Version classiqueVersion mobile
OpenEdition Books

Des Indes occidentales à l’Amérique Latine. Volume 2

Thomas Calvo
Alain Musset

Le travail et l’argent

Prices and the economic history of colonial Mexico

Richard L. Garner

Texte intégral

1Historians have long relied on series of prices to track the behavior of economies. How to analyze and interpret these series has often sparked as much controversy as agreement among scholars. Modem price theory rests on the premise that prices of goods and services will affect the choices that producers and consumers make with regard to what they purchase or do not purchase. By exercising choice buyers and sellers determine the allocation of resources in economies. It cannot be said, of course, that such a premise would fit earlier times. Thus, series of prices from periods and places in which market structures and price mechanisms, as they are understood in the modem world, were in their infancy may represent a less pure application of the theory than is possible in contemporary times. Even now in the most market-oriented economies prices do not always truly reflect the degree of choice that the theory may assume. Still it would be absurd to argue that price series from the sixteenth through the eighteenth century have the same theoretical integrity of price series from the twentieth century.

2The fact remains, however, that some prices from the early modem period and from economies that embraced an array of structures from feudalism to capitalism were based on forms of bargaining over what buyers and sellers would accept. The constraints are numerous: product variety was limited, and since substitutes did not always exist, consumer choice did not exist either: or public policy was designed to regulate economic behavior and to achieve certain social goals irrespective of unfavorable economic consequences. Other examples may well come to mind. In short the study of prices during the early modem period entails burdens and risks that modem price researcher may not have to consider (although they too have their own constraints to consider).

3In this brief essay I will discuss Mexican colonial price history in two different con- texts. First, I want to reexamine some of the series that have been published during the past decade and the findings that they have yielded. As in the past, I am especially interested in long-term patterns that can be identified within the various price series and how these patterns may be related to some larger historical questions. In addition I want to suggest some channels of inquiry and some techniques for analysis that current and future economic historians may consider. As more data are being compiled and published, scholars must be prepared to apply more sophisticated statistical techniques to tease out as much information from these numeric series as we can.

4Let me begin this discussion with the colonial maize price series that I published a decade ago and I have since revised. This series includes prices from different types of maize transactions from various regions across the colony. The colony-wide series has been criticized on technical grounds - that it includes too many different types of maize transactions from regions that had distinct economic system and that it implies the existence of a market. The purpose behind the creation of this series was not to try to substitute an average colonial price for local or regional prices that obviously reflect more accurately local or regional economic activities. Nor is there any design to superimpose a colonial market where one may not exist, although it is debatable in my view that colonial Mexico's economy, even in the sixteenth century, was so oriented to local and regional conditions that a more complex market system could not operate with certain products and transactions. The purpose behind the calculation of colony-wide averages and trends is to establish a standard against which to measure the performance of the economy from the local to the colonial level over time. What we have learned from the comparing of colony-wide series with local or regional series is that they share some common properties. We should not be surprised at this. Colonial Mexico was bound together in a variety of ways, from geography to climate and from commodity production to mercantilistic policy.

5The importance of maize in colonial Mexico is well documented. Even though Spaniards introduced wheat in sixteenth century because their preference for bread corn remained a dietary staple until the end of the colonial period. Maize prices are among the most readily available of all price data. The absence of other long-term price series has forced scholars to use various maize-price series as reasonable substitutes for a more broadly-based commodity-price series. As we construct price series for different commodities, we are learning that growth rates for maize prices tend to be somewhat higher than these other commodities. That suggests, of course, that we need to create an inflation index in which maize will be an essential component but not the only one.

6Our maize series begins in 1525 and continues until 1820. For the sixteenth and seventeenth centuries the data are drawn from Central Mexico, mainly the Valley of Mexico, and from Nueva Galicia. After the seventeenth century the series is widely represented with data from northern, central, and southern Mexico. Drawn from an array of transactions - public auctions. alhóndiga sales, and tithe payments - this series is a composite of as many transactions as could be assembled. Between 1525 and 1820 maize prices rose at of 0.4 percent annually. The R2 value indicates how well the actual observations fit the trend. To avoid using too many statistics we have arbitrary classified R2 values as follows between 75 and above as indicators of very strong trends, 50 and 75 as strong, 25 and 50 as weak, and 25 and below as very weak. We will also note if any findings fall below the confidence level of 99 percent. In the case of this very long maize series the trend is weak. although after correction for autocorrelation it moves much closer to the strong classification. The trend is affected in part by the volatile character of the maize prices. The mean price (1525-1820) was 12.3 reaies per fanega, but prices fell as low as 1 real per fanega and rose as high as 58 reaies per fanega. Maize prices may well have been among the most volatile of all prices. Still, over the colonial era the rise in maize prices was at best modest. At an annual compound rate of 0.4 percent the price of maize would double about every 125 years.

7Figure 1 (in a semi-logarithmic scale) illustrates the path followed by maize prices over almost three centuries. Three distinct phases emerge. Prices rose during the sixteenth century until the middle of the seventeenth century. The absence of seventeenth-century data makes it difficult to determine exactly when prices leveled off and turned downward. The second-phase decline continued until the early eighteenth century after which for almost two quarters it was stuck in a range comparable to maize- price levels of the second half of the sixteenth century, in 1770 an upswing began and continued irregularly until the end of the colonial period.

8We can calculate a annual growth rate for each of the three distinct phases in the long-term colonial trend. From mid-1520s to the mid-1640s maize prices followed a strong upward trend at an annual rate of 2 percent. During the second phase (1640s to 1770s) they fell at a rate of 0.7 percent per annum, although the trend is very weak. Finally in the third phase (to the end of the colonial period) maize prices turned up again at a rate of 2.2 percent yearly with a trend that was close to strong. These rates are mainly useful for comparative purposes. As previously noted, the mean price from 1525 to 1820 was 12.3 reaies per fanega and the rate of increase 0.4 percent per year. In the first phase the mean price was 9.6 fanegas per fanega and the rate of increase was almost five times higher than the colonial rate. In the second phase the mean price of 11.1 reaies per fanega was close to but still below the colonial mean and the rate was negative rather than positive. In the third phase the mean price shot up to 17.1 reaies per fanega and the growth rate, now positive, was nearly six time the colonial rate. Although the rises and the decline appear to be self-evident from Figure 1, the statistical tests indicate the degree to which the trend explains the behavior of the actual data. The very weak downward trend during the second phase may be largely the result of too little data. It must be understood, however, that even if we had more data to analyze the trend may remain weak. Fluctuation rather than downwardness might best characterize the behavior of maize prices during the middle period of the colonial era. At the moment the negative rate for the second phase is at best tentative. It is not so much a question of whether a downward trend existed: rather it is a question of the intensity of the trend.

fig. 1 - Colony-wide maize prices

9A colonial maize series, despite its flaws, can be useful in the analysis of other numeric series. For example, we can now refer to several long tithe series (Puebla, Michoacán, and Oaxaca) and a long silver series. In comparison with maize tithes and silver established much stronger upward trends and much faster growth rates over the long and the very long term. This raises obvious questions about the extent and the impact of price inflation, certainly over any extended time period. Of special interest is the relationship between silver production and price inflation. From time to time, no doubt, a burst in production of silver could push up prices of some goods, but overall the inflationary effect of the silver curve on the price system may have been less than significant. In ail likelihood high levels of bullion exports moderated the inflationary effect of the silver expansion on the colonial economy.

10Publication of price data on maize and other commodities from Nueva Galicia in the second half of the sixteenth century by Woodrow Borah has opened a new line of statistical inquiry and analysis. The commodities that 1 have analyzed are maize, wheat, frijoles, honey, and chili. The region defined as Nueva Galicia includes both the «plateau» and the «coast» areas but excludes Zacatecas, although data from Zacatecas is included in Borah's tables. Not surprisingly volatility is an important characteristic of these series. In this dataset of five commodities the grains - wheat and maize - show less volatility, based on the coefficient of variation, than the other commodities, although maize is slightly more volatile than wheat. In trying to pinpoint more precisely price trends for individual commodities we encounter difficulties. The subsets within the Nueva Galicia price database do not have chronological consistency. The longest series is for maize - from 1557 to 1598 with data for 39 of the 42 years - and the shortest is for chili - 1558-1577 with 18 of the 20 years. The regression statistics are as follow:

Fig. 2 - Nueva Galicia maize and when prices

11Three commodities - maize, wheat, and frijoles - have positive growth rates but except for maize they have very weak trends and unacceptable t values. It is worth noting that the behavior of maize prices does not necessarily influence the behavior of other commodity prices. Using maize prices as a standard for measuring the performance of prices in general and in particular the impact of inflation should not be done without some qualification.

12The relationship between the two grains - maize and wheat - offers some interesting comparative features. We have limited the two series to the period (1557 to 1586) in which they share annual data. The trends (Figure 2) change slightly. For maize the growth rate falls to under 2 percent per annum, but the trend remains strong, although just barely so. For wheat the trend is extremely weak. In an attempt to explore this relationship further we have converted each dataset to its natural logarithm, calculated the yearly changes, and plotted those changes on Figure 3. Year-to-year changes in maize and wheat prices show some similarity, but, when the correlation coefficient is calculated and then converted to an R2 value, it indicates a very weak correlation between the two variables of only.11 Until the mid-1570s the amplitude of the yearly price changes for wheat fell in the range of +/- 20 percent, but in the final decade it reached almost as much as +/- 80 percent. The amplitude of change for maize did not match the wide swings for wheat in the 1570s and 1580s, but overall the amplitude for maize was somewhat greater than for wheat.

13From the supply side the explanation for the behavior of maize prices is more obvious than that for wheat prices. Supplies assigned to royal auctions declined at a rate that almost matched the rate of increase in maize prices. The relationship between the price and supply of maize illustrates a commonly accepted view that as supplies decline and in the absence of a suitable alternative prices will rise. Demand, climate, or disaster may have influenced the maize-price series based on royal-tribute auctions. Wheat supplies declined sharply, but unlike maize wheat prices only rose mildly in response to that decline. Borah suggests that wheat became less important among the royal tribute commodities because as Spaniards increased wheat cultivation to meet local demand, they came to depend less on Indian supplies. That still leaves unexplained the wide fluctuations in wheat prices during the 1570s and 1580s. The transition from dependence on Indian communities to Spanish estates may have led to periodic shortages in supplies and sudden escalations in prices. Moreover, the wheat-supply data year to year from the mid-1570s into the mid-1580s may be a less reliable indicator of overall regional wheat production than in previous decades or for maize supplies.

14Comparisons can be drawn between the Nueva Galicia maize series and one for Central Mexico (the lutter combines data from Charles Gibson and Ulises Beltrán) from 1557 to 1590 (with comparative data for 23 years). The two series do not have a «lockstep» relationship year by year. When the correlation coefficient is calculated for the two variable - Central Mexico maize prices and Nueva Galicia maize prices - it is moderately strong at.64. When the correlation coefficient is converted to an R2 value, it is only.41. In Figure 4 they both show a tendency to rise. They do share similar growth rates, although the 2.6-percent annual rate in Nueva Galicia may be classified as strong and the 2.4-percent in Central Mexico as weak. The upward slope of the two disaggregated maize series reinforces the findings of the colony-wide maize series that prices were rising across the whole colony of the second half of the sixteenth century.

Fig 3 Yearly changes in Nueva Galicia maize and when prices

fig. 4 - Nueva Galicia and central Mexico maize prices

15Lacking a large enough database to allow for a more detailed analysis of the seventeenth century. we now turn to the final century of the colonial period. The maize series is particularly rich, and the wheat series is being added to constantly. The maize series alone could be the subject of a single essay. Let me highlight what the data reveal about the price history of the eighteenth century, and let me underscore what problems exist with the eighteenth-century maize-price series. From 1700 to 1810 or 1820 the rise in the price of maize was less than 1 percent per annum. Depending on the terminal date the rate could be 0.7 or 0.8 percent annually. That is at least two times faster than the rate for the whole colonial period. This was probably the fastest rate of the three centuries, although the gaps in the maize database for the sixteenth and seventeenth centuries make any statistically-satisfying comparisons impossible. No doubt the recovery of the population and the more intensive use of the natural resource base helped to push up maize prices. Whether this constitutes an eroding inflation remains open to debate. In the past I and others have hypothesized that if maize prices were substituted for a broadly-based price series, then rates at or near the 1 percent level probably imposed special economic burdens on the late colonial economy. As more non-maize price data are being analyzed, the role assigned maize within colonial price history may be overstated. Important and essential as maize was to some groups of buyers and consumers in eighteenth-century Mexico, its role as a price pacesetter may need to be reevaluated.

16As was the case in the sixteenth century the existence of series of wheat prices has forced some reappraisal of late colonial price history. In my recent book. Economic Growth and Change in Bourbon Mexico (Chapter I) I have discussed estimates of economic growth in eighteenth-century Mexico. Based on the analysis of maize and wheat prices in the eighteenth century, a long-term rate of inflation should be adjusted downward to between 0.3 or 0.4 percent per year. The reason for this adjustment is that while maize prices moved upward (weakly to strongly) during the century wheat prices (mainly from the Upper Bajío) were flat (see Economic Growth and Change in Bourbon Mexico, pp. 29, 31). Certainly by the eighteenth century urban consumers substituted bread for corn, and higher maize prices did not automatically mean any significant reduction in the consumer's living standards.

17A maize and a wheat series for Mexico City can be compared from 1740 to 1805. It should be noted that maize prices were then regulated at the alhondiga whereas wheat prices were not. Even so the growth rate for maize (0.7 percent year) is higher than the rate for wheat (0.4 percent per year), although in both cases the trends must be classified as very weak. In Figure 5 we find an upward trend is somewhat more identifiable wit maize than wheat. In addition not only were Mexico City maize prices rising faster than wheat prices, but they were considerably more volatile as well. The coefficients of variation were.41 for maize and.25 for wheat. In Figure 6 after charting the yearly changes for the two grains and calculating the correlation (based on an R2 value) we find that almost no relationship between these two variables.

Fig. 5 - Mexico city maize and when prices

18The underlying forces at work in the agricultural economy of the Valley of Mexico are still to be investigated in detail. By the eighteenth century, certainly in the Valley of Mexico, wheat was securely fixed in the hierarchy of agricultural commodities. For the producer wheat may have offered less risk and more stability even though wheat prices were rising less swiftly than maize prices. The volatility of maize prices was such that they tended to be low for longer periods than they were high and this made maize farming less attractive to investors. Rising maize prices may have resulted in part because less was being produced commercially, but the inflationary effects from these rising prices may have been softened simply because wheat prices were rising more slowly than maize prices and bread could be (and was being substituted) for corn by urban consumers.

19The picture of the Latin American colonial economies is evolving. Researchers are becoming more sophisticated about the collection and the analysis of historical numeric data. The argument in behalf of a solidly-constructed price index need not be repeated here. It is generally agreed that without some gauge of the level and movement of prices we are lacking an important tool for economic analysis. Up to now maize prices have served that role in colonial Mexican economic history. The addition of new price series has lessened the dependence on maize. As so often happens with new series and new techniques the water grows muddier before it become clearer. For colonial Mexico long- term price trends based solely on maize may impose economic hardship. When tempered by price trends for other commodities the hardship may be less apparent. Hardship and whether it was more or less constrained must be understood in the context of a society where most probably existed at subsistence or slightly above. The aim is not to dispute the presence of hardship but rather to create as accurate a framework as possible in order to analyze that hardship.

Fig. 6 - Yearly in Mexico city maize and when prices


20Since this essay is designed to provoke discussion rather than offer definitive solutions. I have dispensed with the usual scholarly apparatus of footnotes and citations. Sources on colony-wide maize and wheat prices are given in my Economic Growth and Change in Bourbon Mexico, Gainsville. FL. 1993. Since that publication I have added data from Woodrow Borah, Price Trends of Royal Tribute Commodities in Nueva Galicia, 1557-1598, Berkeley and Los Angeles. CA. 1992. The Mexico City maize prices are from Enrique Florescano, Precios del maiz y crisis agrícolas en México, 1708-1810, Mexico City, 1969. and the Mexico City wheat prices are from Virginia Garcia Acosta. Los precios del trigo en la historia colonial de México. Mexico City. 1988. In the latter case I have used her calendar year series rather than her harvest year series.

21Regression coefficients are calculated from the ordinary least-squares equation in logarithmic form: logy=loga+logbX+e. Growth rates are the equivalent of compound rates. Where autocorrelation existed corrections were made. I pay close attention to R- and Student t values. I have arbitrarily chosen to classify trends from very strong to very weak on the basis of the R2 values. In my view historical datasets, such as the ones analyzed herein, will seldom vield high R2 values, and values of.5 and above should be treated as sufficiently high to take the trends seriously. In addition visual observation remains an important tool for understanding patterns and trends.

Table des illustrations

Légende fig. 1 - Colony-wide maize prices
Fichier image/jpeg, 241k
Titre tableau 1
Fichier image/jpeg, 108k
Légende Fig. 2 - Nueva Galicia maize and when prices
Fichier image/jpeg, 161k
Légende Fig 3 Yearly changes in Nueva Galicia maize and when prices
Fichier image/jpeg, 180k
Légende fig. 4 - Nueva Galicia and central Mexico maize prices
Fichier image/jpeg, 178k
Légende Fig. 5 - Mexico city maize and when prices
Fichier image/jpeg, 141k
Légende Fig. 6 - Yearly in Mexico city maize and when prices
Fichier image/jpeg, 194k

© Centro de estudios mexicanos y centroamericanos, 2006

Conditions d’utilisation :