although the possibility exists that as a result of the crown’s large-scale borrowings and confiscations the pool of capital available to private investors had virtually withered away.
For their comments and criticisms I am indebted to John Coatsworth, University of Chicago, and to Milton Hallberg, Rex Warland and Spiro Stefanou, the Pennsylvania State University. I am also indebted to Glenn Kreider and his staff at Penn State’s Liberal Arts Data Lab for their advice and assistance. Finally, I am grateful to Penn State’s College of Liberal Arts for providing computer funds and to the American Philosophical Society and the American Council of Learned Societies for providing research grants.