1. INTRODUCTION
Were working-class Mexicans better off after the country’s independence or did they have higher standards of living under Spanish domination? How did the late 19th century’s first wave of economic modernisation and globalisation affect the livelihoods of Mexican workers and their families? Today Mexico lags behind the most advanced economies of the world, but was this always the case? If not, when did Mexico start to fall behind? This article seeks to find answers to these simple but fundamental questions about Mexico’s economic history by reconstructing trends in real wages in Mexico City from the mid-18th to the early 20th century. Real wages (the purchasing power of wages in terms of a representative basket of goods and services) serve as an indicator of both material welfare and economic productivity.
Our period of study encompasses the foundational moment of the Mexican nation. During the second half of the 18th century, the Spanish crown sought to gain a tighter rein on the colony through economic and political initiatives known as the Bourbon reforms. In the 1970s and 1980s, historians indicated that living standards had declined in the aforementioned period (Florescano Reference Florescano1969; Coatsworth Reference Coatsworth1978; Tutino Reference Tutino1986; Van Young, Reference Van Young1987), but more recently, revisionist positions have problematised the use of evidence and highlight the undergoing process of economic growth led by the mining sector and trade (see Dobado-González (2015) for an overview of the revisionist arguments). The political crisis unleashed by the French invasion of Spain brought widespread insurrection in the 1810s and the destruction of the most productive mines. The ensuing decades of recurrent instability, civil and foreign wars are truly a terra incognita in Mexican economic and social history (McCaa Reference Mccaa1993). The economy reoriented, peasants gained more autonomy and perhaps political leverage, but the overall trends in productivity and welfare remain hard to decipher (Tutino Reference Tutino1986; Cárdenas Reference Cárdenas2003; Dobado et al. Reference Dobado González, Gómez Galvarriato and Williamson2008; Sánchez Santiró Reference Sánchez Santiró2009). The economic landscape changed in the 1870s when the triumph of liberalism and the Porfiriato (as Porfirio Díaz’s aegis is known) ushered in policies geared to modernising economic institutions, integrating Mexico in the global economy and industrialisation. Historians who examined general economic and social trends have emphasised the growth of the economy, yet the deterioration of living conditions for the majority of the population (Coatsworth Reference Coatsworth1978; Katz Reference Katz1981; Knight Reference Knight1990). However, empirical studies suggest either stagnation or a small improvement of real wages of the working classes during the Porfiriato (López-Alonso Reference López-Alonso2012; Gómez-Galvarriato Reference Gómez-Galvarriato2013). Our period concludes with the Mexican Revolution (1910-1920) and its aftermath characterised by price spikes, increased leverage of organised labour and improvements in the earnings of industrial workers (Bortz and Haber Reference Bortz and Haber2002; Gómez-Galvarriato Reference Gómez-Galvarriato2013). In sum, from Mexico’s perspective these 200 years were nothing short of Polanyi’s Great Transformation experienced throughout the western world, but we know very little about how it correlated with living standards and productivity.
From a global perspective, this is the period when modern disparities in economic welfare and productivity widened (Pomeranz Reference Pomeranz2000). Economic historians working on real wages emphasise that this variable provides a better gauge of «the great divergence» because, at least before the 20th century, GDP estimates are unrepresentative aggregates (of land, capital and labour) based on dubious assumptions. By contrast, real wages allow the comparison of the differences in productivity as experienced in the compensation of workers. In particular, Robert Allen’s methodology has enabled comparisons encompassing different regions of the world by systematising measurements, wages and prices, and devising a standard metric of economic welfare (Allen Reference Allen2001; Allen et al. Reference Allen, Bassino, Ma, Moll-Murata and Van Zanden2011). Studies have applied Allen’s methodology to certain cities in Latin America during the colonial and early national periods (Allen et al. Reference Allen, Murphy and Schneider2012; Arroyo Abad et al. Reference Arroyo Abad, Davies and Van Zanden2012; Arroyo Abad Reference Arroyo Abad2013, Reference Arroyo Abad2014). These works show that real wage levels in non-European regions were comparable to northern European countries before 1800, but that by the second half of the 19th century the divergence in incomes became noticeable.
Our study builds the first internationally comparable index of real wages for Mexico City bridging the 18th and the 20th century. We used the income of unskilled construction workers and weighted price indexes of thirteen products including food, fuel and other everyday items. Mexico City, the capital and a city that surpassed half a million inhabitants by the early 20th century, is a good case for an in-depth study of living standards because of the availability of information throughout the period (Miño Grijalva Reference Miño Grijalva2004, Reference Miño Grijalva2006; INEGI 2009). Moreover, broader conditions in the populous central region influenced labour and goods markets in the city. Mexico City’s population fed from migration from rural communities and towns in the surrounding region and beyond throughout the period (Guerrero 1901, p. 161; Pescador Reference Pescador1992, p. 108; Piccato Reference Piccato2001, pp. 21-26; Pérez Toledo and Klein Reference Pérez Toledo and Klein2004; Scardaville, Reference Scardaville2004). The large size of the city also meant that it drew supplies from surrounding areas. For these reasons, Mexico City was a barometer of labour and goods markets in the central region of the country. Similar studies from other regions will be necessary to establish representative trends for the country.
The article builds long-term series of unskilled real wages based on construction workers’ wage rates, and two alternative consumer price indexes, following Robert Allen’s welfare ratio methodology. Section 2 outlines the procedures and sources and discusses the pitfalls and advantages of the methodology. Section 3 discusses the evolution of real wages in connection to other trends in the economy and the country. Section 4 compares the trends with other economies of the time. In the conclusion, we discuss the broader implications of our empirical findings.
2. METHODOLOGY AND SOURCES
The building of real wage indexes across time and countries requires data on wages and prices, as well as a household basket to weigh the relative importance of items in the overall household budget. In order to build internationally comparable welfare ratios, it is necessary to rely on assumptions that do not generate long-term or country-specific biases. These assumptions are simplistic, yet they provide a reliable broad comparative framework. It is our purpose to provide a general measure of material constraints in everyday life that allows historians to benchmark economic welfare at different points in time. This analysis is not a substitute for more historically rich studies of work, diet and welfare.
2.1. Wages
Our wage sources are hundreds of weekly receipts of construction works that report the daily pay rates of labourers (peones) in Mexico City. Through the mid 19th century, the origin of the reports is varied: churches, convents, hospitals, jails and other public and religious institutions. From the 1830s most observations originate in the construction work contracted by the Vizcaínas school to repair their facilitiesFootnote 1 . Regardless of the source, the format is the same: the overseer reported the names, days worked, pay rates and total weekly compensation. For all weekly reports, we collected information on the number of days worked and daily pay rates, from which we obtained weighted averages of daily wages. In all, we have data for over 240,000 man-days with an annual average for two-thirds of the years.
Our payrolls indicate very little change in the way labourers were paid, their hierarchies, technologies or work descriptionsFootnote 2 . The peones (labourers) were at the bottom of the hierarchy in construction work. The guild of albañiles (masons) regulated construction work during the colonial period, yet peones did not belong to it. Guilds were formally abolished in January 1814, but they remained a guiding principle in labour relations until at least the mid 19th century (Carrera Stampa Reference Carrera Stampa1954, pp. 276-277). Guild members albañiles (masons) hired the peones at a customary (but not fixed) rate. While guild exams were expensive, peons could aspire to become low-rank masons if trained by a maestro and after passing a basic examination of their skills (92, 101). In our payrolls, less than 10 per cent of peons became masons.
The daily wage was the only compensation since there were no payments in rations. According to an anonymous construction treatise written in the late 18th century, overseers discounted the «real de comidas» for the meals that had been given during the week (Schuetz Reference Schuetz1987, p. 89), showing that whenever rations were provided it was discounted from the total salaryFootnote 3 . One real (one-eighth of a peso) represented the approximate weekly cost of food for one person in the colonial period.
It is likely that labourers lived in between the rural and urban worlds. Granados (2008) has shown that construction workers were the single most important occupational group among the indigenous people, representing 16 per cent of workers in a tributary list of 1800, compared with about 3 per cent in the whole population (Pérez Toledo and Klein Reference Pérez Toledo and Klein2004). While many lived in the traza (the Spanish city), one-third of the Indian construction workers, and perhaps many more, lived in the barrios and peripheral areas of the town where subsistence agriculture remained an important activityFootnote 4 . Moreover, we rarely found workers employed more than 150 days at any time in our period, and work tended to slow down during the harvest and sowing time. As we approach the 20th century, such a seasonal pattern becomes less evident, suggesting a greater integration of workers into urban labour markets, as the city population quadrupled between 1895 and 1930.
Given their social background and unregulated compensation, the daily pay rates of construction peones were relatively homogenous in the city and were representative of the earnings of other unskilled workers, who represented more than half of the labour force in the city (Piccato Reference Piccato2001, p. 23). Reported wage rates of construction peones and other unskilled, entry-level workers were often the same and at most diverged by 15 per cent without a systematic bias.
2.2. Prices
Our calculations of cost of living are based on the most extensive combing of prices of which we are aware of in the economic history literature of Mexico. Information was collected on nineteen items that appeared prominently in the expenditures of working-class urban households: maize, tortillas, bread, flour, wheat, frijol beans, sugar, pulque, beef, lamb, pork, ham, lard, soap, candles, tallow, charcoal, firewood and coarse cloth. We converted all prices to silver grams per liter (pulque), squared metre (cloth), million British thermal units or BTUs (charcoal and firewood) and kilograms (all the rest). We used regression analysis to sort out systematic biases of the different sources, types of transactions and seasonality.Footnote 5
For each product annual prices were calculated by year, proceeding in three stages. First, we averaged annual prices using three types of local prices: institutional, retail and wholesale. Institutional sources comprise purchases in convents, hospitals, prisons and schools. Retail prices included posturas from the colonial period, which were prices set by city authorities for a variable period of time (every 2 years in the case of the abasto of beef and lamb, annual for lard and candles, quarterly for flour and bread and daily for maize sold in the alhóndiga)Footnote 6 . More surprisingly, we found that after independence the authorities of the Federal District regulated prices (tasas de efectos) in several years (1832, 1833, 1841, 1856, 1859, 1860, 1867, 1892, 1909 and 1914). By the 1880s, the authorities collected information on retail prices in the different municipalities of the Federal District and reported them to the Secretaría de Fomento, and later to the Departmento del Trabajo and the Departamento de Estadística Nacional. Wholesale prices typically included market updates published in newspapers, and some governmental reports. We also identified scattered sales from haciendas and factories in Mexico City. We found that seasonality did not have a significant effect on price levels, hence we did not apply any seasonal corrections.
Because no single source spans the entire period, we sought to maximise the use of all available information with procedures that were simple and easy to reproduce. For this reason, we averaged the observations of the three types of sources to obtain a single annual price by product. Overlapping observations in some years (288 intersections of year-products distributed throughout the period) did not show the existence of any systematic biases except for a 5.2 per cent markup of retail relative to wholesale prices that we applied to correct the latter.
Finally, missing annual entries were imputed using the prices of correlated products and markets. Although we collected 1,237 annual average prices, we lacked local, direct prices for many years: from three-quarters of the years in the case of candles, charcoal and pulque, to less than one-quarter for maize and sugar. Whenever possible we imputed prices from their inputs (e.g. candles from tallow, charcoal from firewood and cloth from yarn), from other products that shared the same inputs (ham, lard, pork and soap), from substitutes (lamb and beef), or from markets that supplied Mexico City (e.g. cloth from Puebla, sugar from Atlacomulco) or experienced similar market conditions (bean prices from Valladolid, Michoacán, and cloth prices from Zacatecas). In the case of tortilla and bread prices, they were predicted from maize, and wheat and flour, respectively. We obtained the correction factors from years with common observations. The imputations allowed us to add 336 additional annual entries and did not change trends or the variation in the series.
Even with this broad range of sources, we lack information for more than one-third of all possible annual price observations. Data are scarcer in the mid 18th century and the mid-19th century, as indicated in Table 2. In order to fill in missing data, we followed a modified version of Drelichman’s (Reference Drelichman2005) construction of price indexes for early modern Spain: in any given year in which a product’s price was missing, we calculated a high and low band of prices corresponding to the 75th and 25th percentiles of the prices in a 10- or 20-year window. In the case of pulque and soap, products with scarce data but relatively stable prices, we used wider time windows.
2.3. The Consumer Basket and the Consumer Price Indexes
We build two Laspeyres consumer price indexes based on two consumer basketsFootnote 7 . The quantity for each item is fixed throughout the duration of the series, with the exception of rent (see Table 1). The advantage of using this type of index is that it establishes a fixed subsistence or poverty line measured in calories, heating units, clothing, etc., that allows standard comparisons across regions and time periods. The energy represented in the food component of the baskets is constant and meets minimum daily dietary requirements (1,940 calories). However, a Laspeyres index does not account for changes in quantities through time, which could be particularly relevant when analyzing a long period of time.
CONSUMER BASKETS (Rent excluded)
Sources: Allen (Reference Allen2001); Allen et al. (Reference Allen, Murphy and Schneider2012). See text for construction of Mexican baskets.
Notes: The «barebone’s» basket is entirely based on Allen et al. (Reference Allen, Murphy and Schneider2012). For the «respectable» basket, see text and online Appendix B.
The first index is based on a «barebones» or subsistence basket, similar to Allen’s basket for Milan (Allen et al. Reference Allen, Murphy and Schneider2012)Footnote 8 . It includes food products that minimise the total price per calorie and supplies minimum nutritional requirements. It also provides minimum heating, lighting and clothing needs. It includes eight items and represents the minimum cost of subsistence below which individuals could hardly survive — a concept similar to Adam Smith’s price of subsistence or Karl Marx’s cost of reproduction. The basket relies heavily on corn and beans, a consumption pattern that reflects the usual depiction of the customary Mexican diet. Yet, corn was consumed mainly in the form of tortillas, which demanded over 20 hours of weekly work per family (Bauer Reference Bauer1990). Moreover, the urban working classes consumed a remarkable amount of more expensive, but still affordable food that was adapted to the constraints and lifestyles of the city (Quiroz Reference Quiroz2005; Miño Grijalva Reference Miño Grijalva2006; Moncada Reference Moncada González2013).
Our second index better captures the complexity of the urban working-class diet, while keeping the caloric output, services and cost comparable to the «respectable» price indexes in the literature (Allen Reference Allen2001; Bassino and Ma, 2005; Cvrcek Reference Cvrcek2013). This basket includes fourteen products that were considered more appealing by urban workers. We based our basket on average expenditures observed in working-class family budgets from the early 20th century and in the expenses of a convent and school for indigenous peoples from the early 19th century discussed in the online Appendix B. The share of expenditures in our stylised «respectable» basket is roughly comparable to «respectable» budgets in Beijing and London (see Table 1 and Table B1). The basket includes bread, more allowance for beans and meat, maize tortillas, sugar and pulque (a fermented beverage that contributed more than 4 per cent of the daily calories). It does not include chili peppers, chocolate (or coffee) and salt, items that were present in the observed budgets but were not comparable with other baskets, did not provide any calories, or were untraceable over time. The food items were not necessarily the cheapest per calorie, but still add up to the minimum energy requirements (1,940 calories); clothing, lighting and fuel are 60-100 per cent higher than in the subsistence basket. We can consider this basket as a poverty, rather than a subsistence line, necessary to maintain a «respectable» but still very limited lifestyle.
In order to make the results internationally comparable, we consider rent as a constant five per cent of all other expenses in the basket, following Allen (Reference Allen2001). Our research on housing costs suggests that the rent of a room ranged from 15 to 18 per cent of other living expenses. Research in large cities, with the exception of London, indicates housing costs similar to those in Mexico City, suggesting that housing affects the comparison in levelsFootnote 9 . Like Allen (Reference Allen2001, p. 12), we justify this seemingly low allowance in the fact that many workers did not pay for housing, and also in the fact that using higher estimates would not significantly affect the trends presented in the paperFootnote 10 . Nonetheless, we consider actual rental costs that range between 15 and 18 per cent of other expenses in the elaboration of an alternative visualisation of the cost of living.
2.4. The Welfare Ratio
We calculate two welfare ratio indexes based on the two baskets described above, both estimates are based on assumptions about the unit of analysis, employment and the composition of the basket of goods and services. Research in real wages treads a thin line between being accurate to the population being analysed using reliable information and assuring that the results are comparable with other international experiences and time periods. The advantages and pitfalls of different methods and metrics are thoroughly discussed in this issue. Our approach builds on Robert Allen’s methodology that calculates real wages as welfare ratios, or the number of household consumer baskets that could be purchased with a wage-earner’s annual income (Allen Reference Allen2001; Allen et al. Reference Allen, Murphy and Schneider2015). The price and wage observations of a given year are combined with a series of multipliers that result in an annual real wage index constructed as
$${\rm Welfare}\,{\rm ratio=}\left( {{\rm dailywage }{\times}{\rm days}} \right){\rm / }\left( {{\rm basket }{\times}{\rm household}} \right)$$
The numerator (the daily wage multiplied by the number of days worked in 1 year) is the annual income estimate of the reference working-class household, typically urban construction workers. The denominator corresponds to the expenses needed to satisfy minimum consumption needs. The composition of baskets may vary across countries, but they all yield the same caloric output, fuel energy and lighting. To obtain household consumption, the basket is multiplied by the number of adult-equivalent members of the household.
Allen assumes that labourers worked 250 days per year. Although guild regulations support this number as an upper bound figure (Carrera Stampa Reference Carrera Stampa1954, pp. 163-167), the social history of this time period indicates that job insecurity and constant mobility were regular features in the work experience of the urban poor (Guerrero Reference Guerrero1901; Haslip-Viera Reference Haslip-Viera1999, p. 29). In the Vizcaínas payrolls, in which we have systematic records by name, non-eventual workers worked on average 108 days per year before 1910; after 1913, Vizcaínas relied on fewer workers working more days in the year, but still they worked an average of 165. Given that construction work is not a steady form of employment and there was a high turnover, it is likely that workers at Vizcaínas had other jobs through the year. However, if the alternative employment (e.g. subsistence agriculture) was not the equivalent to the construction wage rate, then it is likely that we are introducing an upwards bias in our calculations.
We scale all budgets to represent a nuclear family of two adults and two young children living from one adult income. The ratio of consumers to wage earners is three to one assuming that young children consume half an adult basket of goods. Some studies support this choice of household size (Pescador Reference Pescador1992, pp. 183-190; Dirección General de Estadísticas 1898). However, the single-family household with children was far from being the majority. In 1811, 38 per cent of the households of unskilled and skilled workers were of this kind, but multi-family or extended family households represented almost the same proportion (Brun Martínez Reference Brun Martínez1979; Vera Bolaños Reference Vera Bolaños2004). In Guadalajara (1821-1822) and Oaxaca (1930), households were similarly varied with the nuclear-family structure being the most frequent but not representing the majorityFootnote 11 . Moreover, control of the household size was a common economic strategy of the poor in Mexico City (Scardaville Reference Scardaville2004). These adjustments are a predicted consequence (as the welfare ratio dips below one, households need to boost income or cut costs to the extent of their possibilities), the ranges of possibilities and aggregate choices made by different societies may have varied in substantial ways.
Overall, Allen’s specifications of household size, days worked and rent values are reasonable in relation to our empirical findings, and their use is crucial to allow an international comparison of welfare ratios. However, in order to account for the greater complexity of working-class livelihoods we constructed an alternative representation of welfare: the number of days of work in a year required to buy the basket that supports a four-person annual household budget (see Figure 3).
3. GENERAL TRENDS IN PRICES, NOMINAL AND REAL WAGES
The evolution of welfare ratios follows three well-defined patterns: the erosion of high real wage levels during the last decades of the colonial period, low levels and stagnation through the mid 19th century, and partial recovery with great volatility during the Porfiriato and early 1920s. In the long run the most noticeable feature of the trend is that of stagnation. The «barebones» basket indicates that the high initial 18th-century levels were never reached again, while when using the «respectable» basket these levels were recovered — at least briefly — during the Porfiriato and during the 1920s.
Figures 1 and 2 show the high and low boundaries of annual welfare ratios of labourers using the «barebones» and the «respectable» price indexes, while Table 2 summarises prices, wages and welfare ratios by quarters of a century. The scale is the number of «barebones» or «respectable» household annual baskets (welfare ratios) that could be purchased with 250 days of labourers’ income. Both indexes show a similar wide U-shaped trend, although they differ, expectedly, in their levels and volatility and the trend of «respectable» welfare ratio is even flatter.
REAL WAGES OF LABOURERS USING THE SUBSISTENCE CONSUMER PRICE INDEX (CPI; welfare ratio)
Sources: See Appendix.
REAL WAGES OF LABOURERS USING THE RESPECTABLE CONSUMER PRICE INDEX (CPI; welfare ratio)
Sources: See Appendix.
AVERAGE PRICES AND CPI (silver grams per kilogram)
Notes: CPI: consumer price index.
The CPI values reflect the upper and lower bound in the cost of living. The price of tortillas and bread were derived from maize and wheat (or flour) as described in the text. The percentage of missing data is the missing observations relative to the total number of possible observations (e.g. 12 products×25 years=300 possible observations).
Sources: See Appendix and online Appendix A.
Our barebones price and real wage indexes are highly correlated with those of Allen et al. (Reference Allen, Murphy and Schneider2012) and Arroyo Abad et al. (Reference Arroyo Abad, Davies and Van Zanden2012), resulting in the similar trends shown in Figure 3. Yet the levels and the magnitude of changes were different. Our barebones welfare ratios are above Allen, Murphy and Schneider’s by a small difference (2.7 vs. 2.6) and above Arroyo Abad, Davis and Van Zanden’s by a larger difference (2.7 vs. 2.1). Our series has a smaller range of variation, hence showing a less pronounced, yet significant decrease in real wages from the 1750s to the 1800s. The correlation with price indexes and real wages for the Porfirian and revolutionary periods are also high and significant (see Figure C2)Footnote 12 .
COMPARISON OF REAL WAGE SERIES IN THE COLONIAL PERIOD (Average real wages of the decade)
Sources: Allen et al. (Reference Allen, Murphy and Schneider2012), Arroyo Abad et al. (Reference Arroyo Abad, Davies and Van Zanden2012) and Appendix. Notes: We thank the authors for facilitating the data at the annual level.
During the colonial period there was an increase in welfare ratios towards the 1760s, followed by a steep decline that began in the late 1770s and reached a nadir during the early 1810s. Between 1760 and 1815, real wages decreased by 40 per cent according to the respectable index and by 50 per cent according to the barebones index.
From 1750 to 1812, labourer nominal wages were pegged at three reales per day (slightly more than 9 silver g), with some spread in wage rates noticeable since the 1780s that produced tiny annual variations (lower than 2 per cent) that did not correlate with price changes (see Table 2 and Appendix Table A1 at the end of the article). From the third quarter of the 18th century to the first of the following century, prices increased by 53 per cent. While prices rose to historical highs during the insurrection, inflation was not negligible during the previous decades (average annual inflation of 2 per cent from 1775 to 1809). The trends are remarkably similar in our two price indexes and in other price indexes, indicating a high degree of certitude that prices, both in nominal and real terms, experienced a sustained decline.
Given the stability of nominal wages, the rise of prices eroded the welfare of workers since circa 1780. It was only in 1813 when nominal wages began to increase reaching four reales by 1817. The demand for labour caused by forced levies and steep price inflation during the insurrection (1810-1821) was the likely cause of the spike. However, real wages kept falling due to the disruptions in the city’s supply lines and harvest losses that increased prices to historical records. In sum, these findings reinstall the traditional notion of a significant and generalised increase in prices in the late colonial period and a decline in living standards (Florescano Reference Florescano1969; Tutino Reference Tutino1986; Van Young Reference Van Young1987; García Acosta Reference García Acosta1995) that has been the subject of recent revisionism (Quiroz Reference Quiroz2005; Silva Riquer Reference Silva Riquer2008; Dobado and García Reference Dobado and García Montero2010, Reference Dobado González and García Montero2014).
In the decades after independence, real wages rose from the 1820s to the 1840s, and declined sharply in the 1850s and 1860s. Nominal wages gradually receded from the high rates of the 1810s and stabilised around 3.25 reales (10 silver g) per day for most of the 19th century (the typical annual variation being less than 4 per cent, without a trend). After 1825 prices declined slightly, but they tended to remain higher than in the second half of the 18th century. Welfare ratios recovered ground during the 1820s, but the uncertainty of the information leads to some caution in interpreting the magnitude of the recovery. The most likely scenario is that the real wages of Mexico City remained below the high levels of the mid 18th century and comparable to those of the turn of the 19th century. This picture supports the case of secular stagnation presented by Coatsworth (Reference Coatsworth1978) and, in a more nuanced way, by Salvucci (Reference Salvucci1997). However, the alternative thesis of a strong recovery, and even improvements in living standards (Tutino, Reference Tutino1986; McCaa, Reference Mccaa1993), cannot be entirely ruled out since the upper-level estimates of the «respectable» approached the lower boundary levels of the mid 18th century. Regardless of the strength of the rebound following independence, as Sánchez Santiró has argued (2009), the intense civil war of the 1850s and its continuation into the 1860s had ruinous consequences for the economy and the welfare of workers.
Once peace was restored after 1867, welfare ratios began to increase once again although with great volatility. The recovery continued into the end of the period with pronounced swings. Nominal wages started an upward trek in the last quarter of the century and by the start of the 20th century they increased sharply to almost 5 reales (15 silver g) and beyond (see Table 2). This evolution of Mexico City’s nominal unskilled wages bears a strong relationship with similar data from other regions of the countryFootnote 13 . In the mid and late 19th century, the price trend flattened but during the early 20th century prices increased at higher rates, particularly during the most violent years of the Mexican Revolution (1914-1918) and surrounding years. From the 18th to the late 19th century, the relative prices of manufactured and processed products decreased in relation to food and charcoal prices.
For sustained periods (1868-72, 1885-91, 1904-07 and then 1922-27) real wages reached and even surpassed the highs of the colonial period, yet there were periods when real wages dipped well below those levels (1880-85, 1900-03). The Mexican Revolution (1911-1920) hit welfare ratios hard, reaching their lowest value in 1915 even if our data exaggerates the decline due to monetary anarchy.
The increase of real wages during the last decades of the 19th century and the beginning of the 20th century could be explained by the pacification of the country, its integration into the global economy, and the introduction of new technologies and forms of organisation. Still, it shows that the modernisation process was not deep or wide enough to increase welfare levels above those of the 18th century. In conclusion, the significant and real gains in productivity represented in the growth of GDP per capita only partially contributed to a greater well-being of unskilled workers in Mexico City.
Along with the characteristics of each period, war and climatic disasters are major factors behind the dramatic falls in welfare shown in the series. The deepest plunges occurred in years of both civil wars and climatic disasters: 1810-13, 1867 and 1915-18. In general, years of harsh climatic conditions and agricultural disasters (such as 1785-86, 1801-02, 1874-77, 1902) were years of low real wages that were well below average. In these periods, higher unemployment probably reduced the welfare of the working classes below what is represented in the welfare ratio. These episodes were as frequent and intense in the late 19th and early 20th century as they were in the colonial period, indicating that the Mexican economy was still subject to structural constraints in its food supply.
Throughout the whole period, with the exception of 1915, a labourer living on a barebones budget earned enough to acquire the basic requirements of a nuclear family to survive. In the years with the highest ratio (1760s and 1920s), a labourer’s income could easily afford three baskets, translating into a capacity to spend in a less austere standard of living or to work fewer days. With a «respectable» budget, the income of a labourer was just enough to stay above the poverty line, but in many years (1795-1817; 1856-67; 1896-1903; 1914-21) this was not the case. Under these conditions, households had to work more days (or increase the number of workers), lower housing costs or switch to a basket more similar to the barebones. We seek to shed light on these accommodations by presenting an alternative representation of real wages.
Figure 4 presents this alternative visualisation by describing changes in purchasing power without assuming 250 days of work per labourer and introducing variable housing costsFootnote 14 . This perspective allows a more intuitive interpretation of the absolute levels of poverty or well-being that construction workers in Mexico City faced through the period. According to these results, the number of work days required to purchase an annual household basket increased from the mid 18th (1730-1774) to the mid 19th centuries (1822-1876) and then declined. Labourers were able to afford the subsistence basket working around one-third of the year throughout the period. However, they needed to work most of the year, in some periods more than 250 days, to afford a «respectable» basket, which could be considered a poverty line. During some sub-periods this required more days of labour than the 250 days previously assumed; and this was the case in the entire period if actual housing costs are factored in.
DAYS OF WORK REQUIRED TO PURCHASE A HOUSEHOLD BASKET, 1730S-1920S
Sources: See Appendix and online Appendix B.
This representation also serves to highlight the flexible nature of the actual well-being of working-class households in Mexico City. It is plausible that labourers and their families farmed their own subsistence plots, particularly earlier in the period, in which case maintaining these levels of welfare would require fewer days of wage work. On the other hand, it is possible that labourers did not find work during all the days required, or carried out work that gave them a lower remuneration than the wages paid in construction. If this were the case their welfare would fall below the poverty line. Moreover, there were many expenses not included in these basic baskets, such as other goods (salt, chile, chocolate), rituals (baptisms, burials etc.), medical costs and amusement. Although these costs were highly variable, Figure 4 helps visualise that in order to afford such expenses a small working-class household in Mexico City had to expand the numbers of days worked by its members beyond 250 to afford this type of expenses that in many ways were essential to their well-being.
The trends in real wages correlate well with decadal averages of adult heights, an indicator of nutritional status and biological well-being (see Figure 5). Our comparison uses the long-term series of infantry heights by Challú (Reference Challú2010) and López-Alonso (Reference López-Alonso2012). These are a fraction of a broader anthropometric history literature, but they span a long time period with a relatively homogeneous population (infantry from the central region or the entire country) that encompasses a broad swath of the Mexican working classesFootnote 15 . Real wages and height had parallel trends through the 1810s, evidencing high levels in the mid 18th century, followed by a decline. From the 1820s on, the trends do not track each other; still the rebound in average heights by the 1870s corresponds in broad strokes to the rise in real wages in the Porfiriato. Heights remained flat through the 1920s, as did real wages. It would not be until the mid 20th century that material well-being improved as policies and technology made food more accessible and affordable (Ochoa Reference Ochoa2000; López-Alonso Reference López-Alonso2012). Overall, both anthropometric data and real wages stress that welfare throughout the period did not make substantial progress.
BAREBONES WELFARE RATIOS AND SOLDIER HEIGHTS IN MEXICO, 1730-1930
Sources: Challú (Reference Challú2010), López-Alonso (2012) and see Appendix and online Appendix A. Notes: Soldier heights are represented by decade of birth. The constant in the original regression reports was changed to match unskilled workers and the central region.
4. MEXICO CITY’S REAL WAGES IN INTERNATIONAL PERSPECTIVE
Today, Mexico is considered an «upper middle-income» economy in the World Bank’s classification, which in itself attests to the substantial divergence in incomes that has occurred since the end of the 19th century. Table 3 breaks down the historical trends in «respectable» welfare ratios in Europe and Asia in income level groups. Figures 5 and 6 show the «barebones» and «respectable» welfare ratios. While the selection of cities and the levels in the two figures are different, the overall picture is similar. We mostly follow the «respectable» version here as it provides a more representative gauge.
REAL WAGES IN COMPARISON, BAREBONES CPI, 1750-1930
Sources: Allen et al. (2011), Arroyo Abad (Reference Arroyo Abad2014), Arroyo Abad et al. (2011) and Appendix. Notes: The year represents the center of 25-year periods.
REAL WAGES IN SELECT CITIES RESPECTABLE CPI (welfare ratios)
Notes: CPI: consumer price index.
In most cities (but not Mexico City), the last period ends in 1913.
Sources: For Mexico, our Appendix C. For Beijing, our own calculations of Allen et al.’s (2011) respectable budget using prices distributed in the GPIH website. For Vienna, Cvrcek (Reference Cvrcek2013). For Kyoto/Tokyo, we used Bassino and Ma’s (2005) real wages index adjusted to reflect earnings over 250 days, a basket for three adults, a 5 per cent rent allowance and a 9 per cent allowance for fuel and soap (items excluded from Basino and Ma’s calculation). For Madrid, we relied on Llopis and García (Reference Llopis Agelán and García Montero2011) and Reher and Ballesteros (1993), scaling their index values to Allen’s welfare ratio for 1800 (the only common year in the three series). For all others, see Allen (Reference Allen2001).
Table 3 and Figures 6 and 7 show that from the beginning to the end of the period there were sustained gaps in real wages, although the rankings changed as some cities grew more than others. The high-wage tier included Amsterdam, London and other English cities, although they were not homogeneous in trajectory. The slide of Amsterdam’s real wages through the mid 19th century was remarkable, but well documented in the economic history of the country. Its welfare ratios rebounded in the second half of the 19th century; by the end of the period Amsterdam was below London but still experienced almost a 50 per cent growth in real wages. Mid-wage cities such as Leipzig, Paris and Warsaw experienced solid growth; Leipzig gained parity to Amsterdam by the eve of World War I. Among the group of low-wage cities, Madrid’s decline is remarkable: from levels comparable to London and Amsterdam in the 18th century, it steadily declined through the early 19th century. By the beginning of the 20th century, Vienna and Tokyo moved up in the ranking, while the other low-income cities increased their levels moderately and remained far below.
REAL WAGES IN COMPARISON, RESPECTABLE CONSUMER PRICE INDEX (CPI), 1750-1930
Sources: See Table 3. Notes: The year represents the center of 25-year periods.
In the comparison, Mexico City started out with high real wages, almost at par with Amsterdam and London. The decline through the early 19th century was above the average rate and became more comparable to the mid-income group of cities. It was in the 19th and early 20th centuries, however, when Mexico City’s mild improvement in real wages was no match to the trajectory in any other city with comparable information. Kyoto/Tokyo, Leipzig, Paris, Warsaw were poorer than Mexico City, yet by the early 20th century they had all surpassed it. The three cities in the Hispanic world in this compilation, Madrid, Mexico City and Lima, stand out as the only ones that failed to experience growth in the long term. These data add one more dimension to the scholarship highlighting the connections and resemblances of the economic history of old Spanish empire (Dobado et al. Reference Dobado, Gómez-Galvarriato and Márquez2007).
Mexico City’s trajectory indicates a «reversal of fortune», but its timing does not match well the interpretation of the coiners of the expression. If Acemoglu, Johnson and Robinson (Reference Acemoglu, Johnson and Robinson2002, 2001) focus on the colonial experience as the root of extractive and rent-seeking institutions that stifled growth, our trends indicate that Mexico City in the 18th century was closer to being a wealthy city than to other income tiers. The erosion of Mexico’s position started during the late colonial period. Mexico City was 67 per cent above middle-income cities in 1725-1749, and only 6 per cent ahead by 1800-1824 (see Table 4). The fall was more pronounced at the very end of the 19th century when, even though Mexico’s welfare ratios were growing, they fell behind middle-income cities. By the early 20th century, Mexico City’s welfare ratios became comparable to low-income cities in Europe and Asia. If a «reversal of fortune» took place it could only have been through lagged consequences that produced a failure to carry out the sort of industrialising economic growth that raised welfare ratios in several countries significantly above their 18th-century levels (Sokoloff and Engerman Reference Engerman and Sokoloff1997). In Mexico, although rates of economic growth rose after 1880s they did not translate into increasing welfare ratios in a similar way to industrialising countries.
MEXICO CITY REAL WAGES AS A PERCENTAGE OF OTHER CITIES (respectable CPI)
Notes: CPI: consumer price index.
Sources: See Table 3. The percentage results from dividing Mexico and the group average.
5. CONCLUSIONS
Mexico City’s real wages were high in an international perspective by the mid 18th century, only moderately below those of northwestern European cities. Yet, real wages hit a low point by the end of the colonial period and the decade of insurrection. The late colonial era of Bourbon reformism and mining expansion (1780s-1800s) was not a period of bonanza, but of significant decline in living standards, caused in part by the high extraction of resources by the Crown from the New Spain in order to fight wars against its European rivals (Marichal, 2007). Our real wage series using the «barebones» and «respectable» specifications, other welfare ratio series of the colonial period, and heights agree on this characterisation.
After the late colonial decline, periods of growth were all brief and followed by periods of decline. The estimates from c. 1825 to c. 1850 indicate that welfare may have recovered mid-18th-century levels, to experience a further collapse in the following civil conflicts. When peace was achieved in the late 1860s, real wages regained a higher average level punctuated by some pronounced falls and a dramatic plunge during the Mexican Revolution.
This paper provides a valuable perspective in which to place the long-held views about the Porfiriato (1877-1911) that often focus on its failure to improve general standards of living. Porfirian economic growth allowed a recovery of Mexico City’s workers’ real wages. However, real wages grew slowly and with great volatility, a finding that contrasts with relatively high rates of GDP and GDP per capita growth during this periodFootnote 16 . This could be the result of increasing inequality during the period, and a failure to translate economic growth into better living standards for the working classes, a conjecture that is in line with the historiography of the period, but that requires further research. It is also important to highlight that the upward movement in living standards did not begin during the Porfiriato but in the late 1860s, when the country finally reached an era of greater political stability. It was interrupted by the revolutionary war (1913-1920), but it quickly rebounded and continued in the 1920s.
In the long run, our results show that there was no long-term growth in real wages in Mexico City and, probably, in Mexico as a whole. The collapse of the Spanish empire and the instability and constant warfare that followed were very costly for Mexico City’s workers in terms of their living standards. During the Porfiriato (1876-1910) and in the 1920s real wages recovered ground but did not soar. The close correlation with soldier heights (an indicator of nutrition), the high relative price of food and charcoal throughout the period, and the impact of climatic crises suggest that a low level of agricultural productivity and constraints in the food supply remained a constraint on the capacity to attain higher levels of productivityFootnote 17 .
Finally, as real wages stagnated in the long run and other regions of the world grew, the Mexican economy fell behind. From having real wages moderately below high-wage northern European cities, by the beginning of the 20th century Mexico City found its place below the average middle-income group. The erosion of Mexico’s position in the world economy started in the last two decades of the colonial period yet it became sustained and pronounced after the second half of the 19th century. This outlook agrees with John Coatsworth’s thesis that identifies the 19th century (before the Porfiriato) as period of stagnation of the economy in the context of fast growth in the industrialised nations (1978, 1989). Although Mexico experienced some modernisation, industrial growth and global integration during the Porfiriato, these results show that the degree of this process was modest in relation to those experienced by these countries. In the era of the great divergence, Mexico converged, not to early-industrialisation countries, but to the rest of the world.
Supplementary material
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/S021261091500004X.
Appendix
COST OF LIVING AND REAL WAGES 1730-1930
