A conflicted political will to levy local taxes: inequality and local school politics in Sweden, 1840–1900

ABSTRACT Through an extensive study of 12 parishes in the Sundsvall region, this article, informed by studies in the economic history of education, examines changes and continuities in local school politics during the period of 1840–1900. Using the Sundsvall region in the northern part of Sweden as its point of departure, this article shows how basic political conflicts shifted when political franchise, tax regulations and the social structure of the region changed during the second half of the nineteenth century. At the end of the investigated period, the basic conflict of school politics was no longer between those who owned land and those who did not but rather between high- and low-income groups. Judging from local school politics, the local elites of the Sundsvall region, in contrast to local elites in the USA, England, Spain and Prussia, focused their attention on school funding. The main conflicts between the social groups not only concerned the distribution of school expenditures but also included issues, such as the location of schools.


Introduction
Because the expansion of schooling in the nineteenth and early 20th centuries has been regarded as vital for economic development, numerous studies have been devoted to analysing why schools developed more rapidly in some nations and regions than in others. According to an influential hypothesis promoted by the economic historian Peter Lindert, the decentralized nature of many nineteenth century school systems enabled their expansion. Local school districts were allowed to decide their expenditure levels on their own, enabling decisions on investments in schooling that would have been impossible on a national level. Thus, Lindert argues that schooling was driven less by central government intervention than by 'a spontaneous political will to levy local taxes in thousands of school districts' (Lindert, 2004;p. 121. See also Beadie, 2010;Go & Lindert, 2010).
This spontaneous political will was nevertheless affected by various factors. This article contributes to the growing literature on the relation between schooling and the distribution of wealth, land and political influence. Several studies have noted the negative relationship between inequality and the development of schooling (see, e.g. Engerman, Mariscal, & Sokoloff, 2009;Gallego, 2010;Galor, Moav, & Vollrath, 2009) and have shown how, for instance, landholding elites have blocked the expansion of schooling because they wanted to preserve the existing social order and reduce the mobility of their workforce (see Beltrán Tapia, 2013, and literature cited therein). The negative impact of economic and political inequality has been indicated by quantitative analyses of the distribution of political voice in the US prior to 1850 (Go & Lindert, 2010), the distribution of land in mid-19th-century Spain, the distribution of land in England and Wales in the late 19th century (Beltrán Tapia & Martinez-Galarraga, 2015;Goni, 2013), and landownership concentration in nineteenth-century Prussia (Cinnirella & Hornung, 2016).
However, other findings contradict those results. For example, it is well known that elites have promoted mass education as a means of social control and of fostering a national identity (see, e.g. Green, 1990;Petterson, 1992;Weber, 1976). There are also quantitative analyses that question the scope of the conclusions presented above. For example, Cappelli (2016) has shown that the distribution of political voice in Italian municipalities was not an important factor in the regional inequalities of schooling; Cvrcek and Zajicek (2013) have shown how the elites of the Habsburg Empire did not always oppose mass schooling but instead promoted schooling under certain conditions. In the case of 10 counties in southern Sweden, Andersson and Berger (2016) found that school districts dominated by local elitesthat is, districts with a high concentration of votes in the hands of a few individualsindeed had higher expenditure levels than comparatively egalitarian school districts.
The significance attributed to the political will of school districts, in combination with the differing findings on the impact of inequality on schooling, raises important questions regarding the social and political conflicts of local school districts. The purpose of this article is to examine these conflicts in local school politics in 19th-century Sweden. What questions were central to local political debate? Between which individuals and groups did the main conflicts occur? Did the politically dominant social strata oppose or promote the expansion of schooling in the governing bodies of the school districts? To address these questions, this article presents an extensive study of 12 rural school districts in the Sundsvall region of northern Sweden, focusing on the period of 1840-1900. The primary data sources used in the analysis are minutes from school boards, parish meetings and church meetings, which were the main bodies that governed the school districts.
This investigation of local school politics is well suited to contribute to the research on inequality and schooling. Although economic and political inequality have been recognized as significant determinants of schooling, there is, as noted by Go and Lindert (2010), still a need for further studies of local decision-making processes. There are several likely reasons for this research gap. The growth of schooling has often been perceived as a result of state intervention through school acts, state school inspectors and state subsidies (Grew, Harrigan, & Whitney, 1983), and the main focus has consequently been placed on national politics. For a long time, the historical sciences also primarily attended to national policy rather than local politics (Nydahl, 2010). As a result, questions regarding how inequality affected local school politics and, by extension, the rise of mass schooling remain unanswered.

Changes in society, politics and schooling
In the 19th century, Sweden was a distinctly agrarian and rural society. In 1850, less than 10% of the population lived in towns, and in 1900, two-thirds of the population still lived in the countryside. However, Sweden was undergoing rapid social and economic development. Due to the agricultural revolution and the increase in population and production that this process entailed, social differentiation increased. The number of landholding farmers increased by 25% during the 1750-1870 period, while the number of landless workers, such as crofters (torpare), who rented or owned small farms, cottagers (backstugusittare) and lodgers (inhysehjon), who hardly had access to any land whatsoever, doubled (Gadd, 2005). The industrial revolution, which began in the mid-19th century, led to the emergence of new social classes as well. In addition to the growing working class, a middle class gradually formed, consisting of, among others, industrialists and merchants (see, e.g. Magnusson, 2000).
The Sundsvall region, located in Västernorrland County, 380 km north of the capital of Stockholm, was one of the areas where the impact of these developments was most profound. Although, in 1840 it was merely an area on the periphery in northern Sweden, characterized by small farms and vast areas of wild forests, the Sundsvall region became a centre of rapid industrial development. In the mid-19th century, the region began growing into one of the world's largest cohesive sawmill areas. These sawmills had the character of large-scale industry, with modern steam engines and a great need for manpower. As this change was accompanied by a sharp increase in population, which warrants comparisons with then contemporary developments in the United States, the character of the Sundsvall region changed. Sundsvall, which used to be populated by smallholding farmers, fishermen and timber merchants, became a region dominated by the businessmen of the sawmill industry (Magnusson, 2000;Schön, 1972).
These social and economic developments were accompanied by political changes in political franchise on the local level; these changes affected voting on local matters such as schooling and poor relief. Although initially based primarily on landownership (mantal) or the taxable income of individuals, the number of votes became distributed in direct proportion to fyrkar, which was a measure of taxable income, capital and property, in the early 1860s (Nilsson, 1964;Nydahl, 2010). This change in political franchise had several important consequences, the main one being the extension of the right to participate in political decision-making to all legal persons. Thus, the right to vote was not limited to individuals; it also included companies and the managers of death estates. This change also meant that the majority of Swedish men received the right to vote at the local level. The right to vote was, however, heavily graded. Poor people had to make do with few votes, while wealthy individuals or companies could receive thousands of votes and even attain a majority of votes in a single parish (Nydahl, 2010).
The latter half of the nineteenth century also saw the rise of mass schooling. The Swedish School Act of 1842 defined a decentralized school system, for which Sweden's approximately 2,300 parishes were given the main responsibility. The school act stated that each of Sweden's parishes was required to organize a school district and run at least one, preferably permanent, primary school (folkskola). These school districts had the main responsibility for funding and organizing the Swedish school system, and their operations were mainly funded by local taxes from the populations of the parishes and by central government subsidies (Westberg, 2017). These school districts were governed by a school board (skolstyrelse) and by the parish meeting (sockenstämma), which was later replaced by the church meeting (kyrkostämma). At the parish and church meetings, all taxable inhabitants of the school district had the right to participate and place their votes in accordance with the voting system described above (Boli, 1989).
During the second half of the nineteenth century, the Swedish primary school system expanded. The number of teachers increased from 2,785 in 1847 to 21,585 in 1910, and the estimated enrolment-ratio of school-aged children rose from 21.2 in 1843 to 75. 3 in 1910(Ljungberg & Nilsson, 2009Westberg, 2017). This development was particularly strong in the Sundsvall region. Partly because of the region's rapid increase in population, its number of schools increased from 12 in 1850 to 129 in 1900 (Westberg, 2017).
Regarding the issues of decentralization, local political will and inequality, the school districts of the Sundsvall region are of particular interest because they serve as an example of how local school politics could function in northern Europe, where the development of mass education was relatively strong. Known for its rapid industrialization process, the Sundsvall region is likely to provide particularly fruitful insights into how societal factorscombined with changes in political franchiseaffected local school politics and, by extension, the development of the school system.

Conflicts involving priests, villages and parishes
Despite the aforementioned changes in social structure and suffrage, local school politics in the Sundsvall region were marked by recurrent issues throughout the period under study. This comes as no surprise, as the basic structure of the Swedish school system remained unchanged. Because teachers' salaries and school buildings were the main items of expenditure (Westberg, 2014b) throughout the investigated period, the debate at parish meetings and school boards largely concerned these two topics: Should the parish build additional schools and hire additional teachers? Who will pay for the erection of school buildings and the salaries of teachers? Where should the schools be built? Which teachers should be hired? The answers to these questions varied depending on circumstances and on the positions and the perspectives of the participants in the debate.
When these issues were addressed in school board meetings, parish meetings and church meetings, the debate was characterized by recurrent conflicts between various groups and individuals. Although there were basic conflicts between social groups, as I show below, it is important to notice that the conflicts cannot be reduced to conflicts between individuals with high and low incomes, or between large landowners and small landowners. In the decisionmaking process of the school districts, there were actually several lines of conflict.
Among the most prevalent confrontations were those between the parish priestwho chaired the parish meeting, the church meeting and the school boardand the parishioners. 1 Usually, the priest, who favoured enlightenment and Christianity, wanted to expand schooling to as many parish souls as possible, while the parishioners were more frugal in their approach. The relationship between school boards and teachers had a similar dynamic: the teachers demanded higher salaries and better school facilities, while the school board strove to keep costs down. Conflicts among villages regarding the placement of schools were also common, as were conflicts between parishes.
Such confrontations could sometimes be quite explicit. Allegedly, Swedish local politics during the nineteenth century were characterized by a consensus ideology (Gustafsson, 1989). In the Sundsvall region, however, consensus was not the primary objective of local politics. Due to the wide range of interests voiced by groups and individuals at parish meetings and on school boards, the discussion could become rather heated. Admittedly, unanimity and concordance may very well have been important ideals in Swedish political culture. Still, when parishioners had the opportunity to discuss important andin the eyes of posterityunimportant questions, the discussion easily became heated. On such occasions, the parish priest had to urge those gathered to 'at last unite' or to examine the issue at hand 'in quietness and tranquillity'. 2 In other instances, the discussion was described in the minutes as 'long and disorderly', 'rather lively', or as a 'pesky discourse adjacent to quarrel'. 3 Among the most common types of conflicts were those between parish priests and parishioners. An example of such a conflict, which became rather bitter, resulted in the vicar of Skön parish supposedly telling a parishioner that 'if you don't hold your mouth shut during the next parish meeting, I will shut it for you' (Svedberg, 1947, 84). However, these confrontations were usually less aggressive and more protracted. One such instance was when the priest Pehr Backlund was met with robust resistance from frugal parishioners when he argued in favour of building a new school in the Alnö parish in the 1850s. Instead of expressing rage or discontent, the parishioners simply rejected the plan for new or redeveloped school premises in May 1855. Some parishioners did so without presenting an explanation, some felt that the existing schoolroom was adequate for the needs of the parish, and still others claimed that the matter was simply not particularly urgent. 4 As a result of the parishioners' stance, which was expressed on several occasions, a new school was completed in 1869 (Westberg, 2014a, app. 1).
There were also disputes between villages. These conflicts were conditioned by the significant distances between villages that, in combination with the region's cold winters, hampered travelling and meant that the location of a school often became a hot political issue. An illustrative example of this is the 18-year-long process that preceded the building of a school in the village of Baggböle. Without a doubt, the length of the process was attributable to a conflict between the villages of Indal parish (in which Baggböle was located). Beginning in 1879, the villagers of Baggböle had donated a piece of land where they wanted a school to be built. Judging from the minutes, the villagers' proposal seems to have been fair. The land was located in the middle of the catchment area, and because the land was donated by the villagers, the school could be built at a lower cost. However, all the other villages in that part of the parish voted against the proposal. 5 A similar process was repeated the following year, and it was almost 20 years later, following lengthy and agitated discussion, that the proposal to build a school in Baggböle could finally be implemented. 6 The conflicts between parishes were born from various circumstances. Disputes arose when children from one parish attended the school of another parish without that parish being consulted or compensated. 7 Conflicts also developed when parishes attempted to share a school. Such attempts were made on several occasions without success. 8 There were, for example, plans for a school located in Timrå parish to be shared by the neighbouring parishes, Alnö and Skön. 9 Because these plans involved three parishes, with distinct economies, administrations and varying distances to the school building, conflicts arose almost by necessity, illustrating how difficult it was for parishes to cooperate in matters of schooling. In this case, the parishioners of Alnö and Skön argued that the parish of Timrå should cover both the expenditure of the building plot and a larger share of the teacher's salary, because the proposed school would be built in Timrå. Unsurprisingly, the representatives of Timrå did not share this position. 10 Thus, the parishes chose not to collaborate, arguing that the children's moral education would benefit if each parish had a teacher of its own. 11 These conflicts within and between school districts may certainly be linked to the decentralized organization of the Swedish school system. In addition to enabling disputes between districts, the governing bodies of the school districts gave the majority of the districts' men the opportunity to voice their concerns. Although these conflicts did not hinder school districts from taking decisions on taxes that enabled the expansion of schooling in Sweden, this form of organization allowed opposing positions to be voiced, which meant that decisions could be delayed, sometimes for several years.

School politics between farmers and landless
Despite the obvious continuities in local school politics, shifts occurred due to changes in political franchise and the expansion of the sawmill industry in the Sundsvall region. Above all, the line of conflict between different socio-economic groups shifted, which became apparent when parishioners debated about who should bear the principal burden of the expanding school system.
As studies addressing land inequality suggest (see, e.g. Beltrán Tapia, 2013, and literature cited therein), distribution of landownership marked the local school politics of the Sundsvall region in the years that followed the School Act of 1842. In the context of this region, where agriculture was dominated by smallholding farmers in the absence of a significant class of gentlemen farmers, the main conflicts were between smallholders and the landless population. Compared to the smallholding farmers of the area, the landless were a highly heterogeneous group characterized by virtual lack of landownership as well as by varying housing and working conditions. This group included the sick and old, who were housed on farmers' homesteads (lodgers) without being relatives or employees of the farmers, and crofters, who rented small farms on the homesteads of farmers and had access to small pieces of land, allowing them to engage in agriculture, albeit on a minor scale (Gadd, 2005;Lindström & Mispelaere, 2011).
In the school districts under study, the distinction between farmers and the landless was formulated in various ways. In several parishes, farmers were separated from 'crofters and their counterparts', and in others, this dichotomy was formulated in terms of farmers and the 'other people'. 12 In Alnö, 'crofters and all their counterparts' was defined as a category that included craftsmen, boatmen, lodgers, agricultural workers and industrial workers. In Timrå, the landless population was defined as 'crofters, craftsmen, boatmen, workers in general with their own households, as well as lodgers'. 13 In the Sundsvall region, the issue that divided landowners and the landless was not expenditure levels, and the goal of the smallholding farmers was not to block the rise of mass schooling, which was the concern of wealthy landowners in Prussia, Spain, England and Wales (Beltrán Tapia & Martinez-Galarraga, 2015; Cinnirella & Hornung, 2016;Goni, 2013). Instead, the local debate focused on what may be described as a free-rider problem; that is, that a group may benefit from a service without paying their share (Nilsson & Pettersson, 2008). In this case, the main question debated in the school districts' governing bodies was how the expenditure should be distributed between the farmers and the landless. Should the generally wealthier farmers cover half, two-thirds or even three-quarters of schooling expenditures? To what extent should the number of households be considered when such decisions were made? What significance should the size of households have? Was it even possible for the landless to pay for building materials?
Disagreements over teachers' salaries were resolved in different ways. In line with the stipulations of the School Act of 1842, teachers' salaries during the investigated period usually consisted of two components: non-monetary compensation that could include grain, cow fodder and firewood, and payments in cash (Westberg, 2017). In general, the parishes chose to make the farmers responsible for the cost of the in-kind salary, while the distribution of cash costs between farmers and the landless population varied. In Indal, the cost of the cash portion of the teacher's salary was distributed equally across all taxpayers, regardless of landownership, while the landless population was responsible for threequarters of the cash salary in Timrå, and for all the cash allowances in Alnö. 14 The School Act of 1842 proposed, in accordance with the Building Code of 1734, that farmers should cover the cost of school building materials, and that the farmers and landless should share the costs of labour (Backman, 1831;SFS 1842:19). There were, however, several instances when school districts chose other principles for cost allocation between landholding farmers and the landless. In Alnö parish, the parishioners considered the relative size of the landless population. Because a comparatively large proportion of the inhabitants were landless, the parish meeting decided that this group should cover between one-third and one-fourth of the bodywork needed to furnish a classroom in the existing parish house (sockenstuga), where the parish meetings were held. 15 The landless were, however, not pleased with this decision, believing that too much of the burden had been placed upon their shoulders. This resulted in a compromise: it was decided that the farmers would pay for three-quarters of the building costs, but only if the landless paid a larger part of the teacher's salary. 16 To avoid conflicts during the building process, the parish meeting in Alnö chose to further define how the costs would be split between the landless and the landholding population. Both groups agreed that the farmers should cover the costs of nails, wrought iron, glass, painting work and one ceramic stove (excluding costs for bricks, clay and sand). Discord prevailed, however, regarding the planks. The farmers argued that this cost should be covered by the landless, who objected because the raw material for the planks grew in the farmers' forests. As that particular argument could hardly be disputed, the landless ultimately avoided that charge. 17 Another solution to the thorny issue of distributing building expenditure was found in Ljustorp parish. Prior to the building of a school in 1847, the distribution of expenses between the farmers and the landless was discussed. The initial proposal was similar to that in Alnö: the landless should bear between a third and a quarter of the building costs. Because neither segment of the population approved of this proposal, the parishioners finally agreed upon the model suggested by the School Act of 1842. The farmers paid for building materials, while both the landless and the farmers covered the labour costs. 18 As is evident from these examples, decisions on taxation were seldom easy to reach. In the school district of Skön, the debate was so heated that the time allocated for the parish meeting was not sufficient. The discussions continued for such a long time because no agreement could be reached regarding what a just distribution of costs should be based upon. Considering that the majority of households were landless, the farmers claimed that the crofters and their counterparts should pay for at least onethird of the building costs. The crofters, on the other hand, stressed that the farmers were the largest population group and suggested that the landless should pay for only one-quarter of the building costs. As no consensus could be reached, the issue was resolved by drawing lots, which fell to the advantage of the landless. 19 In the school district of Tuna, the conflicts had to be settled by the county governor. When the district intended to build a new bakehouse, the landless refused to provide for the one-fourth of the costs that the parish meeting had decided they should cover. During a process in which the farmers and the landless accused each other of lies and deceit, the farmers argued that they had actually lowered the demands on the landless from two-thirds to onethird andultimatelyone-fourth of the costs. The landless, in contrast, argued that it was unlawful to force them to pay for one-fourth of the total costs, noting that they were prepared to perform one-fourth of the necessary work. Following a long legal process, the county governor decided in favour of the landless. 20 As is evident from the above, the conflict between landowners and the landless in the Sundsvall region was not about expenditure levels but rather about funding and over deciding who should bear the burden for the schools. Although a thorough explanation of this issue lies beyond the scope of this article, it may be noted that the conflict between landowning farmers and the landless was probably fuelled by the School Act's regulations of teachers' salaries and the Building Code of 1734. The regulations on salaries, which distinguished between non-monetary and monetary remuneration, promoted comparisons between farmers' contributions of grains and cow fodder, and the landless population's contribution in monies. The Building Code's distinction between farmers and the landless probably had a similar consequence. The fact that the conflict did not centre on expenditure levels may perhaps be explained by the comparatively small socio-economic differences between the smallholding farmers and the landless population of the region. It can be assumed that these groups had similar interests in tax-funded education. Therefore, the question was less about whether schools were to be established and more about who should pay for them.

Conflicts structured by the distribution of wealth
During the investigated period, the nature of conflicts shifted. In combination with the wealth created by the sawmill industry, the changes in political franchise and tax regulations in the 1860s altered the character of local school politics. Like the political franchise of local government, the local parish taxes (when no other regulations applied) became based on estimates of wealth (fyrkar) instead of, for example, income, households or landownership (Thulin, 1935). These changes were fully implemented in 1876 when all expenditure on the school system was to be covered by taxes based on fyrkar, with the exceptions of school fees and a minor per capita tax (SFS 1876:47). As a result, the line of conflict was no longer drawn between landholding farmers and the landless population but between the wealthythose who owned many fyrkarand those who were not wealthy.
Under such conditions, the burdens of schooling, as well as the political voice, were increasingly unevenly distributed across the population of the Sundsvall region. In the Njurunda school district, for example, a single individual or company possessed as much as 34% of the votes in 1904; in the Timrå school district, this figure was 29%, and in the Skön school district, 25% (BiSOS R, 1904, tab. 1:21). This uneven distribution of rights, as well as of responsibilities, gave rise to conflicts on several occasions. As noted in the introduction, existing research on political voice has found that restricted political franchise has, in several instances, hampered the rise of mass schooling. The privileged elite, e.g. German Junkers, England's landed Tories, and U.S. plantation owners, also explicitly opposed the expansion of schooling, fearing that it would lead to social unrest (Lindert, 2004).
The sawmill owners and affluent merchants of Sundsvall, however, did not show any signs of such fears. Instead, sawmill owners funded schools themselves, and they probably viewed schools as both a method of social control over their working forces and a part of the patriarchal relationship between sawmill owners and workers (cf. Florin & Johansson, 1984;Michaëlsson, 2016). Such attitudes were, for example, expressed by the county governor Curry Treffenberg (1825-1897), who investigated the Sundsvall strike in 1879 on behalf of the sawmill owners. Treffenberg argued that schooling was one of the measures that could be taken to prevent the strike from reoccurring (Florin & Johansson, 1984).
The sawmill owners were, however, somewhat cautious with regards to the sums that they themselves would have to spend on schooling. Therefore, the fear of bearing an excessive portion of the costs of schooling, rather than the fear of schooling in itself, gave rise to disputes on several occasions. These conflicts could be formulated in terms of what responsibility companies should take for the schools of a parish. In Attmar, the ironworks of Sörfors argued that it was obliged to pay taxes for only the one school in the proximity of the ironworks. 21 In other instances, the debate focused on the tax design and on whether school taxes should be based on landownership or on wealth (fyrkar). Not surprisingly, the wealthy inhabitantswho consequently possessed the majority of the voteswon the vote, which meant that school expenditure would be covered by taxes based on landownership. One such debate concerned the school of Rude, built in 1884. The church meeting was presented with two options. Under the first, the building should be taxed proportionate to wealth (fyrkar), which was fully in accordance with the general regulations on local taxation. Under the second proposal, labour costs should be distributed evenly across the parish households. Because the wealthy voters benefited from the second proposal, it won the vote, even though only 7 out of 26 present parishioners voted for the second proposal. 22 The new legislation on taxes and political franchise also affected the debate on other issues. One important issue was, of course, where new schools were to be located. At the beginning of the investigated period, it was usually difficult to place a school in a distant village because the farmers of a single village usually did not have enough landand thus enough votesto win a poll. However, when voting rights became proportional to wealth, even the most controversial decisions could be implemented. For example, only a minority of the inhabitants of Ljustorp parish wanted to place a new school in the village of Ås, but because the vicar Esbjörn Bergman and the sawmill owner Johan August Enhörning voted in favour of this proposal, the decision was accepted in 1873, with 2,700 votes in favour and only 900 against. 23 Due to the changing fortunes of the sawmill industry, individuals' incomes could change dramatically over time, and these fluctuations created problems in the Sundsvall region. During the second half of the 19th century, the sawmill industry expanded rapidly, concentrating the voting power in the hands of sawmill companies and their owners (Nydahl, 2010;Tedebrand, 1972). However, because the sawmill industry created fortunes, it could also take them away when the economy took a downturn, as it did, for example, during the late 1870s and the early 1890s. As farmers who were critical of this system of suffrage noted, the sawmill owners of Njurunda parish possessed 158,000 votes in 1890, and this number dramatically decreased to 65,000 2 years later 24 .
In the school district of Njurunda in 1892, the debate on schooling stood at the intersection of sawmill industry cycles and wealth-based taxation and suffrage. This debate is illustrative because it indicates how the conflicts at the end of the investigated period were not created by the unequal distribution of land but by the unequal distribution of wealth. This debate also illustrates that the main issue in these conflicts was not whether investments in schooling were necessary but rather how these investments should be funded. In Njurunda, there was an agreement that additional schools were required to cater to the expanding population of the school district. The debated issue was, instead, whether the construction of eight new schools should be funded by taxation or by a mortgage loan of 55,000 Swedish kronor to be paid off over a 20-year period. A completely overwhelming majoritya total of 26,581 votes possessed by only six personsvoted for the latter option, which caused resentment among the 46 parishioners who lost the election because they possessed only 2,788 votes. 25 The reasons for this outcome are obvious. For those with high incomes, such as the famous sawmill owner Oscar Dickson, loans appeared to be the best method of funding school building projects. Because the profits from the sawmill industry were expected to decrease, the incomes of the sawmill owners were likely to decrease as well. Loans that were repaid in the future, by local taxes based on wealth (fyrkar), thus appeared attractive for such rich merchants and industrialists. Sawmill owners also relied on national and international networks, which gave them excellent opportunities to move to other regions in Sweden or abroad in the future, protecting them from having to repay any loans through local taxes. For example, the aforementioned sawmill owner Oscar Dickson had attended Klügmann's trading academy in Lübeck (Germany) and worked at the office of Dickson Brothers in the southern Swedish city of Gothenburg before he moved to Sundsvall (Carlgren, 1945). In contrast, farmers relied on revenues from their lands and thus had few opportunities to relocate elsewhere. Consequently, they preferred that investments in schools be paid through taxation. Because taxes were based on wealth, taxes would also place the burden of a significant portion of the costs on the shoulders of the sawmill owners.
In conclusion, the local political conflicts in the Sundsvall region presented above are interesting examples of how not only school politics but also the main lines of conflict among social groups can change over time. Due to rapid societal developments driven by the expansion of the sawmill industry and due to changing regulations regarding local taxes and political franchise, the main conflict was no longer that between landowners and the landless but rather that between the wealthy and the non-wealthy.

Conclusion
This article has shown how local school politics were characterized by several issues and conflicts, some of which occurred repeatedly throughout the investigated period. Teachers did not always share the interests of the school board, the parishioners did not always have the same objectives as the vicar, and the needs of the parishes' villages often conflicted with each other. Nevertheless, the character of local school politics changed. Most importantly, the basic conflict of school politics was no longer between those who owned land and those who did not; instead, it was between highand low-income groups. In contrast to local elites in other countries, who opposed the expansion of schooling, the elites of the Sundsvall region focused on school funding, and the main conflicts concerned the distribution of school expenditures.
In addition to presenting insights into the impact of inequality on local school politics, this article has shown that the basic conflicts of school politics should not be taken for granted, as they are always the results of specific social, economic and political contexts. I have shown that the conflicts between farmers and the landless may be explained, at least in part, by regulations on building costs and teachers' salaries. The distinctions that these regulations made between in-kind and monetary remuneration, and between the responsibilities of farmers and the landless when building schools, enabled comparisons that fuelled conflicts between these groups. I have also argued that the abovementioned changes in the conflicts of local school politics were due to changes in political franchise, tax regulations and the rise of wealthy sawmill owners following the expansion of the sawmill industry.
These results raise further questions. The actions of local elites on the governing bodies of school districts raise questions about the impact of the inequalities of the Swedish countryside, particularly regarding the distribution of land and political voice. At parish and church meetings in the Sundsvall region, the wealthier strata did not explicitly vote against the expansion of schooling. In line with the findings of Andersson and Berger (2016) on school districts in southern Sweden, sawmill owners could actually vote for large loans to fund investments in new schools. Econometric studies are, however, needed to clarify whether the Swedish rural elites in general did not hamper the rise of mass schooling but instead affected how school taxes were distributed among the population. That is, were the Swedish rural elites generally more concerned with the funding of mass schooling rather mass schooling in itself?
As noted in the Introduction section, Lindert (2004) emphasized that the decentralized nature of nineteenth century school systems was a crucial condition for their expansion. According to Lindert (2004, p. 121), it was the 'spontaneous political will to levy local taxes in thousands of school districts' that made this development possible. Viewed from Lindert's perspective, this article has provided additional insights into how this political will was formed by a particular historical context. More precisely, it has described how the political will to invest in schooling was shaped by regulations on school buildings and teachers' salaries, as well as by changes in social structures, tax regulations and political franchise. Finally, this article has described how this political will emerged from conflicts among different social groups. Although the political will of the school districts in the decentralized Swedish school system was certainly spontaneous, this article has emphasized that this political will was also conflicted and dependent on a specific social, economic and political context.