Introduction
The 1964 Civil Rights Act draws a significant and appropriate level of attention as a crucial landmark in the history of antidiscrimination law in America. It represents the dramatic shift from almost a century of federal legal inertia following the US Supreme Court striking down post–Civil War legislative gains in the Civil Rights Cases of 1883. The interim years saw only incremental progress at the national level, namely a handful of executive orders and the relatively weak 1957 and 1960 Civil Rights Acts. Yet, a parallel thread of that period, the role played by state civil rights statutes, has received scant attention. Not only did several northern states swiftly pass antidiscrimination laws in response to the 1883 judicial invalidation but also three (Massachusetts, Kansas, and New York) had them in place prior to that. In the decades to follow, additional states produced protective legislation. However, despite a recent surge of scholarship that advances understanding of the northern fight for civil rights, we still know little about the function of these laws. Numerous accounts dismiss them as merely symbolic, but without empirical support.
However, it is difficult to argue these statutes were effective without demonstrating if/how citizens or government entities relied upon or enforced them. One indicant of their potency could be the extent to which law enforcement or administrative agencies meaningfully implemented them, a history perhaps impossible now to reconstruct accurately. I have chosen an alternate strategy, taking advantage of systematized legal records to examine whether, and how fruitfully, citizens employed these laws to challenge discriminatory actions through litigation. This research design analyzes racial discrimination cases heard in non-southern state supreme courts from 1906 to 1964, facilitating a comparison of the impact of claims based on statutes versus those based on federal/state constitutional provisions.Footnote 1
My research unveils a small universe of 56 cases heard in northern states in these years, confirming that this arena was hardly a hotbed of activity and progress. Still, this history has yet to be unspooled, and my findings uncover a sharp narrative. At least some citizens took advantage of these policies to combat discrimination, with fewer challenges relying solely on constitutional protections. More importantly, outcomes support the wisdom of that strategy, as plaintiffs armed with statutes typically prevailed, while others generally did not. In the absence of laws, state courts showed little inclination for inferring embedded constitutional rights.
While my quantitative analysis establishes the import of these statutes, my subsequent review of the opinions better illuminates the courts’ general legislative deference. Claims based on statutes tended to elicit faithful judicial application of the provisions, sometimes even broadened in interpretation. When denying statutory-based claims, which represented a minority of outcomes, courts offered seemingly logical explanations, generally pointing out a readily apparent mismatch between the facts at hand and the legislative language. Importantly, this data set includes just one instance of a court striking down a state civil rights law (and even that was an oddity, explained in the following text). By contrast, just a handful of claims made in the absence of statutory support were upheld, and those only on very restricted bases. As the following section illustrates, the story I tell is only a small part of the overall struggle outside of the south. Yet, this investigation reveals a meaningful function for these overlooked state protections in the larger narrative.
Background Scholarship
In this section, I first establish basic context through a summary of literature on the general racial dynamics of non-southern states in (approximately) the first half of the twentieth century. Next, I move to more specific documentation and discussion of state civil rights laws. Finally, I merge this background with the discourse on institutional agents of change and the nature of northern protections, to develop the theoretical underpinning for the research design.
General Northern Context
Starting with the broad backdrop to my inquiry, a significant upsurge of studies in roughly the past 15 years have augmented limited past contributions. Earlier work tended to incorporate northern history solely as a subset of civil rights in general (e.g., Howard Reference Howard1999; Klarman Reference Klarman2004; Litwack Reference Litwack1998; McGerr Reference McGerr2003; Woodward Reference Woodward1974). More recent scholarship, however, provides an exhaustive summary of the scope of northern discrimination, emphasizing its entrenchment and the challenges confronting attempted reforms. Notable examples include Hall Reference Hall2005, Lassiter Reference Lassiter, Lassiter and Crespino2010, Lawson Reference Lawson1991 (an older, but innovative source), Sugrue Reference Sugrue2008, Taylor Reference Taylor2011, and Theoharis Reference Theoharis2018. Synthesizing northern and southern histories, this work authenticates the factual record, documenting the multitude of individuals and groups working for change, and the ways in which that struggle manifested.
Although downplaying the importance of state civil rights laws, these histories and commentaries address two narrative threads that partially inform my research design: the chronology of the northern saga and potential variance by type of discrimination. First, moving beyond the conventional concentration on the late 1950s through 1960s as the peak period for demand and governmental response, more recent accounts trace efforts back to the 1920s and 1930s (Lang Reference Lang2009; Sugrue Reference Sugrue2008), suggesting an evolution that unfolded and strengthened over time. As the Great Migration of the 1920s brought large numbers of African Americans to northern states, both formal and informal discrimination appeared, in turn triggering resistance. However, as Biondi (Reference Biondi2003: 3) points out, “[T]he migration of African Americans to the North and West in the 1940s far surpassed the Great Migration of the World War I era,” fostering intensified discrimination and heightened resistance in those years.
Another crucial trigger to this advancement was World War II, as “Nazi anti-Semitism and racialism destroyed what remained of scientific racism’s political and intellectual respectability among the nation’s elite” (Countryman Reference Countryman2006: 28). African Americans pushed this conceptual progression to reality, seeking “the freedom to perform all of the normal public activities of post-WWII America, including buying ice-cream sodas at the five-and-dime” (Kornbluh Reference Kornbluh, Theoharis and Woodard2003: 201). Taylor (Reference Taylor2011: 3) and numerous others highlight the 1950s Cold War era as the next spark for progress, given the US need to resolve civil rights violations to win “the hearts and minds of nations in Asia, Africa and Latin America.”
Second, concerning variation by context, certain settings are portrayed as more resistant to either official (i.e., state or local policy) or citizen-led (e.g., protests or boycotts) remedies. Typically, the northern narrative is sorted into the realms of public accommodations, schools, employment, and housing. While definitive evidence of significant differences among these categories is lacking, numerous accounts emphasize private, or quasi-public, housing discrimination as the most difficult to tackle, given the entrenched custom and practice of segregated neighborhoods (Boyle Reference Boyle2004; Collins Reference Collins2006; Freund Reference Freund2007; Klarman Reference Klarman2004; Rothstein Reference Rothstein2017; Speltz Reference Speltz2016).Footnote 2
However, Sugrue (Reference Sugrue2008) and others suggest the relative ease of challenging segregation in public accommodations, as it was potentially simpler to demonstrate than unequal housing or employment opportunities. Nevertheless, this is a complicated arena. While gaining the right to dine in previously segregated restaurants might have been comparatively easy, integration of beaches and swimming pools drew some of the angriest white opposition, with Sugrue (Reference Sugrue2008: 138) observing that “the more intimate the institution … the more difficult the fights were.” Biondi (Reference Biondi2003) and Speltz (Reference Speltz2016) echo this point.
Emergence of Antidiscrimination Legislation
This all suggests that the appearance and impact of state civil rights protections may have varied meaningfully over time and by context. As a whole, however, the newer literature on northern discrimination affords minimal significance, and thus pays little attention, to state civil rights statutes. To establish a sense of their number and character, older works provide a better resource, at least in regard to what laws were passed, and when. The most comprehensive catalogs are Murray (Reference Murray1951), Lockard (Reference Lockard1968), and Bardolph (Reference Bardolph1970).Footnote 3 Mangum (Reference Mangum1940) provides a useful history but devotes relatively minimal space to northern efforts.
It can be difficult to reconcile these records, as Murray (Reference Murray1951) organizes by state, Bardolph (Reference Bardolph1970) by year, and Lockard (Reference Lockard1968) and Mangum (Reference Mangum1940) by type of discrimination. However, these sources agree that many of the laws appeared relatively early in this broad period, with 18 statutes passed from 1883 to 1910 alone (Bardolph, Mangum, and Murray all report this number). Some laws adopted within, or just beyond, this period were narrowly tailored, such as an 1899 Indiana prohibition on discriminatory hiring “by posting notices or otherwise” (Murray Reference Murray1951: 147) or Pennsylvania’s 1911 ban on segregated schools. Others were more comprehensive, such as the Michigan statute of 1885 covering all public accommodations and jury service or New York’s similar provision of 1913. While the focus is on protections against racial bias, many of the laws simultaneously shielded creed or religion, and a few incorporated national origin, ancestry, and gender, as evidenced by Connecticut’s broad restrictions on fairness in hiring.
More recent sources offer data updated through the 1960s, but it is unclear whether adoption occurred prior to the federal breakthrough acts of 1964, 1965, and 1968. For example, Collins (Reference Collins2006) reports that 22 states had fair housing laws by 1968, while Chen (Reference Chen2007: 1714) references “more than 50 pieces of civil rights legislation … passed by Northern state legislatures from the 1940’s to the 1960’s.” Both authors conclude that Republican legislative majorities exerted a negative impact on passage of these statutes, while diverse coalitions in support had a positive effect (also see Sugrue [Reference Sugrue2008] on the impact of state-level Republican resistance and Fishel [Reference Fishel1966] for the genesis of Wisconsin’s law).
My inquiry, however, concentrates on the impact of these laws rather than their origins and quantity, and on this point the literature almost universally dismisses them. While Sugrue (Reference Sugrue2008: xviii, 137, 134) acknowledges activists were “optimistic that state civil rights laws could be a powerful tool to challenge segregation,” he refers to the provisions as “watery” and “rarely enforced.” Similarly, on California’s equal protection provisions, Self (Reference Self, Theoharis and Woodard2003: 109, 108) observes, “things looked vastly different, and more hopeful” when the laws were passed, but agrees with criticism of “a naïve faith in legal remedies.” Klarman (Reference Klarman2004: 190) concludes, “[M]ost of the laws had little practical effect.” By contrast, Lockard (Reference Lockard1968), an outlier in this debate, reasoned that it was illogical to assume the impact of discriminatory southern Jim Crow laws, while simply dismissing all northern civil rights protections as inconsequential.
Several authors emphasize how the laws fell short of supporting potential litigation by victims of discrimination. Even in his relatively early volume, Mangum (Reference Mangum1940) recognized the possibility of facilities following the letter of the law by admitting or serving blacks but employing more subtle means of exclusion such as oversalting their food, making it difficult to lodge a legal complaint. Other authors document this sort of “soft” discrimination in various settings. Sugrue (Reference Sugrue2008: 150) further suggests the dampening impact of rulings that “did not award damages sufficient to cover court costs.” And, although it may not have limited the effect of the laws per se, Chen (Reference Chen2007) shows that many of the state fair employment statutes required initial administrative mediation, and some barred litigation altogether.
Importantly, however, what all these conclusions have in common, whether supporting or denying the impact of civil rights laws, is an absence of evidence. In one limited systematic foray into the question, Romero (Reference Romero2009), finds just more than a 50 percent success rate for plaintiffs challenging public accommodations discrimination through state civil rights acts in northern courts from 1907 to 1934, although emphasizing the unwillingness of judges to extend the protections of those acts beyond a narrowly defined public sphere.Footnote 4 Beyond that, there are no systematic investigations.
Viewing through the Legislative/Judicial Dynamics Frame
Therefore, I now turn to scholarship on the legislative/judicial dynamic, to create a framework for my research question. Because the challenge is to document the statutes’ impact, not just their existence, any attempt merely to catalog them will fail to advance the dialog. Rather, my strategy hinges on finding the appropriate venue for assessment. To construct that test, I loosely employ Rosenberg’s (Reference Rosenberg1991) thesis of constrained versus dynamic court models concerning the judiciary’s capacity for social change. While the concept is relevant to a number of policy arenas, its best-known application is to federal efforts to combat southern discrimination. In support of the constrained paradigm, Rosenberg casts courts as the titular “hollow hope,” in comparison to legislative (and, to a lesser extent, executive) actions. While conventional wisdom suggests that historic decisions such as Brown v. Board of Education (1954) exerted a significant impact, this counterargument supports legislative acts as the more meaningful change agents. According to his logic, it is more difficult for judges to “find” rights in constitutions than it is for legislators to declare them within the context of statutes. Furthermore, even when courts do issue declarations of rights, their lack of enforcement authority necessarily hinders the impact of those decisions.
The proposition is significant, and controversial, in its own right. However, as social scientists provide a more complete picture of the civil rights struggle, encompassing the pre-Brown era and the northern states, scholarship might expand Rosenberg’s premise as well. State civil rights laws provide a fruitful setting for an application of the institutional dynamics of change in temporally and geographically broader focus because here we have a case in which scholarship discounts the impact of legislation, without even getting to a judicial/legislative comparison. Yet, while not broaching the non-southern dynamic, Rosenberg hints at the potential for both the strength of state civil rights statutes and a more consequential role for courts in this context. One proposed condition for significant litigation occurs “when legislation supportive of significant social reform has been enacted and courts are asked to interpret it” (31). To this point, however, the lack of attention to these laws, or the assumption of their impotence in the face of institutionalized discrimination, has kept that connection from emerging. This is, in part, attributable to incongruity in labeling northern discrimination.
Many authors tend to focus on the de facto phenomenon in the North, referencing the embedded racism that led to “interlocking patterns of educational inequality and residential exclusion” (Lassiter Reference Lassiter, Lassiter and Crespino2010: 27) and, in turn, economic marginalization. For example, this perspective animates Countryman’s (Reference Countryman2006: 6) discounting of northern civil rights laws, arguing that “by themselves, bans on explicit racial discrimination in employment and government services … could not achieve the kinds of structural changes in the local labor and housing markets that advocates of civil rights liberalism had expected.” While his emphasis is on local efforts in Philadelphia, others apply that basic idea of interred bias more broadly (Burkey Reference Burkey1971; Rothstein Reference Rothstein2017; Theoharis Reference Theoharis, Theoharis and Woodard2003). From a slightly different angle, Theoharis (Reference Theoharis, Lassiter and Crespino2010: 50) challenges the “long imagined dichotomy between legalized Jim Crow segregation in the South and the allegedly non-state-sponsored segregation that existed outside of the South.” Extending that point, Lassiter (Reference Lassiter, Lassiter and Crespino2010: 26) refers to the common de jure/de factor distinction as a “tactical error” that insulated embedded northern practices from effective legislative or judicial remedies by placing them outside the realm of conventionally recognized rights violations.
Applied to Rosenberg’s framework, these arguments suggest both legislatures and courts as an even hollower hope in the northern context, if the expectation was for them to resolve de facto patterns or to foster structural change. However, the northern statutes were not necessarily designed to remedy these factors. Rather, they targeted private but purposeful acts of discrimination, for example workplaces or restaurants that enforced their own whites-only policies. That is quite different from either a de jure example of a southern school or business segregated by state law, or a de facto example of a lunch counter whose customers are all white because the neighborhoods around it are segregated by custom and practice. Sugrue (Reference Sugrue2008: 134) effectively conveys this crucial nuance in the realm of public accommodations, observing, “[T]he exclusion of blacks from hotels, stores, restaurants, and recreation centers in the North operated in a strange gray zone, blurring distinctions between ‘private’ and ‘public.’ Exclusion was the consequence of private actions, sometimes backed by legal sanctions but seldom encoded strictly in the law.”
Thus, while critique of the false de jure/de facto distinction is legally and philosophically compelling, it may represent an oversimplification, ignoring the actual focus of these statutes, which falls somewhere between these two phenomena. In turn, neither did plaintiffs armed with these statutes petition courts to affirm abstract rights against insidious biases. Given their more modest intent, and following Rosenberg’s (Reference Rosenberg1991) predictions, these policies were potentially influential, either on their own merit or as interpreted and applied by courts.
Because a key premise of the constrained argument is that courts can make meaningful decisions only through the direction of strong legislation, I turn that around to use litigation as the venue in which to determine whether these state laws were significant. The evidence that these tribunals were more likely than not to favor challenges premised on state civil rights laws, and to support them more strongly than claims based on other legal provisions, will indicate their significance. To be sure, Rosenberg’s (Reference Rosenberg1991) full thesis concerns a broad, societal vision of policy impact, moving beyond comparison of outputs. My narrower purpose in utilizing the judicial setting is not to conclude whether litigation or legislation more effectively fostered change, but rather the extent to which laws matter in some meaningful context, in this case, a courtroom.
This strategy also allows alternative findings to emerge. One is that northern supreme courts were behaving in a dynamic manner, supporting claims even in the absence of statutory protections. Another is that courts were obstructionist, striking down or failing to enforce existing state laws. Evidence for either of these scenarios would undermine the potency of these northern statutes.
Summary of Context
The strengthening historical record, only briefly summarized in this review, offers an essential backdrop to my narrower focus. Nevertheless, it also provides a challenge, in portraying a context that diminishes the consequence of northern statutory civil rights protections. To be sure, there are signs that their role was limited. Many states adopted policies prior to the periods of strongest resistance, which might mean that they were symbolic gestures rather than meaningful responses to change demands. Furthermore, there is no indication that state governments attempted to educate the public on the substance and implications of these laws, which often included disincentives to litigation. Still, overall conclusions on their role remain anecdotal, and unnecessarily entangled with the challenges of remedying de facto patterns. Given the now dense and sophisticated analysis of the northern civil rights history, it is important to complement it with a systematic review of how, and how successfully, citizens utilized these statutes.
Research Design
The first step toward accomplishing this methodical analysis was to compile a data set of all published opinions in racial discrimination cases from non-southern/border state supreme courts from 1900 to 1964. I employed West’s Decennial Digest (1917, 1927, 1937, 1947, 1957, 1967) to systematize this search, as the Digest references cases by plaintiff’s actual claim, not simply because a phrase (“equal protection,” e.g.) appeared in the opinion. This is a significant difference because, to assess judicial response accurately within a particular legal context, it is crucial to identify cases through the plaintiff’s precise allegation.Footnote 5 Using the Digest’s key numbering system, I selected all cases listed under key numbers 1–15 of the “Civil Rights” heading to capture the state constitutional or statutory claims, and the “discrimination by race, color or condition” key numbers (214–23) under the “Equal Protection” subheading of the “Constitutional Law” heading to capture Fourteenth Amendment and other state constitutional claims.Footnote 6 As noted, many of these state civil rights protections covered other groups besides racial minorities, but because those are not in my area of interest, using these precise key codes effectively eliminates claims based on religion, national origin, and so on.
This protocol identified 76 cases from an array of venues including employment, housing, schools, and public accommodations. Twenty were subsequently rejected for falling outside the scope of this inquiry, and I begin by explaining those exclusions. By disqualifying cases from southern states, challenges to official acts of discrimination (not my topic of interest) generally did not appear in the data. However, in these years, this south/non-south distinction was not always clear-cut and, in 15 cases, the challenge was to an overtly discriminatory law or official act, including mandated public school segregation from a peripheral south state (generally, Oklahoma and Delaware, especially in the early years of the study). I eliminated these from the study.
Another three cases were discarded to avoid the influence of a critical federal precedent. As I report in the following text, my design controls for a potentially strengthening public resistance to discrimination across this period, which could have subtly shaped judicial outcomes. That is quite different, though, from the impact of a particular federal policy or judicial precedent that would have sharply influenced state court decisions, regardless of year and/or the existence of a state law.Footnote 7 For this reason, the study ends just prior to passage of the federal Civil Rights Act in July 1964. As far as clear direction from the US Supreme Court, the civil rights decisions from that body in this period generally concerned only state-mandated discrimination and were thus irrelevant to my cases. However, one decision that was applicable outside of the south occurred in 1948, with the Shelley v. Kraemer opinion concluding that courts could no longer enforce privately negotiated, racially restrictive covenants in sale and rental of property. Therefore, three cases in the original set concerning this precise topic and decided after May 3, 1948, the date of the Shelly decision, were excluded. Finally, I cut two additional cases due to fact sets that veered away from my research question.
This left a final, clean set of 56 decisions, listed chronologically in the supplementary appendix, representing the universe of relevant cases, and replicable using this collection method and with these rules of exclusion. As outlined previously, this small number should be unsurprising. Furthermore, it is highly unlikely that my analysis is missing relevant opinions because, as Gerken (Reference Gerken2004: 19) emphasizes, “all of the decisions of the state’s highest court are typically published.” However, precisely because these comprise the entire record, they are vital to understanding the role these statutes played.Footnote 8
All cases were coded for the dependent, test-independent, and control-independent variables. I then utilized logistic regression to test whether claims based upon state civil rights statutes were more likely to result in antidiscrimination state judicial decisions.Footnote 9 My dependent variable is whether the decision was to strike down the claimed discriminatory practice (“antidiscrimination”) or not (“prodiscrimination”). In most cases, this meant, respectively, that the plaintiff did or did not prevail because almost all were African Americans claiming a violation of their rights. However, in a few cases, the plaintiff was the alleged perpetrator of the discrimination, prosecuted or punished under authority of some state law, and in turn claiming that law to be unconstitutional or in violation of some other state provision.
My test-independent variable is whether the claim was based on a state civil rights statute, and this is most clearly evidenced through the Digest’s presentation protocol. Those that were premised on a statute received the statutory claim code. Those that were not received the nonstatutory claim code. These latter litigations were grounded in some federal and/or state constitutional provision, generally equal protection. While it seems likely, and reasonable, that plaintiffs relying solely on federal constitutional rights would turn to the federal court system, not all did, and I employ these cases as a useful comparison group. To be sure, some cases present a combination of claims, but the opinions clarify that, when statutes are present, this is the controlling factor. If these statutes exerted a significant impact in the judicial context, then my test variable’s coefficient will be positively signed (indicating that statute-based civil rights claims are associated with antidiscrimination judicial decisions) and statistically significant (indicating that the claim/decision pattern is distinguishable from random fluctuation).
Moving to my controls, the next variable is the type of discrimination in question. As noted, evidence on how this factor potentially impacted northern response is limited, but there is some consensus that alleged discrimination in residential housing would be least likely to find judicial relief, whether a protective statute was in place. However, public accommodations were arguably easier to tackle. More simply, for the purposes of my primary question, most of the state civil rights laws did include some provision for equity in public accommodations, so if statutes mattered then claims in that realm would have the strongest chance of success. Overall, the possible values were public accommodations, employment, and residential.Footnote 10 In short, I control for this feature to determine whether potential evidentiary support for the impact of state statutes overall is driven by success in just one or two of these categories. If it were, that would constrain conclusions of their general import.
Finally, each case was coded for the decade in which it occurred, to capture the growing rejection of discrimination in the early to mid-twentieth century. It is impossible to conclude exactly what fostered that change—whether shifting social mores, the evolving precedential background in state and federal courts, or some mix of a variety of forces. The importance of this for my study is that because more states adopted civil rights protections over this period, failure to control for time could falsely inflate the significance of the statutes. If it is the case that there was some sort of diffuse progress toward antidiscrimination decisions, then this variable’s coefficient will be positively signed and statistically significant. By including this control in the same equation as whether the claim was premised on my variable of interest, it will be clear whether statutes exerted an influence independent of the passage of time.
The primary goal of this overall design is to facilitate assessment of whether these laws mattered in a meaningful way, by systematically assessing whether claims based on statutes were more likely successful than not, and if that record was more successful overall than the nonstatutory claims. However, two alternate outcomes would diminish the import of legislation. First, the value of state legislation would be relatively marginalized if we find confirmation of dynamic courts upholding claims even in their absence, suggesting that these tribunals were willing and able to find rights in constitutions. While this outcome would be antithetical both to Rosenberg’s (Reference Rosenberg1991) contention and to recent scholarship on the broadly institutionalized nature of northern discrimination, it is still possible, in the absence of any such analysis to this point.
The other problematic finding for the influence of the statutes would emerge if evidence shows that state courts largely ignored the statutes or even struck them down. This obstructionist scenario would also run counter to the Hollow Hope argument, although it would support narratives that question the impact of these laws. One conceivable legal explanation for such a result, through at least part of my time frame, is judicial adherence to the Lochner v. New York–derived right to contract. Under that principle, for example, courts could have struck down laws prohibiting businesses from barring African American customers as unconstitutional intrusions into private business decisions.Footnote 11
In sum, this inquiry will illuminate the role these statutes played in defining judicial responses to discrimination claims, allowing for a number of alternatives to emerge. The investigation, however, will not stop there as I recognize the subtlety of decisions, in that judges do not simply apply, or decline to apply, a law. After all, their very function is to determine whether, and how, a statute pertains to a given case, an exercise appropriately premised on statutory language and case facts. Furthermore, jurists have the discretion to reign in or expand a law from its original intent. To pursue that question of the nuances of how these laws mattered, I further explore these prospects in the “Opinion Analysis” section.
Statistical Findings
I start the analysis with a review of the basic distribution of cases, as reported in table 1. Most notably, these initial findings emphasize that plaintiffs apparently believed in the power of these laws, since 67.9 percent of all challenges (38 of 56 cases) in the data set proceed on a statutory premise. Chronologically, the cases distribute evenly over the 1900 to 1964 period. As far as type of discrimination, most implicate public accommodations, with residential claims coming in a strong second, and only a few concerning employment. This distribution is not surprising. Because public accommodations were the venue most commonly covered by statutes, that setting would have in turn generated the most suits. Even in the absence of a statute, this sort of discrimination may have appeared the most likely to draw judicial rebuke. By contrast, few state laws protected rights within the context of housing, and there was less expectation of success in challenging residential practices. Finally, as noted, several fair employment laws either barred litigation or required complaints initially be made to a fair employment commission, thus limiting the number of plaintiffs who would have reached the litigation stage. Finally, decisions overall are roughly evenly split between supporting and striking down the alleged discriminatory actions.
TABLE 1. Frequencies
Moving to the causal analysis, table 2 presents results of the determinants of these northern state civil rights decisions. I begin the discussion with how well the equation as a whole explains antidiscriminatory judicial decisions. The model overall, including all independent variables, is modestly successful as demonstrated by the Nagelkerke pseudo R-Square of .26 (which is significant at the. 03 level). This statistic tells us that about a quarter of the “variance” in these decisions is explained by the model, with the probability of this estimate occurring by chance if the model were ineffectual about three times in a hundred.Footnote 12
TABLE 2. Impact on decisions
Note: Dependent variable: decision; case type “public accommodations” serves as the comparison group.
The respective results for the three independent variables add more explanatory depth. Findings for the central test variable, whether the claim was based upon a civil rights statute (Claim) demonstrate that statute-based claims were more likely upheld and the discriminatory action struck down, and this result is statistically significant at the .02 level. In short, this suggests that these statutes did function as estimable weapons against discrimination. Furthermore, this effect is independent of the year in which the case was heard. The time control variable, Decade, is statistically significant at the .02 level, indicating that over the study’s time span (1900–64) these state court decisions grew increasingly supportive of discrimination claims. This likely reflects the aforementioned resistance movement and other factors that slowly changed general public opinion, as clearly explicated in the broad historiography of this topic. Because this is not part of my research question, my findings simply reflect this evolution. However, the fact that both Claim and Decade are significant when included in the same equation shows that the existence of the statue exerted an independent influence on an antidiscrimination outcome. As to these variables’ relative impact, they are very close in their explanatory influence. Although not reported, rerunning this analysis using OLS, the standardized betas are .32 and .26 for Claim and Decade, respectively.Footnote 13
Because logistic regression coefficients are reported in a cumbersome metric (the odds ratio), to get a sense of these variables’ influence on civil rights decisions I calculated the probability of an antidiscriminatory decision for a statutory versus nonstatutory claim, and the probability of an antidiscriminatory decision in the 1960s decade versus the turn of the century decade. For this exercise, I set all other variables at their zero values. I estimate that a statutory claim had a .23 greater probability of an antidiscrimination decision than nonstatutory, and the 1960s decade showed a .38 greater probability of an antidiscrimination decision compared to the turn of the century decade.
Finally, the remainder of the controls for the types of discrimination (residential, employment, public accommodations) collectively show a minimal effect. Because discrimination type is not on a continuum, these categories are tested as dummy variables. Public Accommodations serves as the comparison, or excluded, value to which Residential and Employment are compared. Cases invoking claims of racial bias in the residential realm were relatively least successful (with an antidiscrimination decision estimated at a .07 probability and accommodations and employment at .12 and .22, respectively).Footnote 14 Still, the general lack of statistical significance among these categories mirrors the overall uncertainty in the literature regarding the import of type of discrimination in the northern context. More importantly, their inclusion in the equation did not diminish the overall significance of the import of these statutes in generating antidiscrimination rulings.
Opinion Analysis
While the quantitative findings confirm the impact of these statues, and the aggregated patterns of outcomes, they cannot articulate the path by which judges reached their decisions, or fully illuminate how the laws mattered. To explore this question, I review the content of the opinions, also an underutilized resource.Footnote 15 One advantage of this strategy is to tighten the connection with the Hollow Hope prediction of statutes as the essential spur to judicial protections. Another is to gain a deeper understanding of how these laws performed, or how courts deployed them, in this key setting, engaging with the background literature that discounts these laws as devoid of practical import.
To mine these instructive documents for this shading, I begin with the opinions for challenges premised solely on constitutional provisions. To underscore the significance of state civil rights laws, it is useful to understand what is missing, from the court’s point of view, when a plaintiff’s claim is not based on one of these statutes. There were 18 such cases in my data set, and those grouped mainly into two categories. One cluster concerned the thorny issue of public/private entanglement. Even though I excluded those aforementioned instances of official discrimination, there were several cases in which justices had to rule on whether private acts were effectively of an official nature. Plaintiffs could not invoke state civil rights laws in such situations, if they even existed, precisely because these laws did not contemplate, and so could not remedy, acts of public discrimination.
As a whole, decisions in this group imply a narrow judicial construction of what constitutes public action. Sundry courts failed to find it in the following examples: a whites-only orphanage deeded to city administration (In re: Girard College Trusteeship); a privately run coffee shop housed in a publicly owned parking garage (Wilmington Parking Authority v. Burton; later overturned by the US Supreme Court); or, a municipally chartered development corporation (Dorsey v. Stuyvesant Town Corporation). While other decisions struck down alleged acts of quasi-official discrimination, those opinions are tightly framed. Delaware v. Brown, for example, in which police arrested plaintiff Brown for trespassing while attempting to order a meal at a whites-only restaurant, invoked a complicated legal assessment. In addition to his criminal conviction, Brown challenged the law barring state officials from interfering with private acts of discrimination as a violation of his Fourteenth Amendment rights. As enforced, that law resulted in police officers assisting in the removal of black patrons. The majority opinion labored to conclude that, while the “proprietor may not call upon the state to assist him in enforcing his private policy,” the law did not violate the US Constitution, only striking down the trespassing charge. And, in DeLaYsla v. Publix Theatres Corporation, concerning an equal protection challenge to segregated, balcony-only seating in a movie theater, the court found in favor of the minority plaintiff solely on breach of contract grounds because he had purchased a general admission ticket.
Another significant subset of these nonstatutory decisions is comprised of the seven pre-1948 challenges to private efforts to maintain segregated housing patterns. As none of these plaintiffs was armed with a statute, these claims accentuate the virtual absence of state protections in the housing realm. In all those cases, judges denied Fourteenth Amendment–based challenges to private covenants, usually with little elaboration. The Michigan Supreme Court, in rejecting the plea to apply this constitutional provision to private acts, provided a representative, albeit relatively verbose, decision. The Sipes v. McGhee (630) opinion observed, “[T]he unsettling effect of such a determination by this Court without prior legislative action or a specific Federal mandate would be, in our judgment, improper.” In general, it appears that both legislatures and courts were reticent to confront the entrenched power of residential segregation.
Thus, in the absence of statutory protection, constitutional claims alone failed to evoke judicial inclination to validate plaintiffs’ rights. When courts had the opportunity to strike down an act of discrimination in such cases, they generally did not; when they did, it was on restricted grounds. However, moving to challenges that did hinge on state civil rights laws, decisions present a seemingly straightforward application to private actions, essentially as written. This is illustrated, for instance, in the Washington Supreme Court finding the state civil rights act forbade discrimination at a soda fountain (Goff v. Savage) and in the New York Court of Appeals applying their state act to forbid racially segregated, unequally empowered, labor unions (Railway Mail Association v. Corsi).
More generous, or elastic, interpretations also emerge. For example, the decision in the Pennsylvania case of Everett v. Harron concludes that even though swimming pools were not mentioned on the statutory list of more than 40 types of public accommodations where discrimination was banned, that did not indicate that the legislature meant to exclude them. Similarly, in Erickson v. Sunset Memorial Park, the Minnesota Supreme Court extended a law against housing discrimination to a whites-only cemetery, even though the policy did not mention burial facilities.
Intriguingly, in a few of these statutory cases, judges made a point of justifying decisions through their power of constitutional interpretation. The opinion in Colorado Anti-discrimination Commission v. Case proclaims, “[I]t is the solemn responsibility of the judiciary to fashion a remedy for the violation of a right which is truly inalienable in the event that no remedy has been provided by legislative enactment.” Importantly, however, the claim at hand was based on one of the rare state laws supporting fair housing. Thus, it is critical to keep these sorts of declarations on the importance of constitutional or common law principles in perspective. In this data set, they appear only in cases tethered to a statutory anchor.
There were, of course, cases, albeit in the minority at 34 percent (13 of 38) of the statutory claims, where the decision failed to find the disputed action precluded by the statute. In contrast to the previously mentioned elastic interpretations, these adopted a more restrained approach when the fact set did not exactly match the statute. Examples where the court determined the law could not be stretched to cover the challenged act include a decision that adding a radio to a whites-only Kansas snack bar did not render it a “place of entertainment,” as defined and regulated by the law’s antidiscrimination provisions (Brown v. Sanitary Milk Company). Similarly, the Connecticut Supreme Court, in Faulkner v. Solazzi, found that barbershops refusing to serve black customers did not “possess that peculiar public quality, as places of public accommodation,” that the statute required. Moreover, in the early case of Williams v. Chicago, Rock Island and Pacific Railroad, the court declined to apply a statute prohibiting discrimination in public transportation. Although the conductor barred the plaintiff from the observation car, the opinion pointed out that he was not denied a seat per se. Certainly, these putatively objective applications, with decisions concluding that statutes simply do not apply to the case at hand, could mask a concerted effort effectively to neuter the policies. Still, while interpretation is a subjective exercise, none of the opinions presents a patently arduous justification for dismissal of plaintiff’s claim.
More definitively, and objectively, and with just one telling exception, none of the opinions represented a state supreme court denying a plaintiff’s claim on the basis that the civil rights law violated the state or federal constitution. That exception is O’Meara v. Washington State Board Against Discrimination, where the Washington Supreme Court invalidated a law prohibiting homeowners who held a mortgage financed by the Federal Housing Administration from discriminating in the sale of their home. The decision indicates the law was deficient not for overbreadth but because it failed to include all homeowners in the prohibition.
Conclusions
Employing a research design that advances understanding beyond assumptions, or a simple cataloging of laws, my findings suggest a previously overlooked potency to northern civil rights statutes in this period. State Supreme Court claims based on these protections were significantly likely to result in a decision striking down the challenged act, a record far more successful than for plaintiffs depending solely on constitutional provisions. Furthermore, the majority of cases in the data set were premised upon state laws, implying that these policies provided some incentive to pursue state judicial recourse in the first place. The simultaneous relationship between the year in which a case was heard and the likelihood of an antidiscrimination decision confirmed the abstract, nationwide evolution toward more extensive recognition of civil rights, which may have influenced state court outcomes. Importantly, however, the presence of a state law remained determinative of outcome regardless of year.
My two-step research design was integral to substantiating my claims. While necessarily spare, the quantitative analyses established a statistical criterion for assessing the systematic role these laws played in shaping antidiscrimination decisions. That settled, I was then able to clarify decisional patterns and themes within those parameters, such as judges’ varying willingness to expand statutes to fit novel circumstances. Furthermore, the opinions explicated the significantly lower levels of litigation success for claimants lacking statutory support, demonstrated through an obvious aversion to recognize rights solely in constitutions. In the absence of the data, a discussion of decisions would remain merely anecdotal. However, the data alone would confirm just that the statutes mattered, but not clarify how. In short, my dual method approach offers evidence stronger in its combination than either could standing alone.
As the focus of my research question was the role played by statutes, the judicial setting was primarily a useful venue in which to test the power of legislation. However, it is impossible to downplay the added value the judiciary added to this history, whether simply through the willingness of judges to serve as amenable allies to legislative intent or, more actively, by courts extending statutes to fit case facts, arguably beyond their original scope. More obviously, the judicial challenges, albeit few, show that adoption of these policies did not in and of itself end all prohibited practices. At least some citizens had to turn to courts for relief, but critically armed with these statutes.
If there was a weakness in these laws, it lies more in the small number of citizens who took advantage of their power in litigation than in their substance, or in how courts interpreted them. My findings suggest this aspect as a promising path for future research. Only three factors could have driven this limited set of cases. The background literature on the northern context of discrimination effectively discounts two of those—that the statutes were simply unnecessary because discrimination was not a problem to begin with, or that the statutes were so effective that subsequent litigation was unnecessary.
Instead, the most feasible scenario is that victims of private discrimination were unaware, wary, or simply skeptical of the potentially powerful tools available to them. In this way, my conclusions circle back to the broader literature on the challenge of battling northern discrimination. While my findings correct the record on the role that these statutes played, I remain mindful of Hall’s (Reference Hall2005: 1235) caveat to place all civil rights victories in their larger context, to make it “harder to celebrate as a natural progression of American values … [h]arder to cast as a satisfying morality tale.”
