Exploring State Authorization Stringency
Written by Dr. Jacob Fowles
What is the return on the investment in higher education for a student? A cursory review of both legislative agendas and scholarly research suggests that this question is as important as it is difficult to answer. Evidence conclusively demonstrates that, on average, earnings for those with bachelor’s degrees are roughly twice those of individuals with only a high school degree and 70% higher than those with some college credits but no degree (Broady & Hershbein, 2020). However, those averages only tell part of the story. Recent scholarly work suggests that ROI for students varies widely across colleges and universities as well as degree fields and types.
Further, higher education is what economists call an “experience good.” Experience goods are unique because their quality is difficult to accurately evaluate until well after the good has been consumed. In the context of higher education, it means that enrolling in higher education represents a leap of faith since the returns on that investment of time and money will not be known to the student until the job market reveals them. Instead, students rely on limited information from governments, organizations such as U.S. News and World Report, and, in many cases, colleges and universities to try to make the right decision. Unfortunately, this environment creates space for unscrupulous actors to potentially take advantage by making big promises, cashing tuition checks, and failing to deliver — leaving students holding the bag.
It is against this backdrop that state authorization agencies function, regulating institutions (public, not- for-profit, and for-profit) offering education within the boundaries of the state and serving a critical role protecting students. Despite the important role that authorization agencies play, they have received little scholarly attention. As a result, much remains unknown about how these agencies are organized or how they operate, leaving important questions about efficiency and effectiveness unanswered. With support from SHEEO and Arnold Ventures, I sought to answer two, basic, descriptive questions about authorization agencies: First, are there common patterns in approach to authorization across states? And second, if patterns exist, are there common state characteristics that function to explain those similarities?
A pragmatic issue in seeking to answer the first question is the diverse and multidimensional nature of authorization itself. The regulatory functions of authorization agencies are impressively diverse, ranging from setting fiduciary standards for institutions to monitoring the credentials of faculty and nearly everything in between. Fortunately, recent developments in the field of unsupervised machine learning allow researchers to uncover latent patterns in complex, multidimensional data even when such patterns are not easily seen. Drawing on Ness et al. (2021), whose work surveys state authorization agencies and develops an inventory of the intensity with which authorizers engage in various regulatory activities, I employ a Gaussian Mixture Model that identifies three distinct clusters of states defined by similarity in authorization approach. Analysis of the commonalities of the clusters shows that states largely seem to adopt a one-size-fits-all approach to regulatory efforts. In other words, states that are very stringent in approach to one regulatory function tend to be comparatively stringent in others, while states that seem to have a more relaxed posture in one regulatory domain also tend to exhibit that same level of intensity elsewhere.
This is an interesting finding because literature in public administration and political science suggests that state regulatory effort, in many cases, seems to be more unevenly targeted than uniformly applied. In understanding the logic, consider the choices of a homeowner in protecting their property from break-ins. One option would be to build impenetrable walls around the property. Another would be to invest in a state-of-the-art security system. A third might be to hire vigilant private security guards to patrol the property. But you would not necessarily expect even the most concerned homeowner to do all three. The marginal safety benefit of additional security declines as more measures are adopted. But in the case of authorization, states that take a stringent approach in one regulatory domain seem to uniformly apply that standard of stringency in others: they build the wall, hire the guard, and put in the alarm system (or, in the case of less stringent states, build a less substantial wall, hire the guard part- time, and put in a security camera or two). What I do not observe is a cluster of states that prefer security guards to walls or a cluster of states that put in the high-end security system and supplement it with part-time guard patrols but eschew building walls. While we do not yet have the appropriate data to assess these choices with respect to how they ultimately serve and protect students, important questions are raised about the return on investment associated with authorization regulatory effort. It is my hope that this research will inspire future work to answer these important questions.
In answering the second question, I utilize multinomial logistical regressions to try to understand what factors may be associated with alignment with one of the identified clusters over the others. In so doing, I employ three categories of variables identified in the literature as drivers of higher education policy: state socioeconomic and demographic characteristics, state political dynamics, and characteristics of the state higher education system. Additionally, drawing from the public policy literature, I include a fourth category of variables capturing the broader dynamics of the state’s regulatory environment. These are employed as independent variables in models that test whether differences across states in these four categories are associated with differences in state authorization stringency. Of the four categories, only state socioeconomic and demographic characteristics and the characteristics of the state higher education system emerge as statistically significant.
This is an interesting finding considering the rich body of research across academic disciplines and policy domains exploring the antecedents to policy adoption. If we think of regulatory posture as an output of some process within states, it is inherently valuable to understand what state characteristics seem to be associated with a preference for a particular regulatory posture over other alternatives. This matters because if the public policy literature tells us anything, it is that there are likely multiple pathways that lead to achieving a given policy goal. A broader understanding of the context through which regulatory policies are determined can help connect research to policy by facilitating the development of evidence-based policy solutions that are not only efficacious but politically, socially, and economically palatable. In other words, policies that are likely to both succeed and be adopted. As states continue to face difficult trade-offs due to the economic effects of the pandemic, such work is particularly well-positioned to help states balance the need to protect students and the omnipresent requirement of budgets to balance.
Broady, K., & Hershbein, B. (2020). Major decisions: What graduates earn over their lifetimes. The Hamilton Project. https://research.upjohn.org/externalpapers/91/
Ness, E., Baser, S. & Dean, M., (2021). State authorization landscape and process: An inventory, classification, and analysis. Boulder, CO: State Higher Education Executive Officers Association. https://sheeo.org/wp-content/uploads/2021/09/StateAuthorizationLandscape_NessBaser.pdf