Abstract

Scams involving university degrees are flourishing in many emerging markets. Using a resume experiment in India, this paper studies the impact of gray degrees, or potentially bought academic credentials from questionable universities, on callback rates to job applications. The experiment varied the type of degree (no, gray, and authentic) in online applications to entry-level jobs that require no university qualification. We find that gray degrees increase callback rates by 42 percent or 8 percentage points relative to having no degree. However, we also document that gray degrees fare on average worse than authentic degrees. These empirical patterns are consistent with a model where employers have beliefs about the authenticity of degrees and are discounting gray-degree universities probabilistically. We discuss our findings with respect to the Indian context.

1.  Introduction

It is an open secret that academic degrees can be bought in India. Local media have widely reported on this phenomenon with headlines ranging from “Degrees on sale: Jaipur study centers offer bachelor degree to PhD for money” (Kumar, Khan, and Khan 2011) to “Fake degree scam: No sweat, you can get a university degree in 10 days” (Ullas and Prasher 2013), as well as “PhDs, Bachelor's degrees on sale in Punjab” (Chowdhary 2011).

How are degrees bought? To answer this question, we collected qualitative data through a local market review, as well as through interviews with agents and potential buyers in the state of West Bengal. Agents and intermediaries handle (parts of) the process. They advertise their services via local newspapers, the Internet, flyers, and railway coaches, often using ambiguous language due to the illegality of such services (see one such example in figure 1). Most of the advertised degrees originate from privately funded universities in other states and often by means of distance education.1 Students can obtain degrees within as little as two months without even sitting in exams. However, the market and the offered service packages are very diverse. For instance, one interviewed agent summarized the service as follows:
Figure 1.

Example of an Online Advertisement to Buy Degrees

Notes: This online advertisement is presented as an example and was not used in the experiment. Contact details were blanked. Here, 10th refers to 10th standard, i.e., the board exam after 10 years of studies, and 12th refers to 12th standard, i.e., the next board exam after 12 years of study. Students typically sit in the 10th and 12th standard exams at the age of 16 and 18 years, respectively. The categorization of universities is geographical, clustered in North and South India. Original ad can be accessed at www.clickindia.com/detail.php?id=136428806.

Figure 1.

Example of an Online Advertisement to Buy Degrees

Notes: This online advertisement is presented as an example and was not used in the experiment. Contact details were blanked. Here, 10th refers to 10th standard, i.e., the board exam after 10 years of studies, and 12th refers to 12th standard, i.e., the next board exam after 12 years of study. Students typically sit in the 10th and 12th standard exams at the age of 16 and 18 years, respectively. The categorization of universities is geographical, clustered in North and South India. Original ad can be accessed at www.clickindia.com/detail.php?id=136428806.

We manage her [the student's] signature much before the exam on the answer sheet […]. During exam days she can send anybody to sit in the exam. The only requirement is that the person taking the exam needs to be female if the original student is female. She may wish to write something or not. We manage a certificate. [translated from Bengali by the authors, interview dated 27 July 2016]

In this paper, we focus on the most straightforward rationale for buying a degree, namely, boosting job market success as proxied by callback rates to applications. Employers cannot perfectly differentiate between potentially bought and authentic or more credible credentials. However, they are likely to discount degrees from institutions that offer both authentic and bought degrees, as they cannot tell whether a student has gained human capital. This paper tries to quantify the extent to which this is the case. To the best of our knowledge, there has been no quantitative/experimental work on gray degrees in India or other developing countries. There are also no reliable data on the functioning, size, and extent of the market other than anecdotal evidence. More broadly, we contribute to the literature on the value of different types of college degrees on the labor market. Notably, Darolia et al. (2015) and Deming et al. (2016) performed large correspondence studies to examine the value of for-profit college degrees in the United States. Darolia et al. found no effect of such degrees relative to community college degrees or even to having no college degree at all. In contrast, Deming et al. document a negative effect for some online degrees compared to nonselective public institutions.

We focus on questionable institutions where it is possible to buy degrees. These universities also issue valid and legal degrees. In what follows, we therefore refer to these as gray degrees. Informed by a simple conceptual model, we examine whether employers discount such questionable degrees in applications to low-skilled, entry-level jobs advertised on online platforms.2 Preparatory qualitative work informed the design of the experiment, which took place in the state of West Bengal. We first identified several universities from which gray degrees could easily be bought and we picked three of these. These universities tend to be distant from the local job market (West Bengal) and are clearly questionable; for instance, a simple Google search would reveal that at least one of them was involved in a degree scam. We also searched for comparably low-ranked local (control) universities that issue strictly authentic degrees (details on the choice of institutions are given below). We then performed a resume study similar in procedure to those found in the broader labor economics literature (Bertrand and Mullainathan 2004; Carlsson and Rooth 2007; Correll, Benard, and Paik 2007; Pager, Bonikowski, and Western 2009; Kaas and Manger 2012). We picked job advertisements in sectors that required neither specific skill training nor work experience or academic degrees. This is the relevant segment of the job market for our analysis, allowing us to test the impact of gray degrees on callback rates compared with both having no degree and having an authentic degree (from universities that clearly do not issue gray degrees). In other words, we can examine how gray degrees boost employment chances compared with no degrees, and whether they are discounted by employers.3 In practice, we sent three resumes to 132 identified job openings, varying the type of degree. To assess the influence of gender, we sent female and male resumes to mixed-sector jobs, and only female (male) resumes to female (male)-dominated sector jobs. We then recorded callback rates for interviews.

To preview our results: Resumes featuring a gray degree average 8 percentage points more callbacks than those with no degree. This amounts to a 42 percent increase in the number of callbacks. This difference is significant only for and driven by female applicants. These gendered patterns may be specific to the sampled jobs (and gendered industry differences) as well as the Indian context. That said, at least qualitatively we see similar patterns for men and women with an advantage of gray degrees compared to no degrees. Authentic degrees always fare better than no degrees (in particular among female applicants) and on average than gray ones. Note that these main findings are robust to controlling for sector, firm, and resume/profile fixed effects. In a heterogeneity analysis, we also classified gray degrees into low-, medium- and high-ranked degrees based on the perceived ease of acquiring these degrees according to agents and buyers in qualitative interviews. The lowest ranked university has been involved in scams as discussed on the Internet. We see a corresponding empirical ranking of gray degrees. Reporting a low-ranked degree is statistically the same as reporting no degree at all. Conversely, there are no statistical differences in callback rates between authentic and high-ranked gray-degree universities. Our results imply that employers have consistently formed beliefs—that is, they appropriately discount questionable degrees as a function of degree rank.

We rationalize these results in three ways: First, a degree from a questionable university is better than no degree at all. This finding is compatible with a simple model (outlined in section 2) where employers cannot verify the authenticity of a degree from a questionable university but at least assign some corresponding probability. Second, authentic degrees are clearly preferred by employers compared to no or gray degrees, as indicated by the highest rate of callbacks. Taken together, these findings indicate, at least in part, that the veracity of degrees is judged well by employers. Third, we can understand our findings in light of the literature on fake degrees (Brown 2006; Grolleau, Lakhal, and Mzoughi 2008; Attewell and Domina 2011). Attewell and Domina (2011, p. 59) argue “those who are blocked from attaining degrees through normal means are those most likely to employ false credentials.”4 This is backed by what one potential buyer of a gray degree stated during our qualitative interviews:

I need to take care of my baby, cook, collect water and take care of my parents-in-law … I simply do not have time to study. The duty of a married woman is to take care of her household. If I start studying, who will take care of the household? [translated from Bengali, interview dated 18 July 2016]

The remainder of this paper is structured as follows: Section 2 gives a simple conceptual model, details some qualitative insights, and describes the experiment. Section 3 presents the results. Section 4 discusses the findings, and section 5 concludes with some policy implications.

2.  Empirical Strategy

In this section, we first briefly outline a conceptual model to motivate the experiment and to provide some predictions. Then we outline the experimental design as informed by qualitative information.

Conceptual Model

Applicants report degrees or no degrees on their resumes.5 Universities, denoted as i, vary in their quality Qi. A fraction (1 − xi) of reported degrees is “bought.” This fraction captures that it is possible to both buy and legitimately earn a degree from a gray university. Employers evaluate resumes and then decide whether to interview the candidate at unit cost c. They hold beliefs about the fraction or prevalence of earned degrees xi for each institution. The unit quality of the degree Qi positively correlates with the quality of the candidate. The lower the fraction of bought degrees (1 − xi) and the higher Qi, the higher the likelihood of a callback. In sum, callback probability or expected productivity minus interview costs can be written as yi = xiQi − ci.

In this simple model, differences in callback rates stem from employers’ perceived differences in xi. Say the employer decides between two candidates from universities j and k, so she compares xjQj − cj and xkQk − ck. If we assume equal institutional quality, Qj = Qk and interview costs cj = ck; further, if university k is associated with a lower prevalence of bought degrees so that (1 − xj) > (1 − xk), then university k will be preferred over j. As long as the fraction of bought degrees is less than one, then some degree will be better than none. To sum up, the results from our experiment below are interpretable in terms of beliefs. Employers assign a probability that a degree from a gray institution is “bought” rather than earned. As a result, authentic degrees should receive most callback rates, followed by gray and then no degrees.

Overview of the Experiment

Our aim was to test the impact of those universities from which it is possible to buy gray degrees on callback rates to job applications relative to authentic degree-providing colleges and having no degree at all. Based on insights from qualitative data collection efforts, we ran a resume experiment from July to September 2016 in West Bengal. We sent 396 applications to 132 job postings. We applied to each of the job postings using three generic male or female applicant profiles. For each job posting, the three profiles were arbitrarily distributed to the following three groups: (1) no university degree; (2) gray university degree; and (3) authentic university degree. Our minimum target sample was at least 188 applications.6 We budgeted fieldwork time to achieve this target and in practice we exceeded it.

Background

The design of the experiment required input from qualitative interviews, because there is little information on the functioning of the gray-degree market. The qualitative data were collected in two steps by the first author in the state of West Bengal: First, we reviewed the local gray-degree market and looked for flyers to identify universities from which degrees could be bought. Second, we did qualitative interviews using snowball sampling with sellers and agents in the gray-degree market, as well as with prospective buyers (for details on the ethnographic methodology and results, see Majilla 2016). More specifically, we gathered advertisements for suspect degrees and approached agents pretending to be prospective buyers. We also interviewed actual and potential buyers and accompanied them to the agents. In total, five interviews were done with agents and twenty-three with potential and actual buyers.

How does this gray-degree market function in practice in our study area? A university may have a distance education department, which is legally entitled to provide degrees under the name of the university. A university in a given state (private universities are always state universities—i.e., they are established and governed by laws of the state government) can have study centers across the state. Federally funded universities can also have study centers in other states. However, we did not find any federally funded university providing gray degrees. Study centers have links with agents/offices located in other states, most often located in state capitals. These offices have no legal authority to provide services to students. Recall that only proper study centers within the state have such legal authority. Study centers and these offices located in other state capitals cooperate with agents at the district level, and these in turn may work with agents at the block level.7 Official study centers are meant to provide learning support and also administer exams in the presence of invigilators from the central university. Yet, in reality, we found that buyers of gray degrees never actually seem to visit these real and legal study centers.

We found that in most cases students contact local agents, who forward applications and requests to their contacts at a higher level within the state. Such agents are not common for authentic colleges and students in India. Of course, legitimate and official educational consultants do exist, and their primary customers are students looking for engineering degrees in other states and students looking for foreign degrees. These consultants are merely advising students, but they rarely have any influence on the admissions process. Consultants for arts or social science degrees are less common. The Indian education system is very much prestige-driven and students try to enter the most prestigious colleges they can within their reach and budget.

We found that gray degrees are “earned” through aforementioned informal agents. They not only guide students, they manage the entire “study” process and have substantial influence on the university administration. They literally sell their ability to arrange gray degrees. In India, most undergraduate arts and social science degrees take three years and students sit in exams, either in semester mode or on a yearly basis after their coursework. Gray degrees in principle require coursework but if they are channeled through the distance education mode, then actual attendance in coursework is optional. We also found some agents who arrange gray degrees through full-time mode. We found that most students do not actually sit in exams. In fact, most degrees are acquired through this distance education system. If students do need to sit in exams (as in the case of bachelor of law degrees), they just sign the answer sheet and somebody fills in responses for them; or sometimes answers are provided to them and they copy them onto the exam sheet.

In terms of academic streams, we found arts and social science gray degrees to be most prevalent as these are easier to manipulate. These degrees are cheaper than degrees in science subjects. We found most students were buying gray degrees either (and equally so) at the bachelor's or master's level. We found that students can obtain undergraduate gray degrees within as little as two months. However, it is also important to note that many different business models and service packages exist in this market and one can naturally only capture a small glimpse of these during qualitative research.

During the qualitative work, we identified several universities from which it was possible to (indirectly) buy gray degrees. Out of these we selected three universities that were popular among local agents. Our aim is not to “name and shame” these particular three universities, so we do not disclose their identities here. Details are of course available on request from the authors for replication and research purposes. These three universities are in three different cities and in three different regions. It is important to underline that all three universities also offer authentic degrees, earned through actual academic work. However, it is possible to easily access and buy these questionable or gray degrees with the help of agents. In these cases, students do not write exams and agents take care of the entire process. More specifically, all gray degrees used in our experiment can be accessed through distance learning. The chosen degrees cost around INR 18,000–20,000 (∼USD 275–300). Note that costs are roughly similar across the three ranks. We would also like to note that the three universities tend to be far from the applicant's residence and the job market. For instance, the most questionable gray degree comes from a university that is roughly 1,500 km away from the home address of our job applicants.8 This is important because the qualitative data indicate that agents tend to collaborate with distant universities. Degrees from such universities should be discounted by employers. A local applicant could obtain an authentic degree from a nearby university.

We also ranked these three questionable universities. We did this based on the perceptions of agents. The lowest ranked gray degree comes from a university that easily cooperates with agents and has been involved in scams as documented on Web sites. Naturally, there are no hard quality measures or measures on the prevalence of bought degrees available, so rank-specific results have to be interpreted with caution. We need to assume that the difference in callback rates only stems from the relative prevalence of and beliefs about illegitimate degrees (as suggested by our conceptual framework).

Selection of Authentic (Control) Degrees

Recall that we compare resumes featuring gray degrees to both resumes with authentic university degrees and only high school degrees. Our aim is to compare institutions that are of similar quality, so that differences in callback rates can be attributed to the existence of a market for bought degrees.

To this end, we prepared a list of colleges in the state of West Bengal offering authentic degrees, which we thought would be comparable to the identified universities offering gray degrees. We had a four-step strategy to select those used in the experiment. First, we showed our list of authentic institutions and those known for gray degrees to a human resource consultancy firm and asked a representative to rank and match institutes in terms of job market prospects. Second, we repeated the same exercise with two gray-degree agents operating in the local market. Third, we asked several buyers to examine our list. However, their knowledge of gray-degree universities (apart from their own) turned out to be limited and we did not use their input. In the end, we had a list of five comparable authentic and five gray-degree institutions and we picked three of each for the experiment.

Further Education Characteristics

We did not mention on resumes whether degrees were obtained through distance education or not (which is often the case for gray degrees). Mentioning distance learning just for gray degrees may have confounded our results. However, one limitation is that we cannot test for the impact of reporting distance education on callback rates, which may differ between authentic and gray degrees, and which may provide an additional signal to the employer. Further, we had to make a choice on the academic discipline itself. Based on insights from the qualitative interviews, we opted for two comparable and relatively less prestigious academic disciplines in the Indian context, namely, history and political science. Finally, all resumes featured comparable secondary and high school degrees located near the applicants’ residential address.

Selection of Job Postings and Location

We used three popular online portals in India to search for jobs.9 These portals allow applicants to create a profile, upload a resume, and apply to job postings. We made sure not to apply to the same job twice across portals.

We sent applications to entry-level and relatively low-skilled jobs. We only selected job postings that were open to both inexperienced and experienced applicants. As these jobs did not require a university degree,10 we can test whether gray degrees provide a competitive advantage over having no degree at all. More specifically, we picked job postings ranging from sales, in particular insurance or personal loan sales, to administrative support, clerical support, call centers, and medical representatives. We did not design the experiment to be able to detect impacts by specific sectors. Rather, we simply selected and classified jobs into female (N = 99), male (N = 99), and mixed-sector jobs (N = 198). We gendered the resumes of applicants according to this threefold classification. This allowed us to examine patterns by the gender of the applicant and the gender of the sector. For instance, administrative jobs are typically held by female employees, whereas medical sales representatives are mainly male. Conversely, call centers are known to have a mixed work force. One limitation of our design and the small sample is that we cannot clearly disentangle gender differences (if any) from job and industry differences.

We exclusively applied to jobs in the state of West Bengal (in east India), where the qualitative fieldwork also took place. We mainly applied to jobs based in Kolkata, and in some cases to jobs in district towns within the state of West Bengal. Relatedly, we excluded job postings that would have required applicants to relocate. Our experiment closely mirrors what we found in the qualitative interviews with agents and buyers. Applicants in the local job market can either obtain an authentic local degree or buy degrees outside of their home state (in this case, West Bengal). The three gray universities were in three different regions (East-central, West, and South India) and located in mid- to small-sized cities. The limitation of our design is that we cannot separate out the effect of geography from the overall effects of gray degrees.

Design of Applicant Profiles

We designed the resumes with three considerations in mind: First, we wanted to create realistic applicants for the chosen jobs and sectors. We obtained typical and real resumes from a human resource consultancy. These resumes had Pan-India coverage, had been used in actual applications, and were aimed at jobs where no work experience was necessary. Second, we needed to pick names for applicants that suited the Indian context. Discrimination by caste, gender, and ethnicities is well researched and documented in the literature (Carlsson and Rooth 2007; Pager, Bonikowski, and Western 2009). For instance, there is evidence that employers discriminate using the names of job applicants (Banerjee and Knight 1985; Deshpande 2011). To minimize these kinds of sources of statistical noise and competing drivers of callback rates, we selected upper-caste Hindu names. Furthermore, in India there is also prejudice and discrimination based on the state of origin. So we used only Bengali names. Finally, in mixed-sex sectors we randomly used male and female names. In the case of female (male)-dominated sectors, we only applied with female (male) resumes.

Experimental Procedure and Data Collection

In total we used eighteen resumes: nine male and nine female, with three different names and each with three different types of education levels: gray, authentic, and no university degree (high school only). For each job, we sent three resumes. Across jobs we varied the following features in a balanced way: (1) high school name and residential address and (2) name and gender of applicants (for mixed-gender sector jobs).

We recorded callback rates through e-mails and phone calls. Every resume had a unique e-mail and mobile phone number. We did not apply to jobs requiring applicants to directly call or visit the company. Further, we used authentic residential addresses for all resumes; however, for practical reasons and due to the fictitiousness of applicants we could not record responses by post.11 An applicant's home address may come with labor market discrimination. Bertrand and Mullainathan (2004) found that applicants from predominantly black neighborhoods in the United States receive relatively fewer callbacks than those from white neighborhoods. To minimize such effects, we used residential addresses located in semi-urban areas in West Bengal that are mainly inhabited by Hindus. Sample sizes by degree and gender are summarized in table 1.

Table 1.
Sample Sizes by University Degree and Gender
University DegreeNoGrayAuthenticTotal
Male applicant 66 66 66 198 
Female applicant 66 66 66 198 
Total applications 132 132 132 396 
University DegreeNoGrayAuthenticTotal
Male applicant 66 66 66 198 
Female applicant 66 66 66 198 
Total applications 132 132 132 396 

3.  Results

Gray degrees significantly increase callback rates compared to having no degree. The impact is moderate in size and stronger among female sector jobs and applicants. In addition, authentic degrees have a much larger, relative impact on callback rates. However, we can document heterogeneous impacts of gray degrees according to their ranking (as a function of the subjective ease of obtaining them). In particular, the positive impact of the high-ranked gray degree is not statistically different from the one associated with authentic degrees, while the effect associated with the lowest ranked one is insignificant. In what follows, we first present the main findings graphically, and then investigate the robustness and heterogeneity of the effects in a regression model.

Overall, 30 percent of our applications received a callback. Figure 2 breaks down callback rates by the type of degree. Panel A shows a clear hierarchy: resumes with no degree received callbacks in 19.70 percent of all applications, whereas those with a gray or authentic degree averaged 28.03 percent and 41.67 percent, respectively. The impact of a gray degree on callback rates compared to no degree amounts to 8.33 percentage points (p-value = 0.07, t-statistic = 1.82, N = 264).12 This amounts to a 42 percent increase. In comparison, the impact of authentic degrees over no degrees is 21.97 percentage points (p-value = 0.00, t-statistic = 4.90, N = 264). Finally, the difference between gray and authentic degrees amounts to −13.64 percentage points (p-value = 0.00, t-statistic = −3.29, N = 264).
Figure 2.

Callback Rates by Type of University Degree

Notes: N = 396 (see table 1 for a breakdown by degree and gender). All p-values stem from pairwise regression-based t-tests adjusted for clustering at the job posting level. In panel B, the first p-value refers to differences in the female sample and the second p-value to differences in the male sample.

Figure 2.

Callback Rates by Type of University Degree

Notes: N = 396 (see table 1 for a breakdown by degree and gender). All p-values stem from pairwise regression-based t-tests adjusted for clustering at the job posting level. In panel B, the first p-value refers to differences in the female sample and the second p-value to differences in the male sample.

We can also differentiate callback rates by degree and gender of the applicant (see figure 2, panel B). One limitation is that gender differences may in part be driven by sectoral differences. Further, our experiment was not designed to efficiently test for gender differences. That said, a similar ranking of degrees emerges: resumes with authentic degrees receive more callbacks than those with gray or no degrees. However, differences are stronger among female applicants. Overall, female applicants receive more callbacks than male applicants (36.36 percent versus 23.23 percent). And the difference in callback rates between female resumes with gray and no degrees amounts to 12.12 percentage points (p-value = 0.09, t-statistic = 1.72, N = 132). The corresponding difference for male resumes is small in magnitude (4.55 percentage points) and insignificant (p-value = 0.45, t-statistic = 0.77, N = 132).

Table 2 summarizes estimates stemming from a linear probability model13 in which the dependent variable takes on a value of one in the case of a callback and zero otherwise. Standard errors are clustered at the job posting level. Column 1 presents a stripped-down model with no covariates. “No degree” is the excluded category. The impact associated with gray degrees relative to no degree amounts to 8 percentage points. The corresponding impact of authentic degrees is 22 percentage points. Column 2 gauges the sensitivity of these marginal effects to the inclusion of a gender dummy, as well as sector and applicant profile/resume dummies.14 Point estimates associated with the main variables of interest are extremely stable, confirming that randomization and a balanced sample were achieved in practice. The point estimates associated with gender and the type of sector are insignificant. Column 3 shows that our main findings are also robust to the inclusion of firm/job opening fixed effects. Point estimates and standard errors are stable. Findings are robust within and across firms.

Table 2.
Regression Estimates of the Effects of Types of University Degrees on Callback Rates
Dependent Variable Callback(1)(2)(3)(4)
Degree type: (No Degree excluded category) 
Gray 0.08* 0.08* 0.08*  
 (0.05) (0.05) (0.05)  
Gray (low rank)    0.04 
    (0.08) 
Gray (medium rank)    0.06 
    (0.08) 
Gray (high rank)    0.15* 
    (0.08) 
Authentic 0.22*** 0.22*** 0.22*** 0.22*** 
 (0.04) (0.04) (0.04) (0.05) 
Male applicant  −0.10  −0.09 
  (0.10)  (0.10) 
Male sector  −0.06  −0.06 
  (0.08)  (0.08) 
Female sector  −0.04  −0.04 
  (0.09)  (0.09) 
Constant 0.20*** 0.28*** 0.20*** 0.28*** 
 (0.03) (0.07) (0.03) (0.07) 
Effect equality (p-value) 
Gray = Authentic 0.00 0.00 0.00 0.01; 0.04; 0.41 
Profile dummies   
Job posting fixed effects    
N 396 396 396 396 
Dependent Variable Callback(1)(2)(3)(4)
Degree type: (No Degree excluded category) 
Gray 0.08* 0.08* 0.08*  
 (0.05) (0.05) (0.05)  
Gray (low rank)    0.04 
    (0.08) 
Gray (medium rank)    0.06 
    (0.08) 
Gray (high rank)    0.15* 
    (0.08) 
Authentic 0.22*** 0.22*** 0.22*** 0.22*** 
 (0.04) (0.04) (0.04) (0.05) 
Male applicant  −0.10  −0.09 
  (0.10)  (0.10) 
Male sector  −0.06  −0.06 
  (0.08)  (0.08) 
Female sector  −0.04  −0.04 
  (0.09)  (0.09) 
Constant 0.20*** 0.28*** 0.20*** 0.28*** 
 (0.03) (0.07) (0.03) (0.07) 
Effect equality (p-value) 
Gray = Authentic 0.00 0.00 0.00 0.01; 0.04; 0.41 
Profile dummies   
Job posting fixed effects    
N 396 396 396 396 

Notes: Linear probability models. Standard errors in brackets under point estimates are clustered at the job posting level. In column 4, we report tests of the equality between effects associated with the three gray degrees (low, medium, high) and the authentic degree. Mixed sector is the excluded category in columns 2 and 4.

*p < 0.1; ***p < 0.01.

Until now, we have estimated average effects of gray degrees on callback rates. However, in the experiment, we used three different (subjective) ranks of gray degrees. Column 4 shows the results of this heterogeneity analysis. The effects associated with these degrees are all positive. The effect of low- and medium-ranked degrees is, however, insignificant and small. Only the effect of the high-ranked gray degree is large and significant (at the 10 percent level). What is more, we find that authentic degrees have lost some of their competitive edge over gray degrees. We can no longer statistically differentiate between the effects of authentic versus high-ranked gray degrees (see p-values at the bottom of the table).15 In sum, callback rates are monotonically increasing with gray-degree rank. This is suggestive evidence of proper market discounting by firms with respect to the ease with which gray degrees can be bought.

Table 3 shows separate regression results for female and male applicants. In columns 1 and 4, we show full sample results for female and male applicants, respectively. In the remaining columns, we provide gender-specific results by industry sectors. The effects of gray degrees within gender are similar across the various subsamples. However, we lose precision in the smaller sectoral samples.16 We find that the effect of gray degrees on female callback chances is statistically significant only (at the 10 percent level) in column 1, yet always economically important and larger among female applicants (12 percentage points in columns 1–3 compared with 3 to 6 percentage points in columns 4–6). Finally, note that authentic degrees fare (at least qualitatively so) better than gray degrees and no degrees. It is also interesting to note that the effect of authentic degrees is large among female applicants.

Table 3.
Gender Heterogeneity
Dependent Variable Callback(1)(2)(3)(4)(5)(6)
ApplicantsFemaleMale
Degree type: (No Degree excluded category) 
Gray 0.12* 0.12 0.12 0.05 0.06 0.03 
 (0.07) (0.11) (0.10) (0.06) (0.08) (0.09) 
Authentic 0.29*** 0.27** 0.30*** 0.15*** 0.18** 0.12 
 (0.07) (0.10) (0.10) (0.05) (0.08) (0.07) 
Constant 0.23*** 0.21*** 0.24*** 0.17*** 0.12** 0.21*** 
 (0.05) (0.07) (0.08) (0.05) (0.06) (0.07) 
Effect equality (p-value) 
Gray = Authentic 0.01 0.14 0.01 0.07 0.16 0.26 
Sector(s) Female, Mixed Female Mixed Male, Mixed Male Mixed 
N 198 99 99 198 99 99 
Dependent Variable Callback(1)(2)(3)(4)(5)(6)
ApplicantsFemaleMale
Degree type: (No Degree excluded category) 
Gray 0.12* 0.12 0.12 0.05 0.06 0.03 
 (0.07) (0.11) (0.10) (0.06) (0.08) (0.09) 
Authentic 0.29*** 0.27** 0.30*** 0.15*** 0.18** 0.12 
 (0.07) (0.10) (0.10) (0.05) (0.08) (0.07) 
Constant 0.23*** 0.21*** 0.24*** 0.17*** 0.12** 0.21*** 
 (0.05) (0.07) (0.08) (0.05) (0.06) (0.07) 
Effect equality (p-value) 
Gray = Authentic 0.01 0.14 0.01 0.07 0.16 0.26 
Sector(s) Female, Mixed Female Mixed Male, Mixed Male Mixed 
N 198 99 99 198 99 99 

Notes: Linear probability models. Standard errors in brackets under point estimates are clustered at the job posting level.

*p < 0.1; **p < 0.05; ***p < 0.01.

4.  Discussion

Applicants with degrees from authentic colleges fare significantly better compared with both having no degree and degrees from questionable universities in our simple audit study. However, questionable degrees from gray universities can partially compensate for the lack of authentic credentials; this finding is concentrated among female applicants and female-sector jobs. These gender differences may be explained in that callback rates are higher for women in the first place and that women in India tend to have lower education levels than men and also tend to apply for lower-skilled jobs. A related explanation for the gender differences in callback rates may stem from the fact that employers understand the difficulty for Indian women to accumulate human capital. Consequently, they reward any degree (gray or authentic) obtained by women relatively more.

We rationalize our results in the sense that employers hold beliefs about the probability that a degree from a gray institution is bought rather than earned. It is clear that employers discount degrees from questionable universities. Table 4 breaks down callback rates at the firm level: 13 percent of firms called back applicants with degrees from authentic and gray universities and ignored applicants with no degree (see column 1). At the same time, 14 percent of firms only called back applicants with degrees from authentic universities. Conversely, 5 percent of firms called back applicants with authentic and no degrees, ignoring those with questionable degrees. These patterns are stronger for female resumes (compare columns 2 and 3). Our findings are thus consistent with a simple conceptual model (see section 2) and suggest that most employers lump gray degrees within a broad degree category. Simply put, employers prefer a degree (gray/questionable or authentic) over no degree at all. They cannot perfectly verify the authenticity of a degree from a questionable university but assign a positive probability (however small that may be) that the degree is nevertheless authentic.

Table 4.
Callback Rates at the Job Posting/Firm Level
Sample
(1)(2)(3)
Firm-Level Response Total Male Female 
No applicant 0.48 0.59 0.38 
Only no degree 0.05 0.03 0.06 
Only gray degree 0.05 0.06 0.05 
Only authentic 0.14 0.11 0.17 
Only no degree and gray 0.00 0.00 0.00 
Only no degree and authentic 0.05 0.06 0.05 
Only gray and authentic 0.13 0.08 0.18 
All applicants and degrees 0.10 0.08 0.12 
Number of jobs 132 66 66 
Sample
(1)(2)(3)
Firm-Level Response Total Male Female 
No applicant 0.48 0.59 0.38 
Only no degree 0.05 0.03 0.06 
Only gray degree 0.05 0.06 0.05 
Only authentic 0.14 0.11 0.17 
Only no degree and gray 0.00 0.00 0.00 
Only no degree and authentic 0.05 0.06 0.05 
Only gray and authentic 0.13 0.08 0.18 
All applicants and degrees 0.10 0.08 0.12 
Number of jobs 132 66 66 

Note: This table shows the distribution of callback rates within job postings.

Our qualitative insights complement our experimental findings. One agent summarized the economics of gray bachelor of law degrees (LLBs):

Let's talk about an LLB. You cannot do this on a part-time basis. But honestly, tell me, is it worth spending three years doing an LLB unless you go to a good university? We are providing all LLBs at [Rupees] 60,000 … If you pay, then you'll find that your answer script is ready and you just need to sign. [translated from Bengali, interview dated 12 July, 2016]

Do the benefits of gray degrees really exceed the costs? While our experiment suggests that gray degrees can increase callback rates by 8 percentage points, this does not inform about costs and benefits. That said, the gray degrees in our analysis cost about INR 20,000 (∼USD 300), which can be around a fifth of a yearly salary in an entry-level job.17 But costs vary depending on the degree level, subjects, and the extent to which the degree is managed by agents. For example, an LLB degree costs around INR 60,000 (∼USD 900), compared with INR 18,000 (∼USD 270) for a bachelor's degree in history. The costs of such degrees are actually quite steep compared to the positive impacts on callback rates. Further research should examine the motivations of buyers beyond callback rates, including longer-term economic advantages and social motives.

We would like to point out some immediate limitations of our study and sample: first, we only studied the impact on callback rates. We cannot document whether these callbacks translate into getting an offer and keeping a job. Relatedly, we do not have data on wages for holders of different degrees. Second, we focused on a small set of universities from which gray degrees can be obtained and on just one Indian state. Third, we focused on jobs advertisements that do not require higher-education degrees. We do expect a higher scrutiny of resumes for higher-ranked jobs that require academic credentials.

Our experimental design also has several limitations: Most importantly, we cannot parcel out whether employers make inferences about an applicant's productivity based on the choice of schools—both in terms of size, type, mode of instruction (online, distance, etc.) and location. For instance, gray-degree institutions in our experiment tend to be relatively far away, and employer perceptions of productivity may vary depending on the distance between home and institution. Furthermore, we find a relatively strong effect of gray degrees among women. However, we do not distinguish between mothers and non-mothers, which is an important dimension in the Indian labor market (Das and Zumbyte 2017; Bedi, Majilla, and Rieger 2018). Having a gray degree (potentially involving distance education) may tell the employer that the (female) applicant has children at home and may not have had time to obtain an authentic degree (in line with the above quote from the qualitative interviews). This in turn could imply the applicant has a lower commitment to the job, and will only be available part time (Correll, Benard, and Paik 2007). Several further limitations of our design are worth mentioning: we cannot provide more disaggregated results by study subject (for instance, history compared to political sciences), as well as job board, firm, or sectoral characteristics (see Deming et al. [2016] for a thorough design along these lines in the case of postsecondary qualifications in the United States).

We would like to highlight avenues for future empirical and theoretical work. First, it would be interesting to quantify the screening costs for firms in the (expanding) presence of gray degrees. Second, it would be insightful to study to what extent employers make inference about human capital when reviewing applications (with different degrees). One possibility would be to randomly report additional skills (unrelated to university education) on the resume. Third, gray degrees may theoretically have predictive value of job productivity. For instance, Weaver (2018) documents that bribes to obtain a public sector job in developing countries may actually increase welfare, as the willingness to pay a bribe is positively predictive of subsequent productivity on the job.

5.  Conclusion

In conclusion, mentioning a gray degree from a questionable university on one's resume improved callback rates relative to reporting no degree in our resume experiment. However, employers favored degrees from institutions that offer only authentic degrees. From our results, several policy implications emerge. First, it seems clear that authorities need to more thoroughly regulate and control the Indian higher education sector. We found that services relating to gray degrees are advertised quite openly; thus, either the legal framework is too lax or means to curb such practices are not adequately deployed. Some limited policy measures have been introduced, such as the abolition of the inefficient Distance Education Council and the establishment of the Distance Education Bureau as the new regulatory authority. However, the effectiveness of this new bureau has yet to be demonstrated. Further, authentic colleges need to keep distinguishing themselves clearly from gray institutions. They could, for instance, report gray degrees and agents operating in their areas to local and national authorities. Lessons may come from how other countries, such as the United States, regulate diploma mills or online colleges. In the United States, individual states govern and regulate the education sector and some of them have issued “negative lists of unapproved, unaccredited, or illegal providers” (USNEI 2007, p. 1). Likewise, Indian authorities could list gray degree–issuing institutions in publicly available databases. This may help employers discount questionable degrees even more and may lower the incentives of students to buy degrees.

Notes

1. 

Most Indian states feature at least one state-funded open and distance learning university. These are “open” in the sense that admissions are not selective. The largest and federally funded example of such an institution is the Indira Gandhi National Open University (IGNOU) with more than four million students. Note that the Indian distance education system is not an online system. Rather, students take part-time degrees and are not required to attend lectures. They receive self-study materials in hard copy. They may attend optional lectures, in some cases complete take-home assignments, and finally take exams (for more details, see Gupta and Gupta 2012; Gaba and Li 2015).

2. 

In our local context, firms and applicants tend to use the term CV (curriculum vitae), which in the U.S. context corresponds to a resume. We use the term resume throughout.

3. 

In India, it is perfectly common that job seekers with university credentials apply to jobs that only require high school degrees due to the competitiveness of the job market, and high-skilled jobs are normally occupied by the applicants from elite and established universities. The Indian media has covered this phenomenon (see, e.g., “PhD holders apply for SSC's clerical posts in West Bengal”, available at: http://timesofindia.indiatimes.com/city/kolkata/PhD-holders-apply-for-SSCs-clerical-posts-in-West-Bengal/articleshow/54432085.cms).

4. 

Degrees also convey “a certain prestige or social status” that may motivate buyers (see Groulleau, Kakhal, and Mzoughi 2008, p.680).

5. 

We thank an anonymous referee for suggesting this model.

6. 

We carried out power calculations for a test of two proportions using the pwr package in R. Setting Cohen's h for a binary outcome to 0.5 (medium effect), power to 80 percent, and significance level to 5 percent, the required sample size was 188.

7. 

In an administrative sense, a state consists of several districts, which consist of several blocks.

8. 

We do not have data on the number of students traveling out of their state. Our anecdotal and qualitative evidence suggests that only top students move out-of-state for arts or social science degrees, mainly to study at elite colleges in big cities. One exception appears to be engineering, where many students move out of their state.

9. 

The large majority of jobs were chosen and applied to through one of these three platforms, therefore, we cannot break down results by job portal.

10. 

Unfortunately, we could not find hard data on the percentage of employees without university credentials in the types of jobs included in our experiment.

11. 

In any case, it is unusual for employers to use regular postal services for jobs advertised on online platforms.

12. 

Throughout we present linear regression-based differences in means tests adjusted for clustering at the job posting level.

13. 

Probit or logit models yield similar results and are available on request.

14. 

These dummies account for effects of the six specific resumes (beyond the effects of gender and degree type).

15. 

Also, the difference between authentic and medium-ranked degrees in terms of callback rates is insignificant. However, the economic difference is sizeable. Therefore, we might be lacking power to test this difference.

16. 

We did not design the experiment to be able to document precise gender differences. We cannot efficiently disentangle gender and gendered-industry differences.

17. 

According to our qualitative interviews, a yearly salary in the low-skilled sector could be around 8,000 to 10,000 INR (USD 1,500 to 1,880). Also note that the unskilled minimum wage in West Bengal is 5,962 INR/month (see https://www.wblc.gov.in/sites/default/files/synopsys/01-07-2017/agriculture.pdf).

Acknowledgments

This paper is a substantially reworked version of selected chapters in the MA thesis by Tanmoy Majilla entitled “Grey Academic Degree Market in India: A Mixed Method Study” submitted to the Institute of Social Studies of Erasmus University Rotterdam, available online at: http://hdl.handle.net/2105/37158. The second author, Matthias Rieger, was the supervisor of the thesis. We thank all our participants in our qualitative work. We received valuable comments from Brandon Restrepo and Brigitte Vézina. We thank the editors and two anonymous referees for their detailed and most helpful comments. All remaining errors are our own.

REFERENCES

Attewell
,
Paul
, and
Thurston
Domina
.
2011
.
Educational imposters and fake degrees
.
Research in Social Stratification and Mobility
29
(
1
):
57
69
.
Banerjee
,
Biswajit
, and
John B.
Knight
.
1985
.
Caste discrimination in the Indian urban labour market
.
Journal of Development Economics
17
(
3
):
277
307
.
Bedi
,
Arjun
,
Tanmoy
Majilla
, and
Matthias
Rieger
.
2018
.
Gender norms and the motherhood penalty: Experimental evidence from India
.
IZA Discussion Paper No. 11360
,
Bonn, Germany
.
Bertrand
,
Marianne
, and
Sendhil
Mullainathan
.
2004
.
Are Emily and Greg more employable than Lakisha and Jamal? A field experiment on labor market discrimination
.
American Economic Review
94
(
4
):
991
1013
.
Brown
,
George M.
2006
.
Degrees of doubt: Legitimate, real and fake qualifications in a global market
.
Journal of Higher Education Policy and Management
28
(
1
):
71
79
.
Chowdhary
,
Vikram
.
2011
.
PhDs, Bachelor's degrees on sale in Punjab.
Available:
www.ndtv.com/india-news/phds-bachelors-degrees-on-sale-in-punjab-463936.
Accessed 16 October 2016
.
Carlsson
,
Magnus
, and
Dan-Olof
Rooth
.
2007
.
Evidence of ethnic discrimination in the Swedish labor market using experimental data
.
Labor Economics
14
(
4
):
716
729
.
Correll
,
Shelley J.
,
Stephen
Benard
, and
In
Paik
.
2007
.
Getting a job: Is there a motherhood penalty
?
American Journal of Sociology
112
(
5
):
1297
1338
.
Darolia
,
Rajeev
,
Cory
Koedel
,
Paco
Martorell
,
Katie
Wilson
, and
Francisco Perez
Arce
.
2015
.
Do employers prefer workers who attend for‐profit colleges? Evidence from a field experiment
.
Journal of Policy Analysis and Management
34
(
4
):
881
903
.
Das
,
Maitreyi B.
, and
Ieva
Zumbyte
.
2017
.
The motherhood penalty and female employment in urban India
.
World Bank Policy Research Working Paper No. 8004
.
Washington, DC
.
Deming
,
David J.
,
Noam
Yuchtman
,
Amira
Abulafi
,
Claudia
Goldin
, and
Lawrence F.
Katz
.
2016
.
The value of postsecondary credentials in the labor market: An experimental study
.
American Economic Review
106
(
3
):
778
806
.
Deshpande
,
Ashwini
.
2011
.
The grammar of caste: Economic discrimination in contemporary India
.
Delhi
:
Oxford University Press
.
Gaba
,
Ashok
, and
Wei
Li
.
2015
.
Growth and development of distance education in India and China: A study on policy perspectives
.
Open Praxis
7
(
4
):
311
323
.
Gupta
,
Deepti
, and
Navneet
Gupta
.
2012
.
Higher education in India: Structure, statistics and challenges
.
Journal of Education and Practice
3
(
2
):
17
24
.
Grolleau
,
Gilles
,
Tarik
Lakhal
, and
Naoufel
Mzoughi
.
2008
.
An introduction to the economics of fake degrees
.
Journal of Economic Issues
42
(
3
):
673
693
.
Kaas
,
Leo
, and
Christian
Manger
.
2012
.
Ethnic discrimination in Germany's labor market: A field experiment
.
German Economic Review
13
(
1
):
1
20
.
Kumar
,
Sharat
,
Atir
Khan
, and
Ashhar
Khan
.
2011
.
Degrees on sale: Jaipur study centres offer bachelor degree to PhD for money.
Available
http://indiatoday.intoday.in/story/degrees-on-sale-in-jaipur-phd-in-3-days/1/153852.html.
Accessed 16 October 2016
.
Majilla
,
Tanmoy
.
2016
.
Grey academic degree market in India: A mixed method study
.
Available:
http://hdl.handle.net/2105/37158.
Accessed 1 March 2017
.
Pager
,
Devah
,
Bart
Bonikowski
, and
Bruce
Western
.
2009
.
Discrimination in a low-wage labor market: A field experiment.
American Sociological Review
74
(
5
):
777
799
.
Ullas
,
Sruthy Susan
, and
Garima
Prasher
.
2013
.
Fake degree scam: No sweat, you can get a university degree in 10 days
.
The Times of India
,
24 September
.
U.S. Network for Education Information (USNEI)
.
2007
.
Accreditation and quality assurance: Diploma mills and fraud.
Available:
https://www2.ed.gov/about/offices/list/ous/international/usnei/us/fraud.doc.
Accessed 1 June 2018
.
Weaver
,
Jeff
.
2018
.
Jobs for sale: Bribery and misallocation in hiring
.
Unpublished paper
,
University of California
,
San Diego
.