Regional Inequality in Spain: An Evaluation Based on Taxable Agricultural Wealth (1858–1923) Free

There has been a drastic increase in studies on regional inequality, aimed at delimiting differences in growth between areas. Numerous studies take a regional approach to analyzing economic growth in the nineteenth and twentieth centuries, focusing on the industrialization processes that promote growth and its association with the evolution of inequality. These studies emphasize the evolution of income distribution by region to determine the causes of unequal growth. This research has shown that rising regional disparities in income and the growing divergence between geographical areas were common in the early stages of development. Moreover, research shows that regional convergence typically occurs in more advanced stages of national growth and development.¹

A significant body of literature examines regional inequality in Spain, with several important studies focused on the nineteenth and twentieth centuries. Factors like industrial location, demographic location, geographic location, and geo-institutional circumstances have been reported as responsible for the regional disparities in Spain. The analysis of regional inequality also includes the agricultural sector, although such studies do not specifically refer to land taxation and its support on taxable wealth declarations, but to the privatization of communal land. Studies on the agricultural sector in the nineteenth and twentieth centuries focus, for the most part, on the differences in agricultural productivity and the industry’s participation in general economic growth. The traditional view has been to consider the sector as backward and responsible, in a sense, for its own backwardness, highlighting a slow growth of agricultural production in the nineteenth century. Labor productivity grew slowly (1.1 percent per year between 1891–1895 and 1929–1933).²

Other scholars hold a more positive view of the evolution of the Spanish agricultural sector, emphasizing slight productivity increases that began at the end of the nineteenth century and became more pronounced in the twentieth century. Modest growth in the amount of land cultivated contributed to increased productivity per hectare, largely due to improvements in land use. Technological advances were limited, but not due to farmers’ risk aversion but because the available innovations were not adapted to specific environmental settings or because those in control of production processes hindered their adoption. Although our study does not specifically cover aspects of agricultural development in Spain such as productivity or yield, the analysis of the taxation of certain wealth declarations allows us to assess its impact on regional inequality in the period under study (1858–1923).³

The agricultural sector was the source of employment for most of the population—between 66 and 57 percent—in the years 1877 and 1922 respectively. Land was the main source of wealth in a primarily agricultural economy. A focus on this source of wealth, that is, taxable agricultural wealth (crops and livestock), is the basis of our research.⁴

Taxable wealth, as declared by landowners, was the basis for the establishment the so-called “territorial contribution” tax. We use these declarations to trace the evolution of the inequality that may have derived from this taxation across Spanish provinces in the nineteenth and early twentieth centuries. The main objective of our research is to evaluate the evolution of agrarian wealth declarations, which determined land taxation in each province and to estimate regional inequality in Spain. This wealth was quantified and recorded following the Spanish tax reform of 1845. Driven by the growth of taxable wealth, authorities conducted investigations of rural and urban wealth, giving rise to the formulation and implementation of this new tax, “the tax on real property, crops, and livestock.” To establish land values accurately, our database isolates this information by excluding the total contribution from urban wealth.

The methodology used in this article consists of four stages. First, we collected the information required to construct the database. Second, we conducted a descriptive analysis of the process used to determine the taxable wealth of the Spanish agricultural sector, derived from crop and livestock farming. Third, we categorized Spanish provinces according to a series of variables explaining the agricultural wealth, including cultivated area, irrigated land, and agricultural labor force. Finally, we calculated the Gini coefficient and the Theil index for all Spanish provinces and for the defined groups of provinces, allowing us to assess regional disparities in agricultural wealth. The result is a novel perspective on the study of regional inequality based on the taxable agricultural wealth, using a variable other than the regional gdp, which is traditionally implemented in the scientific literature to measure levels of inequality.

Our research focuses on agricultural wealth based on declarations of taxable wealth by landowners and renters for the purpose of establishing land taxation in forty-five Spanish provinces. This taxation system formed part of the territorial organization planned by the Secretary of State for Public Works, Javier de Burgos, in 1833. Consequently, the fiscal division in this study coincides almost exactly with the current provincial distribution, excluding the Basque Country and Navarre, which have independent tax systems.

The period analyzed begins with the 1845 tax reform, which was crucial due to the sharp decline in state revenue following the independence of the American colonies. Two direct taxes were implemented: the tax on real property, crops, and livestock (“land tax”) and the tax on industry and trade. The former involved assessing the net income from real property derived from rural and urban buildings, as well as crop and livestock farming. Establishing this tax involved analyzing land income from previous years, determining the extension of this land, and more—necessitating the creation of a cadastre.⁵

Implementing the new tax required an assessment of agricultural wealth, which led to the introduction of the amirallamientos—declarations of land ownership and leases made by owners and tenants. Initiated in 1846, these declarations recorded the amount of land, types of crops, production levels, and their yields from previous years. The fiscal nature of these assessments led to underreporting and concealment. Aware of potential omissions, the government attempted to make their own investigations in 1879, known as comprobaciones (verifications). Tax officials revised village declarations, incorporating new, updated assessments when omissions were detected. Although these verifications provided a more accurate picture of land holdings and production, they did not lead to changes to the final tax obligations of the provinces, even in cases where major discrepancies were found.⁶

This partial investigation did not result in an update of the taxes assigned to the provinces, as evidenced by the lack of revisions to taxable wealth in the years following the investigations. Consequently, although the Geographical Institute identified instances of underreporting, these findings did not affect the official valuation of taxable wealth or its impact on the tax assessments.

There has been some debate on whether taxable wealth declarations accurately reflected the agricultural wealth in Spain. Studies across disciplines—such as economic geography and economic history—have analyzed property ownership, tenancy, and farms. The authors acknowledge discrepancies between the declarations made in each province for the construction of the amirallamientos and the actual agricultural wealth. Nevertheless, we argue for the usefulness of this source in the present study.

The declarations formed the basis of the amirallamientos, upon which a tax was established. This tax was distributed among the provinces according to the declarations of ownership and the estimated and yields according to the type of land use. In other words, the fees to be paid by each province were estimated on the basis of the profits from the land after subtracting the production costs from the income of sales. product according to the type of land use. This tax likely deepened regional inequalities as it was not a true reflection of the property holdings or production levels. The potential for arbitrary tax allocation across provinces is a key focus of our study. The nature of the source means that there is the potential for underreporting, yet it remains the only source of agricultural information for this period. Similarly, the Ensenada Cadastre—a widely used and accepted source for eighteenth-century land ownership—was also fiscal in nature and met resistance in provinces where it resulted in higher taxation.

In the case of the amirallamientos, regardless of whether they were fair, they were enforced, generating a differentiated fiscal treatment between provinces that may have hindered economic growth. Though certain biases prevent us from reconstructing an exact measure of agrarian wealth, these records do allow us to study the inequality generated by the taxation system. For this reason, they serve as the basis of our database for the period 1856–1907.⁷

The data used from 1907 onward are more accurate. A 1906 law established a parcel cadastre to replace the amirallamientos, allowing for a fairer distribution of the tax burden. A cadastral commission, operating under the ministry of finance, was appointed to compile and verify tax registers. The new system enabled the direct calculation of the liquid yield of crop and livestock farming, offering a closer approximation to individual wealth. In 1913, the Regulations for the Cadastral Advance were approved for the registry analysis derived from the previous law, making implementation stricter than previous reforms. Initially, graphic and literal data, such as property ownership, crop types, and land value, were recorded. Taxable wealth was then estimated by setting evaluation rates using the extreme values for each crop and the gradations corresponding to the intermediate values. A similar method was used for livestock farming, considering factors such as the labor capacity of each working animal and pasture production.⁸

The data used in our study for the final period (1915 onward) are based on this system of measurement, which considers a slight variation in the tax base. Records are available until 1923, after which new parceling procedures were implemented.

In short, taxable agricultural wealth was the reference for collecting the land tax, which was the most significant revenue source for the Public Treasury. Its importance is evident in its contribution to 24.45 percent of the expected income in the 1845 budget—far exceeding the 14 percent generated by consumption taxes, which were traditionally considered more significant.

To study the evolution of regional inequality in the Spanish agricultural sector, we examine the value of taxable agricultural wealth, on which the land tax was established. This tax applied to both urban immovable assets and agricultural wealth, encompassing cultivated and uncultivated land, recreational land, fallow land, urban and rural constructions, asset censuses, and saltpans. To focus on agricultural wealth, data have been disaggregated, distinguishing urban immovable assets from agricultural wealth. Only the latter is considered in this study.

The data were obtained from the collection of documents belonging to the Instituto Nacional de Estadística, specifically the geographical and statistical reports from the annual yearbooks, for all the years included in our research. In the first period, the data correspond to the years 1858–1860, 1877, 1879, 1884, 1885, 1902, and 1907. The amirallamientos and their audits are the basis of the documentation. The process of compiling the declarations involved an opening inscription (including municipality, judicial district, province, and year), a valuation report (prepared by the provincial tax administration and the town council’s team of surveyors), an alphabetical list of owners, a description of the properties and taxable income, and a summary of the village wealth. These declarations were then used to build a database on annual agricultural wealth, with values originally recorded in reales converted to pesetas for the years 1858–1860.⁹

Inaccuracies were indeed present in the declarations. However, for the objectives of this article, the key point is not that these inaccuracies undermine the value of the inequality, as it does not affect the absolute value of the declarations, but rather how they influence the provincial distribution of the declarations.

To verify the reliability of the distribution of the amirallamientos, we examined the declarations of agricultural taxable wealth in 1879 (the only year with available verification data) together with the verifications made by official experts at the Geographic Institute. Based on these verifications, we confirmed that the distribution of the national total by provinces (average values for all years analyzed) does not differ from the average distribution of the declarations in the years studied (see Table 1). Despite the potential for over- or undervaluations, their impact on the distribution of wealth at the provincial level does not invalidate the use of the available source, the amirallamientos.

From these valuations, we can deduce that the state collected 30 percent less than it could have due to the downward valuations of provincial taxable wealth. However, the relative weight of each province is similar, suggesting that if the taxable wealth had been corrected provinces would have contributed more, but without any observable differences in their percentage share in the overall amount. In short, the overall variation in the amounts allocated to each province was an average of just 0.00005. Notably, some provinces, such as Madrid and Barcelona, would have contributed 1.5 percent more, while others, such as León and Murcia, would have contributed 1.3 percent less. In other words, if the assessments of agricultural taxable wealth had been applied, the variations by province would have been minimal. This finding suggests that the taxes levied on such wealth would have been similar across provinces and therefore do not significantly affect our study of regional agrarian inequality.

The taxable agricultural wealth was calculated differently for the years following the reform implemented in 1913. As determinants, we included the mean values of agricultural production and the mean income from livestock production. The modification of the tax base required the creation of a different database for the years 1915–1923. This modification complicates the transversal evaluation of the agricultural wealth after 1850, but it does not alter the analysis of regional inequality. The data come from the 1915 Annual of the Instituto Nacional de Estadística for the first year after the reform and from the “Memoria de la Gestión” of the General Directorate of Property and Land Tax (1928) for the remaining years. The data available for 1915 were incomplete for the provinces of Albacete, Cádiz, Ciudad Real, Córdoba, Jaén, Madrid and Toledo. We estimated the values by applying the growth in the valuation of taxable income between 1918 and 1923, which remained consistent across these provinces. The annual growth rate was applied to the 1915 data only in the indicated provinces. For the remaining provinces, the wealth valuations were taken from the General Directorate.

Finally, we took into consideration the variations in the price indices over the period under study. Using prices from 1913 (see Table 2), we adjusted the agricultural wealth in constant pesetas according to the agricultural population in each province to obtain the per capita agricultural wealth.¹⁰

To calculate per capita agricultural wealth and consider it as an explanatory variable for wealth, data on the agricultural labor force in each province, as shown in the censuses for 1860, 1877, 1887, 1900, 1910, 1920, and 1930, have been incorporated. The information on the agricultural workforce varies in each census. In 1860, the data includes owners, lessees, and farm laborers. The 1877 census includes data on agriculture, livestock, and land ownership, distinguishing male and female populations. The result includes variations in values compared to the previous census in some provinces (particularly Canary Islands, Lleida, Oviedo, and Seville). The 1887 and 1930 censuses do not report these values but include a combined figure for the population working in agricultural and farming-related activities. For the other years, the agricultural labor force was calculated based on annual variations. Table 3 shows the evolution of annual agricultural wealth according to the labor force.¹¹

Developing the database for the agricultural labor force presented difficulties, particularly due to inconsistencies in how the female population was recorded across different censuses and provinces. In some cases, the count of women working in agriculture was recorded under “female agricultural workers,” and in other instances, they were included in the “family members” group, which was composed only of women. The enormous variations between neighboring years (such as the censuses for 1860 and 1877, or 1910 and 1920, for example) suggest that differences in data registration methods may be responsible for discrepancies. We reconstructed the female agricultural labor force wherever possible by projecting the proportion of women employed in agriculture from nearby censuses, as well as accounting for the annual growth variation between them.¹²

To characterize the agricultural wealth in each province, we also considered the variables of cultivated land and hectares under irrigation. These data are from the 1858 Annual’s statement of agricultural land, the geographical and statistical review of 1885, the data provided by the Agricultural Commission for the elaboration of the documents “Crop Farming” and “Livestock Farming: Meadows and Pastures” from 1904 and the Agricultural Commission’s 1923 report.

The cultivated area played a decisive role in the development of the agricultural sector from 1858, when the farming industry was yet to be modernized. Demands for more land had been frequently voiced for several years and responding to this need was not a simple task. Disentailment was partly a response to those calls for more land for crop growing, whether it met expectations or not. The area dedicated to agriculture expanded from the middle of the nineteenth century, responding to the increased demand for land, which was compatible with the growth in population. We observe this increase—around 20 percent—from 1860 to the end of the century. To determine the availability of land, we considered both agricultural land and areas dedicated to meadows, pastureland, and hills.

Finally, hectares under irrigation could have positively affected agricultural wealth because such land facilitated increased productivity. From 1891 to 1931, land productivity grew due to the modernization of agriculture, with changes related to more modern techniques and the use of chemical fertilizers. In general, productivity growth is associated with technical progress, such as labor intensification, substitution of unprofitable crops, seed selection, use of organic fertilizers, crop rotation, and the adoption of moldboard ploughs. This period also saw a greater combination of production factors.¹³

For this reason, we attempted to ascertain whether the gradual expansion of irrigation that began at the end of the nineteenth century was one of the factors contributing to improved land productivity. The literature holds that the slow growth of irrigation could have reduced productivity growth. Land productivity did, in fact, increase, albeit slowly. However, it was not until the 1920s that the state intervened in the form of more intensive water policies. Our estimations show a 37.7 percent increase in irrigated land, mostly in the twentieth century, with significant differences between provinces.¹⁴

The methodology applied to our database to achieve our objectives may be divided into the following phases. First, to analyze the previously defined characterizing variables, we used a multivariate cluster analysis to stratify and classify the Spanish provinces into homogenous groups based on our variables.

The grouping method chosen was the “K-Means” partitioning algorithm, which is easily applied and commonly used in scientific research. The Euclidian distance was the chosen measure of association to estimate the dissimilarity between provinces. We developed four groups, and both the elbow plot and the Calinski-Harabasz pseudo-F index for the grouping by cultivated land, irrigated land, and agricultural labor force (31.09) confirmed that four groups were appropriate. Table 4 shows the composition of each group based on the control variables.

We calculated the measures of regional equality traditionally used in economic literature, considering all the Spanish provinces and the clusters defined in the previous phase. The first measure was the Gini coefficient, which quantifies the degree of deviation between the wealth distribution of members of a group (in our case, the provinces of Spain) and an equitable distribution. Under this definition, the Gini coefficient ranges from 0 to 1, where 0 represents a perfectly equal distribution and 1 represents a perfectly unequal distribution (where all wealth is held by only one province). The second measure calculated was the Theil Index, a generalized entropy index that measures “the divergence between the share of income and the share of population and weights those divergences in proportion to income.” A value of 0 indicates a perfect distribution of wealth, and higher values reflect greater wealth inequality in the population.¹⁵

The importance of the Theil Index for this study lies in its property of additive composition, which is uncommon in measures of economic inequality. Based on the groups of provinces created in the second phase of the empirical analysis, we can determine whether changes in regional inequality are the result of disparities between the groups or by inequalities within them.

According to the database, agricultural taxable wealth in Spain grew in absolute terms throughout the period of analysis, although the growth was more modest in the first period (14 percent) than in the second (42 percent). The evolution of agricultural wealth, estimated from the lands declared and subject to contributions, reflects the values provided in the declarations and served as the basis for calculating land tax payments. Although there were checks in the middle of the period (1879) that were higher in general terms (except in the provinces of Barcelona and Madrid), they were not used for their intended purpose, which was to establish the land tax. As a result, we evaluated inequality based on the valuations that applied the tax rates.

After adjusting for inflation, the increasing trend in agricultural taxable wealth should be viewed with more nuance. In this case, there was a slight decrease of 1.98 percent in the early years analyzed (1858–1885). Thereafter, there were alternating years of increase and decrease, with an average increase of 3.56 percent. This increase was more significant and continuous after the regulation reforms of 1913, which introduced new assessment systems that resulted in the application of new tax rates. This reform impacted the way that taxable wealth was assessed and collected, making it impossible to establish a strict linear comparison over time.

The analysis of the per capita data confirms this trend. During the first period (1858–1907) there was a reduction in growth, with a decline of 19.31 percent; in the second period (1915–1923), there was an improvement, with an average growth of 117.6 percent. The general decline in the first period conceals the fact that there were provinces in which there was greater wealth, such as Lleida, Gerona, Huesca, Guadalajara, and Segovia. We use agricultural wealth per capita to account for the sector’s active workforce and its significance. The second period shows a positive trend across all provinces. Notably, the Levantine provinces (Castellon, Alicante, and Murcia) experienced particularly strong increases. The increase in wealth per capita during this period likely reflects improvements in agricultural productivity.

The period under study shows a consistent expansion in productive land, with a notable surge occurring in the nineteenth century until 1915, resulting in a remarkable increase of 14 percent in land coverage. However, in the years between 1860 and 1885 there was no new land to cultivate. The repercussion on the expansion of agricultural land of the 1855 Madoz Law—when land was confiscated to raise state revenue—was largely ineffective, it being widely considered that property ownership did not result in greater opportunities for workers. The Madoz Law did not necessarily expand agricultural land effectively or improve opportunities for rural workers. In many cases, land was acquired by the bourgeoisie and large landowners rather than being redistributed to peasants. This law was a result of a need to increase public revenues by increasing public debt, not land reform.

The slow growth of Spanish agriculture in the nineteenth century can be attributed to environmental and institutional factors. Environmental conditions, such as aridity and poor soil quality, made it difficult to achieve good yields. Furthermore, institutional barriers, such as rigid land distribution and limited possibilities for incorporating new land or developing irrigation, hindered agricultural progress. The unbalanced distribution of agricultural property in Spain contributed to slow growth due to the lack of incentives for mechanization, both because of the existence of small properties with little capacity to adopt innovations and because of the large properties whose owners opted for the use of abundant and cheap labor.¹⁶

Other factors, such as a poorly functioning credit market and the badly designed nature of agricultural contracts, may have played a role, preventing improvements in output and productivity. The agricultural labor force, on the other hand, grew in the nineteenth century, particularly in the first years analyzed. Yet, after 1900, there was a decline of 12.5 percent in the agricultural labor force, particularly in provinces with traditionally strong agricultural sectors such as Galicia, Asturias, and Santander. This decline was partly linked to urbanization, with provinces like Madrid and Barcelona seeing a reduction in their agricultural populations due in part to the growth of the service sector. Conversely Castilla-La Mancha and other regions saw increases in their agricultural labor force.¹⁷

From 1885, the sluggish growth in the Spanish agricultural sector was related to a resistance to change, including slow progress in irrigation implementation and a lack of incentives: economic (lack of access to credit for small and medium farmers or scarce public investment in agricultural infrastructure), technology and innovation (little dissemination of new agricultural techniques and machinery), structural and political (lack of agrarian reform policies that would promote a more equitable distribution of land). The sector faced strong competition from abroad due to declining maritime transport costs, and, in the absence of genuine modernization, the sector demanded protectionist measures. Irrigation was not always implemented with the most productive crops, being used on many occasions for traditional products such as cereals, with their limited growth yields. In any case, there was little public support for irrigation as a means of increasing farming productivity. Between 1860 and 1885, irrigated land increased by only 1.07 percent—about 10,000 hectares—mainly in the provinces of the Canary Islands, Badajoz, Albacete, Cádiz, and Madrid. This small expansion was added to an already low base, with limited hectares of irrigated land overall.

The provinces with a tradition of irrigation, such as Valencia and Murcia, maintained their level of irrigation. In contrast, the Catalan provinces were marked by a decrease in cultivated area, especially in the early years of the twentieth century. But there was a spectacular 352 percent increase in investment in irrigation coinciding with a decline in the agricultural workforce, which should have resulted in an increase in agricultural productivity. The provinces of Catalonia also saw trends toward agricultural specialization, particularly in wine production, and expansion overall in 1852.¹⁸

These initiatives were driven by the private sector rather than public investment, which explains the slower growth. Public sector efforts to promote irrigation took longer to materialize, though support for such measures was notably more fervent. Joaquín Costa had advocated for irrigation development since the beginning of the century, and the initiatives were drawn up by Rafael Gasset in the 1902 General Plan for Hydraulic Engineering, although that plan was not adopted by law until 1911. In the case of the Galician provinces, the land devoted to crops declined while livestock farming expanded. Galicia’s agriculture sector became specialized in cattle farming, which was further boosted by the possibility of a railway and the coming of a canning industry.¹⁹

To summarize, in the early years under study, taxable wealth at 1913 prices increased in both periods, although the rise was significantly greater after the 1915 wealth assessment reform. However, some provinces experienced declines. Before the 1915 reform, notable changes occurred in Barcelona, Tarragona, and the Galician provinces. The limited expansion of cultivated land and localized implementation of irrigation hindered the agricultural sector’s dynamism.²⁰

Between 1915 and 1923, taxable agricultural wealth, measured in constant pesetas, grew by 99 percent. When calculated per capita, considering the agricultural labor force, the increase is more pronounced at 117.6 percent. Although differences in the tax collection methods impacted land evaluations in both periods, the ascending and descending trend that characterizes them is not related to this procedure. The growth in taxable wealth coincided with a 4.08 percent expansion in cultivated land, the introduction of new crops and livestock, and the expansion of irrigation. Productivity also increased, as the labor force in agriculture declined by 8.3 percent. External demand, stimulated by World War I, contributed to the development of the sector. When considering wealth per capita, the most significant increase occurred in the provinces of Alicante, Castellon, and Murcia despite a decline in the agricultural workforce in those provinces. But, generally, the increase in taxable wealth per capita was a widespread trend during this period.

The expansion of productive land was accentuated during the period 1913–1923, especially in irrigated areas, continuing a trend that had begun at the beginning of the century. But growth across the provinces was uneven. Irrigated land increased across all of Catalonia, as well as Castellon, Castilla-La Mancha (Toledo, Ciudad Real, and Cuenca), Valladolid, Huesca, and even Madrid, with some areas doubling their irrigated land compared to the beginning of the period. This expansion contributed to greater agricultural wealth. The establishment and maintenance of these newly irrigated areas must have positively affected both land and labor productivity. In other provinces, such as Murcia, Granada, Valencia, and Zaragoza, the percentage increase in irrigated crops was limited, but they had already achieved high levels of irrigated land by the middle of the nineteenth century.

The growth in agricultural output appears to be the main consequence of more intensive use of the land. Structural changes in agricultural output, marked by a trend in reducing the percentage dedicated to bread cereals and increasing that for livestock farming, fresh vegetables, and industrial crops, contributed to improved performance. Nonetheless, irrigation did not always guarantee increased productivity. In some provinces, such as A Coruna and the agrarian regions of León, Palencia, and Soria, the irrigated area declined.²¹

The evolution of the farming labor force is crucial in understanding the provincial distribution of agricultural wealth. The decline in this workforce intensified during the second period, with the active agricultural population shrinking by almost 12 percent between 1910 and 1923. The greatest reductions were in Madrid and Barcelona, where urban population and labor shifts to the secondary and tertiary sectors led to declines of 18.46 percent and 25.87 percent, respectively. Other provinces also lost significant agricultural populations, including the Canary Islands, Pontevedra, Lugo, Oviedo, and Santander. Apart from provincial particularities, there was a growth in the population employed in non-primary sectors, with a decrease in the agricultural labor force by more than 8 percent overall. This shift in the prominence of the productive sectors was greater in Barcelona and Madrid.

In short, the evolution of total taxable wealth in constant pesetas remained stable until the early twentieth century (1907) but grew following the implementation of the 1913 regulations. This trend was accompanied by an expansion of productive land from the 1880s. In the early years, the increase in irrigation failed in making the new productive land more efficient, either due to its concentration in specific provinces or its use for traditional crops. As a result, many early initiatives were abandoned, with only those launched after the 1907 tax reform continuing.

At the same time, the agricultural labor force evolved in the same direction, forming an inverted U-shaped curve where an initial period of growth was followed by a reduction, although the declines in this variable began much earlier, around 1885. This shift made labor productivity gains evident. This effect is further confirmed by per capita wealth, which initially declined and then grew rapidly after 1915.

Applying the Gini coefficient and Theil index to per capita wealth reveals both the extent of inequality across provinces between 1858 and 1923 and the change in the inequality within each of the groups of provinces defined in our empirical analysis (see Fig. 1).

Both the Gini coefficient (Fig. 2) and the Theil index (Fig. 3) reveal two different patterns of inequality change between provinces in the period under study. This difference is associated with the change in the assessment of agricultural wealth. From 1858 to 1907, provincial inequality remained stable. The level of inequity during this time was relatively low, with the Gini coefficient values ranging from 0.18 to 0.19 and Theil index values between 0.05 and 0.06. Yet, within this period of stability, the largest reductions were in 1858–1859 and 1877–1879, with annual declines of 5.9 percent and 1.3 percent in the Gini coefficient and 11.1 percent and 2.1 percent in the Theil index. This stability of inequality suggests that land disentailment did not significantly impact economic growth in any province to the extent that it altered interprovincial disparities. Neither the nineteenth-century crisis nor the competition from overseas agricultural products had a decisive effect on inequality between provinces. The protectionist policy on cereals mitigated the impact, preventing disparities between provinces, even though there was no economic growth.

The second stage of the study period (1907–1923) was characterized by more pronounced fluctuations in inequality. Provincial inequality reached slightly higher levels than in the first stage, with the Gini index rising to 0.23 and the Theil index to approximately 0.08 by 1915. Between 1907 and 1915, inequality increased by 1.7 percent (Gini) and 3.1 percent (Theil), but at the end of the World War I, provincial agrarian wealth declined, leading to a more equitable distribution of provincial agrarian wealth, though pre-war levels of equality were not fully restored. The reduction in inequality continued until 1923, with the Gini index falling to 0.21 and the Theil index to 0.07.

Clustering the Spanish provinces based on the selected control variables—cultivated land, irrigated land, and agricultural labor force—reveals, on the one hand, the disparities in per capita agricultural wealth between provinces with similar characteristics and, on the other, the origin of the inequality in the defined groups. Regarding the source of inequality, the decomposition property of the Theil index, applied to these clusters of provinces, shows that disparities within each group were the primary driver of overall inequality. This disparity accounted for 59 percent of the Theil index in 1885 and 66 percent in 1958 (see Fig. 4).

An analysis of inequality in each group reveals differences in the levels of inequity between the four groups over the study period. In the first group of provinces, inequality remained constant between 1858 and 1907. After 1907, changes in inequality were more pronounced compared to the overall set of provinces. Thus, the annual reduction of the Gini index by 4.3 percent (1916–1918) and then 1.5 percent (1918–1923), reaching 0.17, made this group of provinces the one with the greatest provincial disparity.

The first group (see Fig. 1) was considerably heterogeneous, with an average area of agricultural production around 600,000 hectares. The agricultural labor force in this group was the smallest in the country. The differences between provinces generate wealth disparities, with inequality peaking in 1915. The war disrupted foreign trade, weakening irrigated agriculture in the provinces of Levante (Alicante and Castellon) and Catalonia (Barcelona, Girona, and Tarragona). The same occurred in other export-oriented provinces such as Almeria and Logroño.

The development of the agricultural sector in Catalonia, particularly its wine specialization at the end of the nineteenth century, was essential for broader economic development. The reduction of irrigated land in inland provinces such as Avila, Burgos, Segovia, Soria, Palencia, Salamanca, and Soria demonstrate the difficulties for the internal wheat trade. Despite the protectionist measures enacted by the 1891 tariff, competition from imported wheat displaced wheat from the plateau to the coast. The agrarian depression reached the interior of the country, with the fall in production, prices, and land income severely affecting agrarian workers, many of whom emigrated to America.²²

In the twentieth century, the reduction of irrigable areas continued, particularly in the inland provinces, where the loss amounted to nearly 50 percent. Loss of the active agricultural population was also common in all provinces, showing the low productivity of the land. A slight recovery after 1915, was accompanied by changes in composition of the supply of agricultural products (with more intensive crops like beets and more livestock) and the use of moldboard plows, winnowers, and mowers. The increase in protectionism, reinforced in 1915 by World War I, led to a significant expansion of cultivated hectares and, therefore, production. Avila, Burgos, Salamanca, Segovia, Soria, and Valladolid saw more than 50 percent of previously abandoned land being cultivated again. The years of the World War I came with new opportunities to export agricultural production.

Madrid is typically not included in studies on agriculture as it is the economic capital of the country, and studies tend to focus on its significant increase in the service sector. However, it cannot be ignored that, despite the considerable reduction of the farming labor force (27 percent lower in 1923 compared to 1860), the land in Madrid included irrigated areas with a mean size of 34,000 hectares. Although the greatest growth in irrigated land occurred in the early years (until 1885), even during the expansion of 1915–1923 irrigated land increased by 6 percent. More land was not incorporated into the sector, but the existing land was more productive.

The first group includes the island provinces, the Balearic and the Canary Islands. The agricultural sector in the Balearic Islands remained constant, with an increase in land area only occurring in the later years (1915–1923), along with the similar progress in increasing irrigable area. There was, however, a significant decline in the agricultural labor force, with a 22 percent reduction, despite an overall increase in the working population. A similar trend was observed in the Canary Islands, where the introduction of new crops led to a combination of periods of growth and periods of emigration. These factors explain the inequality between provinces in this group after 1915.

The second group of provinces exhibited a distinct pattern of inequality over time, characterized by a downward trend between 1858 and 1907. This trend of inequality reduction was only interrupted between 1859 and 1860 (Gini: 5.9 percent increase; Theil: 0.6 percent increase) and between 1884 and 1902 (Gini: 11.5 percent increase; Theil: 1.0 percent increase). This group includes the provinces of the southern plateau, La Mancha, Extremadura, and parts of Andalusia (Cordoba, Seville, and Jaen). These regions had a land ownership structure linked to large parcels of land dedicated to livestock or cereal production. The variety in the agricultural product was due more to greater land fallowing than to technical innovation. For example, in Extremadura, both livestock and crop areas were expanded, but there was little irrigated land. A significant increase in the active agricultural labor force resulted in lower productivity levels than in other groups where there was greater increase of crops. In La Mancha, the climate and limited government action impeded greater progress in the sector. Traditional crops and the extent of fallow land both declined, while livestock farming was common. As a result, physical yields in the region were lower than the national average in the first third of the twentieth century.²³

The 1913 reform in the assessment of agricultural wealth led to a noticeable rise in inequality within the second group, with the Gini index increasing by 4.2 percent and the Theil index by 8.4 percent annually between 1907 and 1915. By 1923, these indices reached values of 0.15 and 0.04, respectively. The differences between them were driven by the specialization made possible by foreign market connections made in response to the economic demand during World War I. Provinces like Badajoz and Teruel expanded their cultivated land, while the others capitalized on existing farmland by increasing labor inputs. Other provinces, particularly in Andalusia (Córdoba, Jaen, and Seville) and León, displayed less dynamism. After the war, inequality persisted as certain provinces, such as León, saw cultivated land abandoned; in contrast, provinces such as Teruel increased their land under irrigation by 15 percent.

The third group of provinces exhibited consistently low levels of inequality throughout the period analyzed. Its evolution is marked by two periods of pronounced change. The first (1860–1885), was characterized by large consecutive annual increases and decreases in inequality, with provincial inequality reaching values of approximately 0.14 (Gini) and 0.03 (Theil). Inequality then declined until 1918. The second period (1918–1923) saw a more equitable distribution of provincial agrarian wealth driven by an annual decrease of 7.5 percent in the Gini coefficient and 14.6 percent in the Theil index.

In the third group, the provinces with more than 100,000 hectares of irrigated land—Lleida, Valencia, and Zaragoza—stood out due to their high agricultural yields. The fertile Granada plain, reoriented to the production of tropical crops on its more than 60,000 hectares of irrigated land, justifies the province’s presence in this group. In the second half of the nineteenth century, Zaragoza consolidated its position as a center of cereal production, linked to the economic center of Barcelona, whose supply it provided. The rail connection to that city facilitated trade. Although the province faced competition from external markets at the end of the nineteenth century, new tariffs permitted specialization in agriculture. By 1923, Zaragoza had increased its productive area and the areas of irrigated land to 120,000 hectares, shifting toward new crops, such as beet and alfalfa, with improved productivity. The agricultural labor force, which had been growing until 1915, declined by 7.07 percent in line with the century’s demographical modernization.²⁴

There was a strong presence of agriculture in Valencia, but at the end of the nineteenth century, cereals were still the dominant crop despite the higher profitability of oranges. The transition from the cultivation of cereals was complicated due, on the one hand, to the structure of agricultural property—mostly small to medium-sized properties and leased land—and, on the other hand, to the high capital costs required. Although the evolution of the inequality indicators in this group mirrored those of Spain as a whole, disparities between provinces grew at the fastest annual rates after 1918, corresponding to the end of World War I. This divergence between provinces in these years coincided with the increase of irrigated land in Lleida, while other provinces in the group experienced slower growth in irrigation.²⁵

Finally, a fourth group of provinces—comprising Galicia and Asturias—experienced three distinct phases in the distribution of per capita agricultural wealth. Between 1860 and 1877, inequality declined annually by 2.5 percent (Gini) and 5.4 percent (Theil), reducing the Gini coefficient from 0.15 to 0.10. This level remained constant until 1885 but then increased at an annual rate of 2.2 percent, reaching 0.14 (Gini) and 0.03 (Theil) in 1902. Despite a reduction in irrigated land in this period, the impact on wealth was limited due to the region’s climatic characteristics. Instead, progress was likely constrained by the structure of agricultural property, which was dominated by smallholdings.

After 1907, continuous reductions in inequality resulted in this group achieving the most equitable distribution of wealth among the four, with Gini and Theil values close to 0.07 and 0.01, respectively. This fact is not linked to economic growth, however. In fact, since 1915, land abandonment was widespread in all Galician provinces. The expansive process of land clearing due to World War I did not affect these areas. Except in Pontevedra, irrigated areas were also reduced and the active agricultural population shrank, except in A Coruña.

In short, provincial inequality in terms of taxable agricultural wealth remained in general terms constant until 1907, when the distancing between them began to increase. A key event was World War I, which linked agricultural activity more closely to foreign markets, which in turn differentiated the provinces.

Our empirical approach has allowed us to define and estimate Spanish taxable agricultural wealth in the nineteenth and twentieth centuries and to determine the level of inequality between provinces based on this agricultural wealth variable. Achieving these two objectives represents a novel contribution to the literature. First, estimating taxable agricultural wealth as a measure of inequality in Spain’s predominantly agrarian society provides more approximate evidence of the concentration of agrarian wealth in a few provinces. This concentration shaped disparities in agrarian growth between Spanish provinces, particularly at the beginning of the twentieth century. Second, the use of taxable agricultural wealth as a variable to measure inequality is innovative; other studies on inequality typically analyze different regional growth based on migratory movements or industrial development.

Taxable agricultural wealth per capita declined progressively throughout the first years of our analysis (until 1907). The amount of land devoted to agriculture and livestock increased slowly (1860–1885), with significant employment in the sector. The economic crisis at the end of the century pushed people progressively away from rural areas. Irrigation was implemented slowly during this period and not always successfully, so that some initiatives were abandoned. The subsequent implementation of the 1907 reform in 1913 not only brought tax rates up to date with the new assessment of agricultural wealth but was the beginning of a period in which increases in assessed agricultural wealth were continuous until at least 1923. In the early years of the twentieth century, increased investments and successful irrigation projects, along with the specialization and reorientation of production in agricultural areas (with a greater role for livestock, for example), stressed the differences between provinces, particularly in the provinces of Catalonia and the eastern coast. The result was a reduction in the labor force and better agricultural performance and productivity.