6.1. South Carolina: division and cooptation

In this period, South Carolina was divided between two broad areas.Footnote ³⁰ To the north was the state’s “upcountry,” defined by the hills of the Piedmont. The region was majority white, in part because the poorer soil quality had prevented both the establishment of large plantations and the importation of large numbers of enslaved people. In the decades before the Civil War, the region had transitioned from subsistence agriculture to cotton production, enmeshing it in new national markets. White farmers generally eked out an existence on small farms worked by their families and a few hired employees.Footnote ³¹ To the south lay the state’s coastal plain, the “lowcountry.” The lowcountry was geographically larger—about two-thirds of the state’s land area and a majority of its counties. This region was majority Black, and dominated by white-owned plantations. Its continuing political dominance partially obscured its economic decline as poor infrastructure and inefficient farming took their toll.Footnote ³²

6.2. Louisiana: fusion and repression

Louisiana was more geographically and demographically fractured than South Carolina was during the post-Reconstruction era of 1877–1901. The state’s northern parishes, which largely produced cotton, were divided between majority white parishes in the state’s less fertile interior and majority Black parishes that bordered the Mississippi River and were dominated by large plantations. The southeastern region of the state, bounded to the west by the Bayou Teche and the east by the Mississippi River, grew sugar on large plantations worked by Black laborers who worked for wages, rather than a share of the crop.Footnote ⁶¹ The region’s French and Spanish heritage had left a lasting imprint in the form of Catholicism. New Orleans, the state’s cultural and commercial capital, was, during this period, by far the largest urban center in the South. The city’s political leaders allied with the state’s conservative landowners in exchange for the resources the city needed to grow and sustain itself.Footnote ⁶² I will lay out some of the political consequences of these regional and cultural cleavages in more detail below.

6.2.2. Incorporation

Unlike in South Carolina, Louisiana’s economic and political elite demanded a constitutional convention, over the objection of rural agrarians, in order to entrench their rule. A sizable group of non-Democrats—Populists, Republicans, progressive reformers from New Orleans—were elected to the Louisiana legislature in 1896. However, Democrats still held sizable majorities in both houses of the legislature. They used these majorities to call a constitutional convention. However, the legislature first passed several laws designed to shrink the electorate.Footnote ⁹⁴ It was this shrunken electorate that elected the delegates charged with crafting the state’s new constitution. Only one Populist, Bryant Bailey of Winn Parish, and one Republican, J. Dubliex of the sugar parish of Iberville, managed to secure election. The remainder of the elected delegates were Democrats.

Agrarian radicals, despite their defeat in the 1896 elections and subsequent legal repression, continued to inspire fear in prominent Democrats and influence the workings of the convention as it opened in 1898. E.B. Kruttschnitt, elected by his fellow delegates as president of the convention, declared in his opening address that the task of the convention was “to eliminate from the electorate the mass of the corrupt and illiterate voters who have during the last quarter century degraded our politics.” Kruttschnitt clarified in two other sets of public remarks that he believed in both the necessity and the desirability of limiting both Black and white suffrage.Footnote ⁹⁵ A durable cleavage had opened up between poor, rural white voters and the state’s conservative political and economic elite. Kruttschnitt was also clear that the agrarian electoral mobilization influenced the timing of the convention. He explained in a speech that “This convention was called midway between the exciting gubernatorial campaign of 1896, and that of 1900, which bids fair to be just as exciting, in both national and state politics, as that of 1896. Thus the legislature set the time when political antagonisms were at their lowest.”Footnote ⁹⁶ Kruttschnitt and his allies did not see the Populist movement as a thing of the past—instead they saw it as a continuing threat requiring both repression and, crucially, conciliation.

As in South Carolina, the suffrage committee was a focal point of the convention. Kruttschnitt and his allies stacked the committee with delegates from black majority counties, as well as urban conservatives from New Orleans. However, the convention’s deliberations on the suffrage provisions were fraught. Fights about the poll tax threatened to “split the party wide open.” At one point, two opposing caucuses formed in the convention—one supportive and the other opposed to the measure.Footnote ⁹⁷ An “understanding” clause divided the convention as well. In another example of the enduring influence of Populism, Kruttschnitt argued that the “understanding clause” posed a danger because the political opposition (the Populists) might win control of local election machinery and then allow their voters to register.Footnote ⁹⁸

Many scholars of this period of Louisiana’s history, despite other disagreements, largely see the convention as mostly a formality—the constitutional entrenchment of an already existing nondemocratic reality. They see Populists as exercising little influence, and the new constitution as the reincarnation of existing documents, with suffrage restrictions welded on.Footnote ⁹⁹ Perman paints a picture of a convention riven by factionalism and internal tumult that saw different coalitions control the outcomes of different provisions that, at critical moments, left its final outcome in doubt. However, despite outlining the conflict that characterized the process of the convention, Perman largely hews to conventional wisdom about the result. The “other changes” the convention made to the 1879 constitution, Perman argues, “were relatively insignificant.”Footnote ¹⁰⁰

While these accounts capture key aspects of the political and coalitional dynamics that influenced the outcome of Louisiana’s 1898 constitutional convention, they do not fully describe the lingering influence of the Populist insurgency on the design of Louisiana’s governing institutions. Perhaps the most important oversight of existing scholarship is the neglect of the apportionment committee—the committee charged with updating the formula by which the constitution apportioned seats in the Louisiana House and Senate. The members of this committee represented a different set of regions and constituencies than the members of equivalent committees in South Carolina. In South Carolina, the apportionment committee was stacked with representatives from majority-black counties who tended to be political opponents of Tillman—a double incentive to ensure that apportionment favored the Black Belt. In Louisiana, the committee drew a narrow majority of its members from Parishes that were majority-white—many of which had voted for the Populist-Republican ticket in 1896. Delegates from New Orleans also secured representation on the committee—the chair of the committee was from New Orleans.Footnote ¹⁰¹

The apportionment scheme proposed by the committee and ultimately included in the constitution by the convention afforded fairer representation to white voters who had backed the Populist-Republican ticket in 1896. Utilizing the Loosemore–Hanby index as a measurement tool, the Louisiana Senate was more fairly apportioned on the basis of white population in 1900 than state senates in South Carolina. When compared to the apportionment included in Louisiana’s 1879 constitution, the 1898 constitution actually apportioned Louisiana’s state senate more fairly. In contrast, the total malapportionment of Louisiana’s lower house was similar to the distortion of representation in South Carolina’s lower house.

However, a metric of the total degree of malapportionment does not tell the full story. It does not capture which constituencies or regions are disenfranchised and which receive representation disproportionate to their numbers. In Louisiana, the constituencies that formed the core of the Populist-Republican coalition were much better represented after their state’s authoritarian founding than their Deep South Peers. In 1900, the 50 percent of the white population of Louisiana that lived in the parishes where the Populist-Republican ticket performed the best (which did not include New Orleans) in 1896 had about 45 percent of the seats in the Louisiana House of Representatives and 46 percent of the seats in the Louisiana State Senate. Combined, this was only about a 4.5 percent gap between the chosen apportionment formula and an equal apportionment on the basis of white population. In Louisiana, a potential future political faction that reassembled the surviving elements of the Populist-Republican coalition would only need to pick off a few opportunistic representatives from New Orleans to field a working majority in the state legislature.

It is possible that rural white landowners from majority-black parishes accepted this apportionment scheme because they believed they had entrenched their dominance of state politics in other ways. Right before the convention passed the final suffrage provision, delegates from majority-black parishes inserted a final modification. The last-minute amendment stated that the delegates to party conventions called to select nominees be apportioned based on total population, rather than the vote at the last election.Footnote ¹⁰² This would ensure that party conventions—in 1898 still the method by which Louisiana Democrats selected their nominees for statewide office—would continue to be dominated by rural landowners from majority-black counties. However, this factional advantage would last only a few years, until the introduction of the white primary in Louisiana between 1904–1906. Tellingly, Kruttschnitt vociferously opposed the white primary. The New Orleans Times Democrat, summarizing a public statement he made, said that Kruttschnitt “regards general primaries as revolutionary and as tending to produce discord and other party evils to a greater degree than the good sought by the advocates of direct elections.”Footnote ¹⁰³ He likely realized that the adoption of the white primary destabilized the delicate balance achieved in the convention and opened up a path—however narrow—for a coalition of agrarian reformers and urban dissenters to achieve control over both the executive and legislative branches of state government.

The continuing influence of the Populist mobilization is also evident in how the convention approached the issue of railroad regulation. Originally, the proposal to create a railroad commission was referred to the committee on corporations.Footnote ¹⁰⁴ However, on March 29, 1898, Phanor Breazeale, a delegate from Natchitoches, motioned to refer the proposal to the committee on general provisions instead. Breazeale supported the creation of a strong railroad commission, and before making his motion, he urged those with a financial interest in the question to recuse themselves—a signal that he believed that railroads favored the corporations committee maintaining jurisdiction over the issue. His motion passed with sixty-four votes in support, and thirty-five opposing. Twenty-Eight delegates abstained. Figure 1 plots the relationship beteween Parish-level vote share received by the Populist-Republican ticket and the propensity of delegates from those parishes to vote in favor of Breazeale’s motion. Of the sixty-four yes votes, twenty-eight came from parishes where the Populist-Fusion ticket received at least 49 percent of the vote, including delegates from twelve parishes that were also majority-white. Of the thirty-five no votes, only four came from parishes where the Populists had received above 49 percent of the vote—the rest came from parishes where Foster Murphy had won a clear majority of the vote in 1896.

Figure 1.

Relationship between the Parish-Level Vote Share for the Populist-Republican Ticket in Louisiana and Propensity of Convention Delegates to Assign the Proposal to Create a Railroad Commission to the Committee Favored by the Commission’s Chief Supporter.

Source: Authors own calculations based on roll call votes from Louisiana’s 1898 constitutional convention.

Five of the eleven members of the corporations committee came from majority-black parishes—a sixth represented Rapides, which was just 51 percent white. Moreover, only two of the delegates came from parishes where the fusion ticket had won a majority, and only one of those parishes (Calcasieu) was also majority-white. By contrast, seven of the nine members of the committee on general provision came from majority-white parishes, two of which had also given a majority to the Populist ticket. This was potentially more favorable ground for reform. The subsequent article that the committee drafted was surprisingly robust. The commission itself, including an outline of its powers and responsibilities, was entrenched in the constitution. Just as importantly, its political structure and method of election were also constitutionally established—cutting off avenues for the legislature to dominate the commission via alterations to its mandate and method of election.Footnote ¹⁰⁵

7.1. Legislative malapportionment

Between 1900 and 1930, variation in legislative malapportionment between South Carolina and Louisiana increased. Wealthy elites in South Carolina took advantage of how they had structured and apportioned the legislature in the 1895 state constitution to further expand their power in subsequent decades. Between 1895 and 1919, South Carolina Democrats created eleven new counties. Seven of them were majority Black and dominated by wealthy, white plantation owners. Consequently, even as the white population of the upcountry region grew much faster than that of the lowcountry during this period, wealthy lowcountry whites strengthened their domination over the powerful state senate, in which each county had one senator. In contrast, in Louisiana, a more favorable institutional starting point allowed poor whites to maintain, and even strengthen, their position in the legislature. All five new parishes created by the legislature after 1898 were majority white, and Louisiana’s 1921 legislative reapportionment contributed to an overall decrease in malapportionment between 1900 and 1930. Consequently, in Louisiana, poorer whites who lived in white-majority, rural counties—the electoral base of populism—achieved fairer representation in the legislature than did their peers in South Carolina.

The stability of the discrepancy in legislative malapportionment between South Carolina and Louisiana points to the durability of the constitutional equilibriums established in 1895 and 1898. In Miller’s terms, legislative malapportionment was a “self-referencing” institution—it allowed those who benefitted from it to protect it.Footnote ¹⁰⁶ Any change in malapportionment—via either legislation or constitutional amendment—had to originate from the malapportioned legislature.

Malapportionment interacted with the political system to block and distort the distribution of public goods in three ways. First, it prevented majority coalitions from translating their numbers into commensurate political power. If a majority coalition within the voting public resides in legislative districts that together constitute only a minority of seats, they will not be able to elect a legislature responsive to their preferences. Second, malapportionment allowed legislative power to be exercised by a minimum winning coalition that was a minority of the voting public. This minority often has distinct interests and preferences that conflict with the preferences of the majority. Sometimes, these minority coalitions prefer lower taxes and less redistribution overall.Footnote ¹⁰⁷ At other points, they may prefer to use their outsize political power to secure a greater share of resource transfers for themselves.Footnote ¹⁰⁸ Third, malapportionment shielded recalcitrant legislators from either electoral backlash or political pressure from otherwise popular statewide political figures. If an antiredistribution legislator knows that his or her voters approve of their actions, and the legislative majority they are a part of is secure, they can ignore both popular protest and attempted arm-twisting by a pro-redistribution governor.

I observe all three of these mechanisms at work in the South during this period and offer examples of each in the analysis below. When wealthy, conservative landowners entrenched nondemocratic rule in the South in the 1890s, they designed legislative apportionment schemes that would allow them to wield legislative power far in excess of their numbers. Consequently, wealthy white landowners were over-represented in the South Carolina and Louisiana legislatures. This allowed them to use legislative institutions in their states to impose their preferences with regard to redistribution on the rest of the voting public. Protected from electoral sanction, they could also ignore popular pressure and the entreaties of pro-redistribution governors. Consequently, the degree of malapportionment mattered a great deal. In Louisiana, the majority coalition that Huey Long constructed was eventually able to command legislative majorities and enact much of his redistributionist agenda. In South Carolina, the popular majorities that supported Governor Olin Johnston were not able to win control of South Carolina’s legislature. Consequently, his policy agenda with regard to public goods provision largely was not enacted into law.

I pay special attention to the state senate in both Louisiana and South Carolina for two additional reasons. In South Carolina, the Senate was the more malapportioned of the two houses. Consequently, it was the limiting factor on what legislation could pass—a fact I observe repeatedly in the 1920s and 1930s. Moreover, the constitution delegated to the Senate in both states the power to hold trials for impeached governors, rendering it the final arbiter of many of the most intensive conflicts between the executive and legislative branches. Table 1 contains a side by side comparison of legislative malaportionment in South Carolina and Louisiana before and after their constitutional conventions. The table uses several metrics to quantify post-convention legislative malapportionment.

Table 1.

Effect of Malapportionment on Populism’s Former Voting Base

Table 1 Long description

Malapportionment levels are listed for South Carolina and Louisiana before their constitutional conventions and again for the upper house in 1900 and 1920, alongside who dominated each convention’s apportionment committee. South Carolina’s pre-convention malapportionment is 0.160 for the upper house in 1880, with 9 of 11 committee members from majority-Black counties; Louisiana’s is 0.167 in 1880, with 8 of 15 members from majority-white parishes. In South Carolina, upper-house malapportionment increases from 0.174 in 1900 to 0.284 in 1920, indicating a substantial rise in unequal representation over time. In Louisiana, upper-house malapportionment is lower and changes little, moving from 0.145 in 1900 to 0.137 in 1920. For 1920 lower-house seat allocation in Louisiana, the half of the white population living in parishes with the highest Populist vote share in 1896 receives 48 percent of seats; South Carolina has no comparable Populist measure. Also in 1920, the half of the white population living in the whitest counties or parishes receives 41 percent of lower-house seats in Louisiana and 36 percent in South Carolina, with South Carolina’s state senate figure noted as 28 percent. Values are based on white population counts, reflecting that the effective electorate in these states at the time was overwhelmingly white.

* To compute these numbers, I utilized census data detailing white population by county in 1900 and 1920. I accessed the data online (https://www.socialexplorer.com/a9676d974c/explore). I aggregated the data up to the legislative district when necessary. I used the white population because after the consolidation of southern authoritarianism, the electorate in South Carolina and Louisiana was practically entirely white. In the context of an all-white electorate, white agrarians sought to achieve legislative apportionment that would be fairer relative to past formulas when calculated on the basis of white population alone. The Loosemore–Hanby index—a popular metric for computing and comparing malapportionment—adds up the gaps between the percentage of a population in a legislative district and the percentage of seats that district is awarded. The final score is derived by summing the absolute value of all of those numbers and dividing the resulting number in half, for the measure of how much legislative power favored districts would have to give up to generate a proportional legislature (Mickey, Reference Mickey2015; Samuels and Snyder Reference Samuels and Snyder2001).

† I included measurements for both the Louisiana House and the South Carolina Senate because both were limiting factors among their state’s political institutions. The Louisiana House was more malapportioned than the Louisiana Senate, and the South Carolina Senate was far more malapportioned than the South Carolina House.

7.2. Tax capacity

The variation in southern states’ fiscal capacity, now entrenched in newly durable state constitutions, persisted during the early decades of the twentieth century. Louisiana’s constitution contained stronger limits on the local taxation of property than South Carolina’s constitution did. Under the state’s 1898 constitution, the state was forbidden to levy more than six mills of taxation on property.Footnote ¹⁰⁹ Parishes could raise their property taxes as high as ten mills, with municipalities allowed to impose another ten mills of taxation. Thus, Louisiana’s limit on property taxation overall was twenty-six mills. South Carolina, by contrast, had laxer limits on local property taxes. The state’s 1895 constitution imposed fewer limits on how county and local governments could tax and spend. Consequently, by the mid-1920s, the rate of property taxation levied ranged from a high of fifty-two mills in Abbeville to 38.6 mills in Georgetown.Footnote ¹¹⁰ Higher rates masked the fragility and inequality that defined South Carolina’s tax system. Revenue from property taxes was highly sensitive to economic downturns, and varied tremendously between rich and poor counties. In contrast, the relative lack of property tax revenue in Louisiana pushed the state to utilize other sources of revenue collected directly by the state.

7.3. The balance of power between legislatures and governors

There was variation in how the Louisiana and South Carolina constitutions distributed power between the legislative and executive branches. The Louisiana constitution provided for a stronger executive branch than the South Carolina constitution did. For instance, governors in South Carolina had, relative to their counterparts in Louisiana, very weak powers of appointment and little patronage to distribute, as the state senate exercised an unusually high degree of influence and control over executive branch appointments.Footnote ¹¹¹ In Louisiana, by contrast, governors had reasonably expansive appointive powers, including over state boards that controlled large budgets and workforces.Footnote ¹¹² As a result, it was harder for South Carolina governors to cement factional loyalty via patronage or to get the state’s relatively weak executive agencies to align with the governor’s priorities. After he was elected in 1910, Cole Blease sought to reward his friends with jobs and patronage so as to strengthen his influence in the South Carolina Democratic Party. He broke precedent by ignoring the recommendations of the Spartanburg legislative delegation and appointed his own choices as local magistrates. The Senate refused to confirm his magistrate appointees, and a subsequent ruling from the state Supreme Court ejected those he had already installed.Footnote ¹¹³

This variation in how power was allocated among the branches of government had implications for the balance of factional power within the Democratic Party because of the newly instituted white primary. After first appearing at the local level in the 1870s, by the first decade of the twentieth century the white primary was entrenched throughout the South—including in South Carolina and Louisiana.Footnote ¹¹⁴ As its name suggests, the white primary restricted voting in Democratic Primaries only to white voters—provided they had paid their poll tax. It served as an additional tool of disenfranchisement, since winning the Democratic nomination was tantamount to election in the South during this period. For white voters, however, the white primary was a democratizing reform because the power to determine party nominees was taken away from county and state party conventions and given to the (white) electorate. Party primaries allowed candidates disfavored by elite actors and interests to win the Democratic nomination and, subsequently, election. This is why, as I outline above, E.B. Kruttschnitt (the conservative chair of Louisiana’s 1898 constitutional convention) publicly opposed the introduction of the white primary in Louisiana.

Statewide primaries, relatively speaking, favored candidates supported by poor workers and agrarians because these constituencies (if united) formed a majority of the electorate. Malapportioned legislatures favored conservative, elite interests because they were largely insulated from majoritarian pressure. A constitutional system that allocated more power to the governor at the expense of the legislature (as was the case in Louisiana) was one where pro-redistributionist governors, when elected, would likely wield more power in service of their agendas. In states such as South Carolina, where the governor was very weak, even a popular mandate for increased public goods provision would likely founder on the shoals of an intransigent, insulated conservative legislative majority. As I detail below, both states followed this theoretical expectation.

Differences in the distribution of power among the branches of government affected other executive branch agencies beyond the office of the governor. The differences between each state’s railroad commission were also stark. South Carolina’s commission could be strengthened or weakened at will by the legislature. In Louisiana, the commission’s powers and method of election were entrenched in the constitution, granting the agency an unusual degree of power and autonomy from interference by the state’s legislature. That power attracted savvy, ambitious politicians to the office, which proved a powerful launching pad. T. Harry Williams describes how Huey Long, prior to his 1918 run for railroad commissioner, looked up the powers and responsibilities of the agency in the state constitution and, seeing its potential, decided to run for the position.Footnote ¹¹⁵ He narrowly won the election.

8.1. Public schools and southern political development

The development of public education in the South is heavily bound up with the expansion and contraction of southern democracy.Footnote ¹¹⁹ The final consolidation of single-party, authoritarian rule in the South via the passage of each state’s authoritarian constitution freed Democratic Party elites from the dual threats of black political action and interparty competition.Footnote ¹²⁰ Consequently, elite Democratic lawmakers were free to target their austerity and neglect more precisely without the fear of electoral blowback. The result was a public education system defined by grievous racial inequality, achieved via a variety of formal and informal methods.Footnote ¹²¹ However, the oligarchic nature of authoritarian rule in the South resulted in large disparities in public school funding among whites as well.Footnote ¹²² Poor white agrarians, who had evinced a preference for more generous funding of public schools since the antebellum period, chafed at this arrangement and struggled to change it.Footnote ¹²³

Their efforts were often frustrated by powerful, malapportioned legislatures keen on protecting the property and prerogatives of wealthy conservatives. Cole Blease was elected governor of South Carolina in 1910 and 1912 thanks to strong support among mill workers in South Carolina’s majority white upcountry region.Footnote ¹²⁴ In 1913, Blease proposed a one-mill statewide property tax to increase funding for public schools. Channeling the preferences of his supporters, Blease wanted the money to be held in a general fund and distributed by the State Board of Education. Doing so would ensure that white students in poorer counties got a larger share of the new funding. The legislature drastically altered the legislation before passing it, decreeing that counties would retain half the revenue from the new tax within their borders—making the measure far less redistributive.Footnote ¹²⁵ Blease’s popular mandate—and the preferences of his supporters—was no match for the institutionally insulated legislators of South Carolina’s General Assembly.

Beginning in the early 1920s, several political and economic shifts combined with the longstanding preference of white agrarians for more school funding to produce large expansions of state support for education. Political elites became increasingly embarrassed by white illiteracy, and so-called New South boosters believed a better educated workforce would draw in investment and cure the ills of incipient industrialization.Footnote ¹²⁶ South Carolina enacted its expansion of state support for its public schools in 1923. Louisiana followed suit in 1930. Consequently, public school systems came to be some of the most capacious and best-funded state bureaucracies in the region, and a vital nexus between state and local political authority. This section utilizes a mixed-methods approach, employing both historical process tracing and statistical analysis. I explore three different types of variation in each state’s distribution of public school funds between 1920 and 1938: (1) the strength of the linkage between vote choice and the provision of school funding, (2) variation in the geographic distribution of funding for white schools among each state’s counties, (3) the durability of the reforms in the face of economic dislocation from the Great Depression.

8.2. The politics of public education in the Deep South

The politics surrounding the enactment of expanded state support for public education in Louisiana and South Carolina differed. In Louisiana, Huey Long rallied popular support for expanded funding; first during his unsuccessful bid for the Democratic nomination for governor in 1924, and then during his successful run in 1928. Long campaigned against illiteracy—particularly white illiteracy—and touted his proposal for free school textbooks as a way of decreasing inequality among students in the public school system.Footnote ¹²⁷ In an example of the rhetoric he employed in his campaign, Long drew on the imagery of Longfellow’s poem Evangeline to remind his audience that “Evangeline is not the only one who has waited here in disappointment. Where are the schools that you have waited for your children to have, that have never come?”Footnote ¹²⁸ Long’s campaign events drew large and enthusiastic crowds. In Terrebourne Parish, supporters packed a theatre to hear him speak, with a “large number” unable to enter the venue due to lack of space.Footnote ¹²⁹ A newspaper in Vermillion Parish estimated that one thousand people came to hear Long speak on Christmas Eve, 1927—right before the election.Footnote ¹³⁰ Turnout in the Democratic primary for governor doubled between 1920—the last election where Long was not on the ballot—and 1928, when he won the primary.

In South Carolina, by contrast, expanded state support for public education did not arise out of political campaigns rooted in agrarian discontent. Thomas McLeod, the governor elected in 1922, was no agrarian champion. Rather, reform was an elite-driven effort.Footnote ¹³¹ Despite efforts by reformers, during the 1923 legislative session, to create an appearance of unified and widespread white support for education reform via rallies and mass meetings, “few South Carolinians were involved in these well-orchestrated but largely ceremonial gatherings.”Footnote ¹³²

Enthusiasm aside, who were the voters who supported the governors who ultimately oversaw the passage and implementation of the expansion and reform of public education funding in their respective states? First, I examine how closely county-level support for Populist candidates for governor approximated county-level support for each state’s governor during their campaign. Second, using data from the annual reports of each state’s auditor, I compiled a county-level dataset of per capita taxable wealth.Footnote ¹³³ My results show important differences in the political coalitions of the governors who implemented education funding expansions. As shown in Figure 3, Thomas McLeod drew his strongest support from wealthier counties, as measured by per capita white wealth, in his successful 1922 campaign for governor. In contrast, as shown in Figures 2 and 4 Huey Long’s drew his strongest support from the poorest parishes and parishes that had supported the Populists in 1896. My analysis confirms claims made by Key and Sindler about the composition of Long’s coalition.Footnote ¹³⁴

Figure 2.

Correlation between the Parish-Level Vote Share for Huey Long in 1928 and Populist Candidate John Pharr in 1896.

Source: Election returns contained in Heard and Strong (1950).

Figure 3.

Correlation between Parish-Level Per Capita Wealth and 1922 McLeod Vote Share.

Source: Author’s own analysis based on county-level data on property values from SC.

Figure 4.

Correlation between Parish-Level Per Capita Wealth and 1928 Long Vote Share.

Source: Author’s own analysis based on data from LA auditor report.

However, winning the gubernatorial election was necessary but not sufficient for Long, and the coalition he represented, to enact an expansion of public school funding. Long needed the cooperation of the Louisiana legislature. He enjoyed more success on this front than his peers in South Carolina for two reasons. The legislature itself was less malapportioned, and Louisiana governors controlled a significant amount of patronage via influence over large state agencies. Long was considerably more aggressive than his predecessors at bringing agencies such as the state Highway Commission and Orleans Parish Levee Board under his personal control. This allowed him to offer jobs to those who agreed to go along with his agenda. As a result, Long was able to get his favored candidates chosen as leaders of both the upper and lower houses of the legislature, and exert considerable influence over the chairs of key committees.Footnote ¹³⁵

During his first year in office, Long benefitted even more directly from Louisiana’s more fairly apportioned legislature when his political opponents in the legislature, despite the opposition of legislators Long had elevated, initiated impeachment proceedings against him. He faced nineteen indictments in total, including bribery and abuse of executive authority, of which only eight were brought as formal charges.Footnote ¹³⁶ The lower house of the Louisiana legislature impeached him, sending the charges to the state senate for a trial. Working with his allies in the senate, Long managed to convince fifteen senators—enough to acquit him—to sign a statement declaring that they would vote to find him innocent. Eleven of these senators were from parishes that had backed the Populists in 1896, and then backed Long in his election as governor. Two others were from New Orleans. Had Louisiana’s legislature—particularly its senate—been apportioned less favorably to poor, rural whites, Long would likely have been unable to reach this crucial one-third threshold to prevent his removal from office. Louisiana’s political institutions created opportunities for political leaders, like Long, to win control over powerful offices in the executive branch. The state’s legislative apportionment—which was, at least compared to South Carolina, relatively favorable to poor whites—then played a role in enabling these leaders to stay in office and maintain their power.

8.3. The design of Deep South education reforms

Despite their different paths to winning and maintaining their respective offices, both Long and McLeod faced a common challenge of designing legislation that could satisfy their respective coalitions and be enacted by their legislatures. South Carolina reformed its educational system in 1923, with Louisiana following in 1930. Though these plans had similar goals, reformers—responding to the different political and institutional constraints of their respective states—designed them differently.

Both states legislated minimum education standards. For the first time, these states would guarantee the necessary funding to close any gap between what the county or parish could appropriate for its white students and what was needed to implement the new state standard. South Carolina legislated a minimum school term of 7 months (Laws of South Carolina 1923). Louisiana chose to tie the standard to a minimum dollar amount that would be appropriated for each student.Footnote ¹³⁷ Louisiana included an “equalization fund” designed to direct money to less wealthy, predominantly white and rural counties and parishes that lacked local sources of revenue.Footnote ¹³⁸ South Carolina structured their legislation specifically to pay, and equalize, the salaries of all the teachers in the state for 6 months, provided counties put up the money for an additional month.

To fund these new minimum standards, reformers enacted new taxes that would serve as a dedicated funding stream for the public schools. Louisiana included the new taxes necessary to fund the expansion of state support for public education in the reform bill itself. Long and the legislature structured the bill as a constitutional amendment, which opened up more policy possibilities. One of them was an increase in the state limit on property taxes, allowing the state to levy an additional two and a half mills of property taxation. Louisiana also imposed new taxes on malt liquor and natural resource extraction.Footnote ¹³⁹ South Carolina, in contrast, chose to rely almost entirely on new property taxes in order to fund their education legislation.Footnote ¹⁴⁰ Section two of the legislation stipulated that “To meet the amount provided for in Section 1 of this Act, there is hereby levied upon all the taxable property of each county of this State (4) mills” of additional taxation.Footnote ¹⁴¹

8.4. Linkage between vote choice and the provision of school funding

Did these reforms allow Long to deliver expanded public school funding to his poor, white rural supporters? Put more broadly: Did the political choices of voters in South Carolina and Louisiana have material stakes? To answer this question, I examine whether the voters who supported the governors who enacted funding expansions received larger funding increases for their local schools after the relevant election. I constructed a panel of county- or parish-level, per-capita education spending on white students in both states. Each panel includes 3 years before the election of the reforming governor, and 3 years after the implementation of the reform package—encompassing the period both before and during the Great Depression. I utilize a postelection dummy and interaction term regression model to test whether any subsequent changes in funding patterns are plausibly attributable to the election of a reforming governor. The model specifications, including year and county fixed effects, are as follows:

\begin{align*} \text{Funding per capita}_{it}& = {\unicode{x03B2}}_{0} + {\unicode{x03B2}}_{1}(\text{Post}_{t}) + {\unicode{x03B2}}_{2}(\text{VoteShare}_{i})\\ &\quad + {\unicode{x03B2}}_{3}(\text{Post}_{t} \times \text{VoteShare}_{i}) + {\unicode{x03B1}}_{i} + {\unicode{x03B3}}_{t} + {\unicode{x03B5}}_{it} \end{align*}

Here, FundingPerCapita_it is the county level, per capita spending on white students in a given year, β₀ is the intercept, and β₁(Post₁) is the postelection dummy, which equals 1 for years after the election and 0 for years before. β₂(VoteShare) is the county-level vote share for the winning candidate in the relevant primary. It is time-invariant (the same value for all years within a county), so it is absorbed by county fixed effects αi in models that include them. β₃(Post × VoteShare) is the interaction term and thus the key coefficient in the model. It is the product of the postelection dummy and vote share. β₃ is the central estimate: how much more (or less) funding a county received in the postelection period for each additional unit of vote share, relative to its own pre-election baseline. This is the difference-in-differences coefficient. α_i represents county fixed effects, which absorb all time-invariant differences between counties. γ_t represents year fixed effects, which absorb statewide shocks that affect every county equally in a given year. Lastly, ε_it is an error term, a residual. Tables 2 and 3 contain the results of the regression analysis for Louisiana and South Carolina:

Table 2.

Effect of 1928 Gubernatorial Election Outcomes on Per-Capita White School Funding in Louisiana

Table 2 Long description

Regression results relate per-capita white school funding in dollars to an interaction between the post-election period and Huey Long’s 1928 county vote share, estimated under four specifications. The key interaction term is consistently positive and statistically significant in every model, with an estimate of about 36.145 and standard errors around 12.4 to 13.7. In the model without fixed effects, the post-election indicator alone is negative at about 11.844 and marginally significant, while Long vote share alone is negative at about 56.773 and statistically significant; these main effects are omitted when absorbed by fixed effects in other specifications. Constants range from about 32.784 to 71.903 and are statistically significant across models. Model fit varies by specification: R-squared is highest with two-way fixed effects at 0.759, compared with 0.205 with no fixed effects, 0.414 with year fixed effects only, and 0.549 with county fixed effects only. All models use 384 observations covering 64 counties. Standard errors are clustered at the county level, so statistical significance reflects within-county correlation adjustments.

Notes: Clustered standard errors at the Parish level (HC1) in parentheses. The dependent variable is per-capita white school funding in dollars. Long Vote Share is the county-level vote share for Huey Long in the 1928 Democratic gubernatorial primary. Post-Election Dummy equals 1 for years 1933, 1935, and 1938, and 0 for years 1920, 1921, and 1927. Model (1) is the preferred specification (shaded).

† Post-Election Dummy is absorbed by year fixed effects and omitted.

‡ Long Vote Share is time-invariant and absorbed by county fixed effects and omitted.

*** p < 0.01; **p < 0.05; *p < 0.10

Table 3.

Effect of 1922 Gubernatorial Election Outcomes on Per-Capita White School Funding in South Carolina

Table 3 Long description

Regression results estimate how the winning candidate’s 1922 county vote share relates to later changes in per-capita white school funding, using 276 county-year observations from 46 counties. The main interaction of post-election period with candidate vote share is 10.765 in every specification, with standard errors around 18 to 20, indicating no statistically reliable effect. The post-election indicator is about 5.8 dollars in models without year fixed effects, but it is omitted when year fixed effects are included. Candidate vote share alone is about 32.7 dollars and statistically significant at the 10 percent level in models without county fixed effects, but it is omitted when county fixed effects are included. Model fit varies: R-squared is highest with two-way fixed effects at 0.712, compared with 0.186 with no fixed effects, 0.445 with year fixed effects only, and 0.453 with county fixed effects only. Constants range from about 8.3 to 24.8 dollars, with statistical significance in the two-way and county fixed-effects models. Standard errors are clustered by county, and the key takeaway is that the post-election interaction does not show a robust, statistically significant association across specifications.

Notes: Clustered standard errors at the county level (HC1) in parentheses. The dependent variable is per-capita white school funding in dollars. Candidate Vote Share is the county-level vote share for the winning candidate in the 1922 South Carolina Democratic gubernatorial primary. Post-Election Dummy equals 1 for years 1929, 1933, and 1937, and 0 for years 1920, 1921, and 1922. Model (1) is the preferred specification (shaded). The interaction term is not statistically significant in any specification.

† Post-Election Dummy is absorbed by year fixed effects and omitted.

‡ Candidate Vote Share is time-invariant and absorbed by county fixed effects and omitted.

*** p < 0.01; **p < 0.05; *p < 0.10

The key coefficient is positive and statistically significant for Louisiana. For every ten percentage point increase in a county’s vote share for Huey Long, that county received approximately $3.61 more per capita in funding for white public schools in the years after Long’s election (and the subsequent reform and expansion of Louisiana’s public education system) relative to its own pre-election baseline. This analysis is robust to the inclusion of two-way fixed effects. In South Carolina, by contrast, the key coefficient is not statistically significant. This signifies that there is not a statistically significant relationship between the election results immediately preceding the enactment of South Carolina’s reform of its public school system and subsequent public school funding increases during my period of interest.

8.5. Decreasing inequality and decoupling school funding from local wealth

Prior to these reforms, school funding for whites in both South Carolina and Louisiana was highly unequal. Since most funding came from local property taxes, wealthier counties and parishes were able to appropriate and spend far more per enrolled white pupil than their poorer counterparts. This discrepancy worsened after the onset of authoritarianism in the South, a material reflection of the unequal distribution of political power among whites.Footnote ¹⁴² Moreover, the continuing exclusion of African Americans from the political sphere enabled wealthy counties with large black populations to take, with no fear of political backlash, money nominally apportioned by the state for black students and direct it instead to schools that served whites. The result, shown in Figures 5 and 6, was stark intrawhite funding gaps in Louisiana and South Carolina, driven by a pernicious combination of economic inequality and racialized hoarding of resources. In South Carolina for the 1922 school year (just before the state passed its reform package), Beaufort County, a wealthy, black majority county in the lowcountry, spent \$101.35 per enrolled white pupil. That same year, Cherokee County, a poor majority-white county in the upcountry, spent just \$20.37. In Louisiana for the 1927 school year (just before Long was elected), poor, majority-white Lincoln Parish spent \$29.14 per enrolled white pupil while wealthy, majority-black Tensas Parish spent \$79.08.

Figure 5.

Correlation between White Wealth and School Funding in South Carolina Prereform.

Source: Author’s own analysis based on data from the SC Superintendent.

Figure 6.

Correlation between White Wealth and School Funding in Louisiana Prereform.

Source: Author’s own analysis based on data from the LA Superintendent.

Both states’ reforms decreased intercounty funding inequality, as they directed more money to poorer counties and parishes. However, the scale and aggressiveness of Louisiana’s efforts to close interparish funding gaps far exceeded South Carolina’s more limited efforts. While South Carolina narrowed the gaps in funding that existed between poor and wealthy counties, it was not able to break the relationship between local wealth and school funding. In 1923, prior to the reforms, per capita spending on white students was moderately correlated with how much taxable property existed per white person in the county. In 1938, 15 years after the state had implemented the reforms, that correlation (documented in Figure 7) strengthened (the correlation coefficient rose from .53 to .76), even as its scale shrank somewhat. The poorer the county, the less the local public schools spent per white student.

Figure 7.

Correlation between White Wealth and School Funding in South Carolina Postreform.

Source: Author’s own analysis based on data from the SC Superintendent.

The shifts in Louisiana, visualized in Figure 8, were far more extensive. In 1927, on the eve of Long’s ascension to the governorship, large white landowners in black belt counties appropriated and spent far more per white student than other regions of the state. Education spending correlated with local wealth, just as it did in South Carolina. In particular, the geographic heart of the old populist movement suffered from some of the lowest levels of per capita education funding in the state. The reforms of the Long administration decreased funding inequality (it fell 37 percent between 1927 and 1938) and helped poor, majority white parishes increase their per capita funding. Critically, Long’s reforms succeeded in decoupling the relative amount of school funding available to each parish from the value of the taxable property in that parish. The correlation coefficient between wealth and school funding declined from .51 to just .13.

Figure 8.

Correlation between White Wealth and School Funding in Louisiana Postreform.

Source: author’s own analysis based on data from the LA Superintendent.

The trajectory of education spending in Jackson and Union parishes, both white majority parishes in the state’s north, is illustrative of how this change played out in individual parishes. In 1927, Union and Jackson ranked near the bottom of all Louisiana parishes in both per capita white school funding and per capita taxable property. In 1938, the two parishes remained among the poorest in the state, but now ranked 8th and 24th out of sixty-four parishes in per capita white school funding. Louisiana accomplished this, in part, by having the state take on a greater share of the responsibility for funding schools. By 1938, the Louisiana state government provided about half of all school funding. Parishes and municipalities split the remainder. In the same year, the South Carolina state government only provided about 38 percent of the state’s education funding.

8.6. The great depression and durability of southern education reform

The Great Depression was a calamity for the southern economy.Footnote ¹⁴³ The economic crisis was particularly harmful to the region’s public schools. Property tax revenue that played a huge role in sustaining public schools cratered as land values fell and farmers, short on cash, failed to pay their taxes. How durable would the region’s reforms of its education systems be in the face of economic disruption and heightened scarcity?

Overall, between 1929 and 1934, property tax revenue collected by counties and local districts in South Carolina fell by 25 percent. State appropriations for education in South Carolina only fell by around 7.5 percent during the same time period (Report of South Carolina State Superintendent of Education 1934).Footnote ¹⁴⁴ These declines in revenue led to sharp cuts in spending. In 1934, per capita funding per enrolled student fell 39.6 percent from 1929. Funding remained depressed through the end of the decade. By 1940, per capita funding per attending student was still 25 percent below 1929 levels.

The political and institutional context for the trajectory of South Carolina spending on its public schools is Governor Olin Johnston’s losing battle with South Carolina’s powerful state legislature. I elaborate on Johnston’s struggle against South Carolina’s political institutions in the following section. He was elected in 1934 as a champion of poor upcountry whites in the state (Simon, Reference Simon1998).Footnote ¹⁴⁵ Johnston pledged to shift state resources from highway construction to the public school system. He aimed to raise the salaries of public school teachers. He bemoaned the state of affairs where the state spent tens of millions of dollars to build roads in black-majority parishes dominated by wealthy whites and the public school system only received a few million dollars per year. In a 1935 message to the South Carolina House of Representatives, Johnston wrote,

The South Carolina legislature rejected Johnston’s proposals to shift state resources from the highway commission to state support for the public schools.Footnote ¹⁴⁷ Consequently, he was not able to reshape and redirect education funding to the degree that he wanted to.

Louisiana followed a different trajectory. Tax revenue collected by counties and municipalities only declined by about 16 percent between 1929 and 1934. Thanks to the implementation of the state’s reforms, revenue from the state actually increased 20 percent during the same period. The funding decline in Louisiana, while sharp, was less severe. Due in part to the loss of “nonrevenue receipts” (primarily short-term loans from banks no longer capable of lending), spending fell 27.3 percent for enrolled students from predepression peaks. The central state’s greater role in financing public education helped shield school funding from the declines in local property tax revenue. Consequently, education funding in Louisiana also recovered more rapidly, as state tax revenue bounced back more quickly than local tax collection. The central state was able to tax a wider variety of sectors and activities, providing a shield against the catastrophic revenue shortfalls experienced by counties and localities. Moreover, state tax collectors were far less subject to capture by powerful local elites than local property tax assessors. Reformers all across the South sought to shift funding and tax collection responsibilities from localities to the state for this reason.Footnote ¹⁴⁸ Long’s administration, after the enactment of their initial reform package, passed additional legislation to shift the responsibility for collecting school taxes from localities to the state.Footnote ¹⁴⁹ Louisiana began to appropriate more money per enrolled white student than South Carolina in 1931, and by 1938 Louisiana spent 28 percent more per white student than South Carolina. Louisiana exceeded its previous peak of per capita education funding in 1938, while also achieving a more equal distribution of funding among whites. Counties maintained a far greater degree of their power and fiscal autonomy in South Carolina, reflecting the preferences of the state’s wealthy, landowning elite. A greater degree of political accountability and centralization of funding and tax collection helped Louisiana surpass South Carolina in education funding during this period.

The fact that Louisiana expanded its education funding—usually thought of as a longer-term investment—in the teeth of an economic crisis that caused immediate hardship is one measure of how intensely the state’s rural agrarians desired increased school funding. As the 1930s drew to a close, the Depression notwithstanding, Louisiana was spending more money per capita on white public schools than it ever had before. The state paired this increased outlay of funds with a far more equal distribution of money between wealthy and poor white students. Consequently, education spending in poorer, white parishes increased significantly. In South Carolina, the drop in education funding was much steeper, and the recovery was much slower. The state continued to spend less than it had in the final years of the 1920s, and made no further progress toward a more equal distribution of funding among white students.

9.1. Louisiana: a highway construction program for small-time farmers

Louisiana established a state highway commission in 1910, 6 years before the federal government made it a precondition to receiving federal highway aid. However, it remained an underfunded, administratively threadbare agency for most of the next decade. Consequently, newly available automobiles were largely restricted to the state’s urban areas where more extensive, higher quality roads already existed.Footnote ¹⁵³ That changed in 1921, when newly elected governor John Parker, a good-roads enthusiast with progressive, centralizing instincts, worked with the legislature to increase the Highway Commission’s capacity and give it a dedicated revenue stream. The legislature instituted a gasoline tax and a registration fee on trucks operating within the state. The highway commission would be led by three commissioners, all appointed by the governor. The governor also had the power to designate one of the three commissioners as chair, and to fill any vacancies that arose if a commissioner resigned early or died in office. Crucially, the governor could also remove a commissioner for “inefficiency, malfeasance, or neglect of duty.”Footnote ¹⁵⁴ Taken together, the new Highway Commission constituted a significant expansion of gubernatorial power, as the commissioners effectively served at his pleasure. In insulating the highway commission from the legislature, Louisiana followed patterns of institutional design entrenched in the state’s 1898 constitution.

However, in a concession to conservative opponents of taxes and government debt, the legislature refused to supplement these tax- and fee-based streams with bonded debt. Although the state could now construct and manage a statewide system of public highways, it would do so via a mechanism of “pay as you go.”Footnote ¹⁵⁵ Road building accelerated, with Louisiana putting an average of about five hundred miles of new public highways under construction per year between 1922 and 1928. However, the sluggishness and unequal distribution of the state’s highway construction program created an opening for Huey Long, who promised voters living in poorer, more isolated parishes more and better quality roads (Sindler, Reference Sindler1956: 54).Footnote ¹⁵⁶

In his 1928 gubernatorial campaign, Long promised an improved state highway system alongside his calls for increased school funding.Footnote ¹⁵⁷ During his time in office, he greatly accelerated aggregate highway construction. Louisiana put an average of 1,710 miles of new highways under construction per year—more than three times higher than the state average in the previous 6 years. Long’s administration also improved the quality of public highways. About 29 percent of the roads constructed during Long’s governorship were paved, as opposed to being surfaced with packed dirt or gravel. In the 6 years prior to his term in office, the share of newly constructed highways that was paved was around 10 percent. Highway funding quadrupled from roughly twenty-four million between 1924 and 1928 to a little over ninety-seven million between 1928 and 1932. Louisiana’s agrarian coalition succeeded in greatly increasing the overall provision of essential public infrastructure.Footnote ¹⁵⁸

How was highway construction distributed among Louisiana’s parishes? Did that distribution change after Long’s election? I find that after Long’s election in 1928, the distribution of both total miles of highway construction and state highway funding shifted toward both poorer and whiter parishes that constituted the core of Long’s electoral coalition. As shown in Figure 9, in the 4 years before 1928, black majority parishes received about 29 percent of all new miles of highway construction in Louisiana, despite containing only 15.7 percent of the state’s total population. Between 1928 and 1932, that percentage dropped to 25 percent. A similar shift occurred in the distribution of spending on new highways, with the share of money going to black majority parishes dropping from 25 to 21 percent.

Figure 9.

The Graph on the Left Shows the Correlation between the Share of a Parish’s Population that was Black and Miles of Highway Construction in Louisiana between 1924 and 1928. On the Right Shows the Same Correlation for the Period 1928–1932. Note the Increase in Total Construction and Distributional Shift Toward Whiter Parishes.

Source: author’s analysis of data from LA Highway Commission reports.

Long and his political faction also shifted both new spending and construction away from wealthier parishes and toward poorer ones. In the 4 years before 1928, parishes above the state average in taxable property per white resident received about 28.4 percent of all new miles of highway construction in Louisiana, even though they contained roughly 50 percent of the state’s total population. Between 1928 and 1932, that percentage dropped slightly to 26.5 percent. A more pronounced shift occurred with respect to the distribution of spending on new highways, with the share of money going to the richest parishes dropping from 38.1 to 31 percent. This shift was even more pronounced among the wealthiest quarter of the state’s parishes. In the 4 years prior to 1928, the fifteen wealthiest parishes in the state (out of sixty-four in total) received about 32.5 percent of all of Louisiana’s highway construction dollars. In the 4 years after 1928, that percentage dropped to 22 percent. Between 1928 and 1932, new highway construction and funding were negatively correlated with both the black share of a parish’s population and its per capita wealth.

9.2. South Carolina: accelerated highway construction leaves poor whites behind

South Carolina’s highway construction program developed differently than Louisiana’s. The state’s constitution simultaneously centralized power in a highly gerrymandered state legislature while devolving a considerable amount of taxing authority from the state to the county.Footnote ¹⁵⁹ The result was a highly uneven pattern of highway construction as individual counties (and their legislative delegations) pursued highway construction largely on their own. This took the form of local bills passed by the state legislature that allowed county governments to borrow money and charged them with building and maintaining new highways. The title of a bill passed during the 1921 testifies to the degree of micromanagement the State Senate exercised over counties. The bill was entitled

The highway building programs of larger and more populous counties could be quite substantial. Greenville County, a majority white county in the state’s northwestern upcountry, passed a million-dollar road bond several years before the outbreak of World War I.Footnote ¹⁶¹ Rural white landowners in majority black counties resisted increases in the local taxes necessary to fund highway construction. The result in the aggregate was the inverse of Louisiana. Before 1929, the number of miles of public roads in a county was negatively correlated with the share of the population that was Black. The wealthy white landowners who dominated these counties did not want to pay for public infrastructure, whereas the small farmers and textile factory operators who populated the state’s upcountry were willing to finance public highways.

South Carolina did not establish a highway commission until incentivized to do so by the federal government in 1916.Footnote ¹⁶² When the legislature established the Commission, it ensured that it would not become a source of gubernatorial power and patronage. The initial Highway Commission was composed of five members, with only two appointed by the governor. Throughout the 1920s, the legislature modified the structure of the highway commission several times. It increased the number of commissioners to seven, then to fourteen. However, while the governor now had the power to appoint members of the commission when their terms expired, those appointments had to be approved by the state senate.Footnote ¹⁶³

Unlike in Louisiana, South Carolina’s major expansion of highway construction was a project of the state’s conservative landowners. In 1929, the state’s highway commissioners formulated a proposal to float sixty-five million dollars in highway bonds to provide more financing for statewide highway construction in South Carolina. The proposal was controversial and split the state legislature. The state senate, gerrymandered to favor white elites from black majority counties which stood to benefit from the bond issue, passed the legislation easily. In the lower house, a body less favorable to rural, Black majority counties, opposition was stronger. Figure 10 showcases the geographic split in the support and opposition for the measure. In the end, it passed over significant legislative opposition, 60-41.

Figure 10.

Geographic Breakdown of Legislative Vote on 1929 South Carolina Highway Bond.

Source: proceedings of the South Carolina legislature.

Why was the plan controversial? White majority counties that had already committed resources to highway construction resented having their taxes support state bonds that would disproportionately fund highway construction in the landowner-dominated, black majority lowcountry. John G. Richards, who won the 1926 governor election and supported the legislation, came to office backed by a constituency drawn mostly from white voters from black majority counties. In essence, political leaders from poorer, white majority counties were angry that wealthy conservative elites from black majority counties had engineered a way to get the whole state to pay to build highways in their counties.

Political leaders from the poorer, white majority upcountry were right to be suspicious. As shown in Figure 11, in the 2 years after the passage of the 1929 legislation authorizing the bond issue, overall highway construction accelerated, and the distribution shifted toward black majority counties. Before 1929, black majority counties contained about 44 percent of all the state’s public highways. Between 1929 and 1931, black majority counties received 52 percent of all new miles of public highways constructed in the state. The richest half of South Carolina’s counties, by taxable property per white resident, went from receiving 42 percent of all new construction before the bond issue to 46 percent of all new construction after. Louisiana’s state-driven buildout of its highways benefitted the state’s poorer white agrarians. In South Carolina, increased state involvement in public infrastructure benefitted the wealthiest, most politically influential of the state’s white voters.

Figure 11.

The Graph on the Left Shows the Correlation between the Share of a County’s Population that Was Black and Miles of Highway Construction in South Carolina up Until 1929. The Graph on the Right Shows the Same Correlation for 1930–1931. Note the Sharp Distributional Shift *Away* from White Majority Counties.

Source: author’s analysis of data from SC Highway Department.

Wealthy elites’ use of South Carolina’s fiscal and bureaucratic capacity for their own narrow benefit shaped the rhetoric and priorities of Olin Johnston’s successful 1934 campaign for governor. Johnston had been born in Anderson County and worked in a Spartanburg textile mill in his youth. Both counties were located in South Carolina’s upcountry region, and the demographics and political economy of the area shaped Johnston’s politics. He gained widespread attention for his vociferous opposition, as a member of the state house, to the 1929 bond issue. He ran for governor as a champion of the state’s poorer, white voters. In his first attempt, in 1930, Johnston narrowly lost while assembling a coalition anchored by the state’s poorest voters. He won in 1934 after expanding his coalition and reassuring voters outside his base that he could be trusted.Footnote ¹⁶⁴ Johnston most enthusiastic supporters—poor whites from the upcountry—were angry about the distribution of state resources, including the unfair apportionment of new highway construction funded by the 1929 bond issue. His ultimately successful 1934 campaign included promises to build more roads in poorer, white majority counties and redirect a portion of the funding for highway construction to the state’s struggling public schools (Simon, Reference Simon1998: 135).Footnote ¹⁶⁵ Once elected, Johnston sought to bring the highway commission under his control by firing several commissioners. This touched off a dramatic conflict between the South Carolina legislature—which backed the commissioners and wanted them to remain in their posts—and Johnston. South Carolina’s political institutions ultimately stymied Johnston’s ambitions to reorder the balance of power in the state.Footnote ¹⁶⁶ The highly malapportioned legislature reinstated the original commissioners, and then further weakened the governor’s power over the commission. The state senate, the institutional citadel of entrenched conservative power in the state, also rejected nearly all of Johnston’s legislative initiatives. In one instance, a bill supported by Johnston to limit the workload that could be imposed on any individual millworker (the “stretch out” in industrial parlance) passed the lower house 94-6 only to be killed by the Senate after a brief discussion.Footnote ¹⁶⁷ Unlike Huey Long, Johnston was neither able to enact his legislative agenda nor overcome institutional impediments to his broader program. Johnston and his allies were never able to seriously challenge the legislature during the remainder of his term.

In 1936, 2 years into his term, Johnston made a concerted effort to defeat legislators hostile to his agenda by supporting primary challenges by candidates aligned with him. His efforts made little headway in the heavily malapportioned senate—only one senator hostile to Johnston lost his election. Johnston had more success in the lower house, to the point that soon after the election Johnston believed that his faction in the legislature would be able to elect its favored candidate speaker of the state house. However, in the end divisions within the Johnston faction allowed the conservative candidate for speaker to win by five votes (Miller, Reference Miller1976: 237–243).Footnote ¹⁶⁸ For the last 2 years of his term, Johnston would once again face a legislature organized and led by hostile forces.

Consequently, Johnston failed to shift the distribution of highway construction and was also unable to speed up the slow and unequal recovery in South Carolina’s support for its public schools. Funding for public goods in general, and benefits for poor whites in particular, languished in South Carolina during Johnston’s term. Johnston’s defeat at the hands of the legislature illustrates the power of institutions, rather than technocratic need or the expressed preferences of electoral majorities, in determining the geographic distribution of key public goods in South Carolina.