How does Economic Discrimination of Schedule Castes in India compare with Economic Discrimination of the Black Race in the USA?

May 8, 2024

808

This paper aims to find if there are any similarities between the Eco- nomic Discrimination of African Americans in the US and Schedule Castes in India. While there is a lot of literature about both these topics individ- ually, few have tried to compare the plausibly similar aspects of both. The results indicate that there has been a greater betterment in the conditions of Schedule Castes and Schedule Tribes in India than the Black commu- nity on some socio-economic metrics, including school enrolment, income levels, and arrests. Data sourced from several governmental sources from the USA and India were utilised, normalised, and expressed — in some cases — as a percentage of the population to account for demographic dif- ferences between the two countries. Following the consolidation of data, regression has been utilised to find the correlation between the trends of the two communities, and eventually compare them.

Introduction

The United States and India often find themselves facing similar issues at dif- ferent points in their political and economic trajectories. Some of these issues are new, but most of these are old skeletons that years of development haven’t been able to keep in the cupboard.

One of these skeletons is the ever-lasting problem of race and caste-based discrimination in the US and India respectively. I began writing this paper soon after the widespread Black Lives Matter protests, which stirred a debate in the Indian mainstream media as well as in the academic community about whether similar forms of discrimination manifest themselves in present-day Indian society.

The problem we’re facing is clear, but it’s hard to draw parallels in the problems that two of the world’s most populous countries — with very different economic, social, and political conditions — face. A comparison of arrests, for example, might not account for the very different conditions of policing techniques, crime reporting, etc. This causes objectively perfect comparisons to be virtually impossible. Regardless of that fact, controlling for years and minorities might allow us to get close to a complete understanding of the reasons and repercussions of caste and race-based discrimination. The knowledge of these trends not only allows us to plan out the future for the safe inclusion of these minorities but also allows us the privilege of acknowledging the very real progress that has been made in including minorities so far.

Literature Review

Berreman (1960)¹ is one of the first pieces of Academic literature on the intersection of Caste Discrimination in India and Racial Discrimination in the United States. Rather than conducting quantitative analysis or data regressions, the author takes a more qualitative view, elaborating on how the societies of historically discriminated sections consider themselves and their position. It concludes that racial and caste-based discrimination tends to vary little across religions and that no one appreciates, or recognises, themselves as belonging to a lower section of the society, regardless of the justifications given to them for this dis- parity by their ”superiors” or themselves. While analysing quantitative data, it can be very easy to forget the real human beings reflected, sometimes imper- fectly, by that data. This paper, however, takes the story behind the data into account, and hence, provides a much more holistic analysis.

Wohlstetter and Coleman (1970)² was a breakthrough paper, exclusively analysing wage/income differences between Non-whites and Whites using math- ematics to quantitatively understand if there is a measurable disparity in income for both categories of race. While the paper comes to the conclusion that a wage discrepancy does exist, it may not necessarily depict the true picture of the eco- nomic condition of both types of races in the United States. It points out that merely lessening the wage gap may not give the results desired, i.e. eliminating inequality in the society as a whole. I also use the paper to figure out some methodologies of analysis while also drawing from its diverse conclusions.

Antecol and Bedard (2004)³ goes into the very niche area of focusing chiefly on Black, Mexican, and White males and how the wage gap between the former 2 and the latter can be explained by the two factors of labour force participation and education. The study, by focusing exclusively on these two factors and the male population of three races, comes up with very specific estimates, and explanations of the problem at hand, i.e. Racial Wage gap. It finds that Labour force participation can explain 55-56% of the wage gap between Mexican and White males while the number stands at 44-50% between Black and White males. I use this paper to understand the niche area of Gender-Race inequality. While I am not using gender as a differentiating factor amongst races, the paper succeeds, from my point of view, in establishing quantitative results unlike any other.

Kijima (2006)⁴ endeavours to explain the caste-based wage gap in India using a variety of factors and observes how they explain their relevance to income disparity. One of its unique findings was the fact that geographical disparity plays a considerable role in the differences in living standards amongst Sched- ule Castes (and Schedule Tribes) and Non-SCs. While geography is a factor, it is not the only one, with income differences arising between SCs/STs and Non-SCs/STs, even in the same village. The paper also finds that the return on Education of Schedule Caste households is lower than that non-SCs, amounting to lower household expenditure due to the lower income of SCs, with respect to general castes.

Zacharias and Vakulabharanam (2009)⁵ and Zacharias and Vakulabharanam (2011)⁶ concern themselves with more intricate distinctions within the SC/ST group, covering differences in religion and place of habitat, rural or urban. Both studies, the second one a research paper building upon the first, a working pa- per, find out that over the two years of their study, i.e. 1991 and 2002 conditions haven’t changed much for the discriminated, while getting worse in some cases. Hindu Forward Castes lead the income levels, irrespective of rural or urban areas, with non-Hindus and Other Backward Classes (OBCs) occupying a no- ticeably distant second though considerably safe from last. While I don’t take religion into account in this paper, it has been really interesting to note how there are major differences in the condition of Castes, even as we separate them into broader religious sects.

Rawal and Swaminathan (2011)⁷ takes, as microcosms of the general In- dian layout, some villages, 8 to be precise, in varied states, and then explains their survey and data analysis results. They highlight the fact that there is a severe lack of data about Income disparities between castes, which is also ev- ident in some other papers on this topic. The paper finds that more Dalits, or lower castes, are found in the lower quantile in all villages but one, while under-represented in higher quantiles. The paper has been really insightful in describing some methods that can be used to explain Economic Disparity, some of which I use, while also drawing out reasonable outcomes from the somewhat lacking public data.

Sharma (2015)⁸ focuses on another relatively unexplored control variable in considering the Caste-based wage gap, Crime. Through thorough data analysis, the author reaches the conclusion that cases of crimes against Schedule Castes rose in tandem with income inequality, more so the standard of living. Since the Standard of Living is usually affected by non-body crimes or crimes against property, it was these non-body crimes that tended to rise, while bodily crimes weren’t as effective in predicting standard of living inequality. This paper moti- vated me to take up Crime as a factor in the analysis of Casteist discrimination in India.

Wilson and Rodgers III (2016)⁹ revisits decomposition analysis methods of the mid and late 1900s to re-evaluate the racial and gender income gap in the US. The study finds that while the general Black-White wage gap was higher in 2016 than in 1979, but that doesn’t reflect the full picture. The wage gap expanded in the 1980s, narrowed considerably in the 1990s, but then increased slightly, even through the great recession. The only exception to this conclusion seems to be women, who have been the most racially discriminated against in income terms in the 2007-2015 period when compared to all other periods in the study. I use various factors analysed and mentioned in this paper as control measures of my own to fully understand the factors that might work in the background of Economic Racism.

Data and Methodologies

The data used in this research paper comes from various sources. While the majority of Data comes from Indian and American Governments and Surveys, some essential data is taken from Research Papers as well. Following is the list of main data sources:

Some data was only available in the form of ’.pdf’ or ’.xls’, but, it is con- sidered best practice to have all data in the ’.csv’ format. To get such a uni- form file format, I converted the ’.pdf’ data by hand by using Apple Num- bers and Microsoft Excel while I used https://cloudconvert.com/ to convert ’.xls’ (and similar ’.txt’, ’.xslx’) to ’.csv’. For pre-processing of so collected Data and conversion of ’.dta’ files to ’.csv’, I used Python’s Pandas Library (https://pandas.pydata.org)

Obvious questions arise on the reasoning behind picking the three factors we’ve chosen to explain the income inequality between the minorities and the majority. These factors, namely Reading Proficiency, School Enrolment, and Arrests were chosen due to two main reasons:

Pre-existing literature has proven that there is a causal relationship be- tween education (which we’ve expressed through school enrolment and reading proficiency) and income distribution Tinbergen (1972)¹⁰ and a strong correlation between arrests and income, which in turn are related to edu- cation, amongst other factors LaFree and Drass (1996)¹¹.

In an attempt to find clear avenues for comparison, several factors were considered to allow for a more comprehensive analysis of the problem at hand. Some of these considered factors were reported incidents of police brutality, the predominant political ideology at a given time, welfare schemes, etc. However, for reasons such as lack of comparable data, minimal explanatory effect, and ineffective quantisation of data, a short but put powerful list of these three factors was decided upon.

A note about how I analysed the impact of these factors on income inequality also warrants mention. The normalised data on almost all factors taken into ac- count followed a straight-line trend, either linearly falling or rising with respect to time. This led to the decision to choose Simple Linear Regression to find the relationships between the various factors and then express it comprehensively through the use of the Ordinary Least Squares (OLS) estimator to indicate the values of factor coefficients. The OLS method outputs the coefficients of the line of best fit between income (dependent variable) and the independent variables. These coefficients are, then, further interpreted to decide on the correlation of the independent variable with the dependent variable. Due to the independent variables being causative of income in general allows us the liberty to assume that a correlation between independent and dependent variables can be construed as causality.

Results

Discrimination against the Black Race in the US

This subsection will focus specifically on the Economic and Social discrimina- tion against the Black Race in the United States, using control variables like Education and Crime to predict Income or Wealth levels. Figures 1, 2, 3, 4, 5,6 focus just on trends over the last few years, while other figures try predicting their co-relation, and causation to income.

Figure 1: Source: https://data.census.gov/cedsci/

Figure 1 suggests that there is a disparity of around 30 points between the proficiency in reading simple passages by African American (or black) students and White students, calculated by taking an average of the points scored by the participants in a government census over a 300 point scale.

Literature on this inherent difference in reading proficiency from early school- ing is vast. While some prescribe this difference to the differential access to — and interest in — reading material from a young age Ortiz (1986)¹², others point out that economic and social factors affect the average Black Family’s choice of middle and high school: forcing them to choose worse schools with a homo- geneous mix of other students whose families suffer from similar circumstances Levitt and Dubner (2008).

Figure 2 gives the trend for school admissions of Black and White students over the last 10 years. Here too, there is a wide gap, roughly mimicking the previous figure. What needs to be remembered, however, is that this graph shows the number of students in absolute terms, and not as a percentage of the population. The school enrolment by Percentage of Applicable population is shown in the next graph.

Figure 3 displays the percentage of the population of eligible persons currently enrolled in high school by race. The graph shows that out of the total population of Blacks and Whites under the age of 18, a greater percentage of Black Persons are enrolled in high school than Whites under the same context. The gap here is considerable, coming in at around 8-9 percentage points between Black Students and White Students. The Pearson’s R correlation value of En- rolment in Schools by Percentage of Population and Income of Black Americans is -0.982, showing a negative correlation between education and income, though not showing anything about causality, as shown by (Angrist, J. D. and Pischke, J.-S. (2014). Mastering’ metrics: The path from cause to effect. Princeton uni- versity press).

Figure2: Source: https://nces.ed.gov/ccd/elsi/tableGenerator.aspx

Figure 3: Source: https://nces.ed.gov/ccd/elsi/tableGenerator.aspx and https://www.census.gov/data/tables/time-series/demo/popest/ 2010s-national-detail.html

Figure 4 is the all-important graph depicting individual incomes of the Black and White Race in America. Like the other 2 figures, there is an inexcusable gap in both lines, depicting the inherent difference in wages/income of both races. It is interesting to note that there is an almost constant difference between both the lines, and while they are separated by a huge margin, they roughly mimic each other’s shape. The matching shapes of these graphs is particularly worrisome. It shows that while little to no progress has been made in alleviating the historic income divide between Blacks and White, the median income itself has grown to an extent that percentages would misleadingly hide. A sensible takeaway from this graph, thus, is that the increasingly frequent claims of income disparity being reduced stem from an incomplete understanding of the real trend. Over periods of similar growth, it is likely that Black and White individuals would not be competing in the market on level ground. The White parent of the future would have the means to send their child to a better school than a Black parent, which predictably increases the likelihood of the White child performing better through College and beyond than the Black child, causing the vicious cycle to continue despite no explicit foul-play by any individual.

Figure 5 portrays the number of arrests, in absolute terms, of African Americans and White people over the last 10 years. The graph is meant for the sole purpose of showing how the trends in arrests of blacks and whites mimic each other to a great extent, much like income or reading proficiency. While the reader might find it redundant, it is my duty to repeat that this graph shows the number of arrests in absolute terms, and not as a Percentage of the Population, which is shown in the following graph.

Figure 5: Source: https://www.ojjdp.gov/ojstatbb/crime/ucr.asp? table_in=2

Figure 6: Source: https://www.ojjdp.gov/ojstatbb/crime/ucr. asp?table_in=2 and https://www.kff.org/other/state-indicator/ distribution-by-raceethnicity/

Figure 6 displays the percentage of the population of a race arrested for crimes in the US. While in the previous graph, it was seen that more White Americans were apprehended for crimes in absolute terms, this graph gives a better insight into the fact that a greater percentage of the black population is arrested for crime than the percentage of the white population arrested for crime. The delta between the two graphs also is substantial at around 5-4 percentage points on average over a decade. It is interesting to note, however, that the arrests by % of the population are dropping faster for African Americans than White Americans. Underlying reasons for this accelerated decline in Arrests might stem from the influence of the Black Lives Matter protests — which put social pressure on the police to limit unwarranted, race-driven arrests of Black individuals in the US Phelps et al. (2021)¹³. The Pearson’s R correlation value of Arrests by Percentage of Population and Income of Black Americans is – 0.913, establishing a negative correlation between income and arrests, though not showing any causal effect amongst both the variables, meaning that arrests could be affecting income, or income could be affecting arrests.

Regression Results

To get some detailed insight into the causation effects that Educational En- rolment, Arrests, and Race might have on Income, I decided to run an OLS Regression using equation (1)…

Income = $\beta_{1}Education + \beta_{2}Arrests + \beta_{3}1(White)$

… and obtained the following result table.

Dep. Variable	Income	R-squared:	0.991
Model	OLS	Adj R-squared:	0.988
Method	Least Squares	F-statistic:	341.1

Variable	Coef	Std Error	t-value
Enrolment	2083.0226	369.868	5.632
Arrests	-3542.6400	1149.063	-3.083
Race	9351.5940	2384.630	3.922

Table 1: OLS Regression Results

Table 1 sums up all the important results of the OLS regression conducted on the values. The dependent variable is income, and the independent variables are Arrests, Educational Enrolment, and Race. The values that the reader should especially keep in mind are the t value, the standard error, the coefficient, the R- squared, and the Adjusted R-squared of the model. The R-squared and the Adj R-squared show the efficacy of the model, i.e. the capability of the independent variable to predict the dependent variable. For both cases, a higher (closer to 1) value is better. Both the values are very close to 1, hence, we can say that Arrests, Enrolment in high school, and Race are good predictors of income. Now coming to the coefficient values of all the independent variables, the only major thing that we need to remember is that a negative coefficient means a negative causation between the independent and dependent variable, i.e. an increase in the independent variable would result in a decrease in the dependent variable. The opposite is true in the case of positive coefficients. As seen in the table, only arrests seem to show a negative coefficient, backing up the intuitive argument that an increase in Arrests would result in a decrease in Income. For the other variables of Race and high school enrolment, there is a positive coefficient, again, supporting the hypothesis that a higher school enrolment % would result in a higher income, and that belonging to the white race (which is beta 3 in equation (1)) would lead to a higher income. Next, we can look at the ’t’ value, which is obtained by equation (2)

$t = \frac{Coef}{Std Error}$

We usually check the t value to find out if the results we obtained are sig- nificant enough to consider their independent variable as predictive or not. As a best practice, the absolute value of the t value should be greater than 2 to be considered significant. As we can see, then, that the absolute values of all the t values are greater than 2, we can consider every involved variable to be a good predictor of the Income.

In summation, table 1 provides substantial evidence that:

Belonging to the White Race results in some inherent benefit in income.
The variables of Educational Enrolment, Arrests, and Race, are good pre- dictors of Income in the United States, especially between African Amer- icans and White Americans.

Discrimination against Schedule Castes in India

Figure 7 shows the number of crimes against people belonging to Schedule Castes compared to total crimes under the IPC (Indian Penal Code) for that year. While both variables are kept on two different y-axes, it is important to note that, unlike most other graphs in this paper, there seems to be no mimicking or similarity in the shape of both curves, showing that there is minimal co-relation between both values. It is worth noting that one of the many avenues where the nature of the discrimination faced by African Americans in the US differs from that faced by SC, STs in India is social attitude. It is more frequent for a member of the SC community to be abused, assaulted, or lynched due to their caste than for a Black person to have to face direct violence due to their race. While the nature of American discrimination is systemic rather than personal, it is almost the opposite in India, where political empowerment precedes social equality.

Figure 7: Source: https://ncrb.gov.in/en/ crime-in-india-table-addtional-table-and-chapter-contents and https://ncrb.gov.in/en/crime-india-year-2013

Figure 8 shows the enrolment of eligible students to school, with the blue curve representing the total enrolment for that year, and the orange curve rep- resenting the enrolment of Schedule Caste Students in school. Even with a des- ignated quota in schools, the volume of SC students enrolling in schools doesn’t seem to increase as much as the total eligible population of students. That being said, the curve is not in the negative direction, nor is it straight, proving the fact that some improvement in enrolment of Schedule Caste students has been made over the previous decade.

	Median Income by Caste Groups
Group	1991	2002	% change
Schedule Castes	14550	20300	40
Total	25300	29850	18

Table 2: Source: Zacharias and Vakulabharanam (2011)⁶

Table 2 displays the wealth of Schedule Castes and all other groups. The data regarding this, which is admittedly sparse, has been calculated from⁶ which provides a more diverse array of data that is not required for this specific analysis.

As shown by the last column of the table, there has been a greater change (increase) in the median wealth of schedule castes than the median wealth of all other groups, consolidated. That being said, the median wealth of SCs in 2002 was almost 25% less than that of other castes in 1991, a stark difference to say the least. A rounded change (increase) of 18% is seen in the median wealth of caste groups except Schedule Castes, less than half the percentage point increase seen for SCs, at 40%, but substantial nevertheless.

Parallels between Discrimination of Schedule Castes and African Americans

Now that we have interpreted the trends of Education, Crime, and Income of both the groups in question, it is now easy to consider how they both pan out against each other. Using the results in the above sections with the assistance of insights from⁹, it seems to converge on the conclusion that while certain improvements have been made it racial parity, there is still a greater degree of discrimination between races in the US than Castes in India, while improvements are being made over time. It was, however, interesting to note that while arrests by PoP (Percentage of Population) in the US, as shown by figure 6 seems to show that a greater percentage of Black people are arrested, there were more African Americans enrolled in Schools by PoP than White people in the same context, as shown by figure 3. This goes against the general conception that I, and many others, had about more White students going to school, hence their lower arrest rates. This seems to be a false hypothesis, perhaps because arrests and enrolment in schools are not co-related and non-causal, leaving room for some other unexplored vari- able, one of which might be inherent racism in the policing and arresting system. One must remember, however, that mere racism in the policing system, while capable of playing a large role in greater arrests despite more school enrolment, cannot be the sole reason for the inconsistency and/or dichotomy. One thing that might play a, perhaps minor, role in explaining the aforesaid inconsistency is the proficiency or knowledge levels of a student in school, which as shown by figure 1, greatly varies by race, at an average of around 30 points.

One method to explain the fairly noticeable gap in the wealth of SCs and the non-SC population, although not empirically tested, may be attributed to the very mindset of SC adults and/or parents. The main occupation among the Schedule Castes in India is (Gang, I. N., Sen, K., and Yun, M.-S. (2017). Is caste destiny? occupational diversification among dalits in rural india. The European Journal of Devel- opment Research, 29(2):476–492.) agriculture and as proven above, there is a very substantial disparity between their wealth and those not belonging to SCs. Hence, the parents of SC children would usually consider it more beneficial for them, their family, and even their children, to engage them in agriculture and agriculture-related fields, rather than sending these children for any formal education. This leads to a vicious cycle, where each generation of SCs is disadvantaged from birth and are engaged in agriculture (a volatile profession in the Indian economy) rather than obtaining formal education anywhere, leading, again, to lower income/wealth, and the cycle re- peats. I must remind the reader, again, that this is a hypothesis, rather than a proven (or provable) observation.

Scope for Further Research

Despite this attempt at a comprehensive, loss-minimizing analysis of the economic divides that exist between the considered communities, certain lim- itations persist. Future research will need to, primarily, utilise first-hand or extensive second-hand data that:

Exists for both the communities across similar explanatory factors.
Is continuous across the considered time period.

Furthermore, the scope of this paper was limited by a myopic view of direct influences of a select-few factors on income. Underlying — yet influential — metrics such as changes in percentage of population might also be considered to account for demographic conditions affecting larger economic trends.

Alongside the inclusion of newer explanatory factors, the extent of the influ- ence of existing factors can be better judged by considering graduation rates, common college-majors, family cohesion and size, amongst others. A broader analysis of the tried-and-tested predictors can lead to concrete takeaways for policy-makers that need clear directions on where to focus their attention to alleviate systemic racism and caste-ism.

Conclusion

Through the table 1, it seems fair to say that Income, in America, does, in fact, vary inherently from race to race, in addition to changes due to other variables like High School Enrolment, Arrests, and perhaps even Educational Proficiency. One can think of the same ideals applying to Schedule Castes in In- dia, although any material regression or evaluation is really hard due to a severe scarcity of publicly available data. Considering the aforementioned statement, I hypothesise that there has been a greater improvement in the Economic dispar- ity between the Schedule and General Castes in India than between the Black and White races in the USA. I derive this belief from an observation of the yearly trends of our income-describing factors, and income itself. Through the latter’s trend in both the nations, the improvement in median income of SCs, etc is significantly better when compared to blacks. The income growth among SCs was more than twice that of the total population of India, whereas growth in the income of blacks mimics that of the white race. According to the data I have utilised, however, I think it is safe to consider that the initial wealth differences between African Americans and White Americans were greater than between SCs and GCs. We also found that Educational Enrolment and Arrests are good predictors of Income, at least in the US.

References

G. D. Berreman, American Journal of Sociology, 1960, 66, 120–127. [↩]
A. Wohlstetter and S. B. Coleman, Race Differences in Income, Rand
Corporation, 1970. [↩]
H. Antecol and K. Bedard, The Journal of Human Resources, 2001, XXXIX,
year. [↩]
Y. Kijima, Economic Development and Cultural Change, 2006, 54, 369–
404. [↩]
A. Zacharias and V. Vakulabharanam, Caste and Wealth Inequality in India,
2009, https://papers.ssrn.com/abstract=1410660. [↩]
A. Zacharias and V. Vakulabharanam, World Development, 2011, 39, 1820–
1833. [↩] [↩] [↩]
V. Rawal, RAS | Review of Agrar-ian Studies, 2011, https://ras.org.in/income inequality and caste in village india. [↩]
S. Sharma, Journal of Comparative Economics, 2015, 43, 204–226. [↩]
Wilson, Black-white wage gaps expand with rising wage inequal-ity | Economic Policy Institute, 2016, https://www.epi.org/publication/black-white-wage-gaps-expand-with-rising-wage-inequality/. [↩] [↩]
J. Tinbergen, THE IMPACT OF EDUCATION ON INCOME DISTRIBU-TION – Tinbergen – 1972 – Review of Income and Wealth – Wiley Online Library, 1972, https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1475-4991.1972.tb00865.x. [↩]
Lafree, SafetyLit: The Effect of Changes in Interracial Income In-equality and Educational Attainment on Changes in Arrest Rates for African-Americans and Whites, 1957 to 1990, 1996, https://www.safetylit.org/citations/index.php?fuseaction=citations.viewdetails&citationIds[]=citjournalarticle 418304 38. [↩]
Ortiz, American Journal of Education, 1986, 95, 58–76. [↩]
Phelps, “We’re Still Dying Quicker Than We Can Effect Change”: BlackLivesMatter and the Limits of 21st-Century Policing Reform — Experts@Minnesota, 2021, https://experts.umn.edu/en/
publications/were-still-dying-quicker-than-we-can-effect-change-blacklivesmatt. [↩]

LEAVE A REPLY Cancel reply

POPULAR CATEGORIES

NAVIGATION

ABOUT US