↳ Human_capital

March 9th, 2019

↳ Human_capital

Incomplete Squares

CONTEXT ALLOCATION

Expanding the frame for formalizing fairness

In the digital ethics literature, there's a consistent back-and-forth between attempts at designing algorithmic tools that promote fair outcomes in decision-making processes, and critiques that enumerate the limits of such attempts. A December paper by ANDREW SELBST, dana boyd, SORELLE FRIEDLER, SURESH VENKATASUBRAMANIAN, and JANET VERTESI—delivered at FAT* 2019—contributes to the latter genre. The authors build on insights from Science and Technology Studies and offer a list of five "traps"—Framing, Portability, Formalism, Ripple Effect, and Solutionism—that fair-ML work is susceptible to as it aims for context-aware systems design. From the paper:

"We contend that by abstracting away the social context in which these systems will be deployed, fair-ML researchers miss the broader context, including information necessary to create fairer outcomes, or even to understand fairness as a concept. Ultimately, this is because while performance metrics are properties of systems in total, technical systems are subsystems. Fairness and justice are properties of social and legal systems like employment and criminal justice, not properties of the technical tools within. To treat fairness and justice as terms that have meaningful application to technology separate from a social context is therefore to make a category error, or as we posit here, an abstraction error."

In their critique of what is left out in the formalization process, the authors argue that, by "moving decisions made by humans and human institutions within the abstraction boundary, fairness of the system can again be analyzed as an end-to-end property of the sociotechnical frame." Link to the paper.

  • A brand new paper by HODA HEIDARI, VEDANT NANDA, and KRISHNA GUMMADI attempts to produce fairness metrics that look beyond "allocative equality," and directly grapples with the above mentioned "ripple effect trap." The authors "propose an effort-based measure of fairness and present a data-driven framework for characterizing the long-term impact of algorithmic policies on reshaping the underlying population." Link.
  • In the footnotes to the paper by Selbst et al, a 1997 chapter by early AI researcher turned sociologist Phil Agre. In the chapter: institutional and intellectual history of early AI; sociological study of the AI field at the time; Agre’s departure from the field; discussions of developing "critical technical practice." Link.
⤷ Full Article

March 2nd, 2019

Weak Local Lineament

REACH ARREARS

Charting the significance of credit and debt throughout society

Household debt has proliferated in the past decade. In the final quarter of 2018, it reached $13.54 trillion—an $869 billion increase since the previous peak in 2008 and a 21.4% increase since the post-crisis trough. While it is now widely recognized that the financialization of household consumption set the groundwork for the Recession (see for example this chapter by Manuel Aalbers), credit markets seem immune to structural reform.

On the one hand, access to credit enables purchases and investments crucial to long term financial mobility; on the other, it incorporates those who lack resources into a cycle of obligations to lenders. In her most recent publication in the Annual Review of Sociology, RACHEL E. DWYER questions how debt has shaped the American social landscape. She develops a two dimensional model of formal debt relationships which categorizes contracts according to the source of credit (state vs. market) and the nature of the obligation (prospective vs. retrospective). The model integrates the logic of debt and credit relationships with an analysis of distributional politics:

"The top row of prospective credit offers are more likely made to affluent or middle-class and disproportionately white populations, and the bottom row of retrospective financial obligations are more likely to fall on lower-income or poor and disproportionately racial/ethnic minority populations. The experience of debt and financial fragility is thus different across these social groups defined by class, race/ethnicity, and other social status, though also tied together by similar logics of financialization and individualized accountability for life conditions."

Dwyer's research shows how credit and debt relations vary geographically and temporally, encouraging a comparative analysis of debt relationships in countries with different political economies. Link to the article.

  • On the unique role that credit markets play in the American economy, see Monica Prasad on the credit-welfare state tradeoff, and Colin Crouch on privatized-mortgage Keynesianism. Link to the first; link to the second.
  • For a pre-crisis examination of credit and inequality, see Patrick Bolton and Howard Rosenthal's Credit Markets for the Poor. Link.
  • Vicki Been, Ingrid Ellen, and Josiah Madar explore the relationship between urban segregation and subprime mortgage lending. Link.

    New Researchers: VISIBILITY PREMIUM

Political effects of celebrity exposure

In a novel paper, HEYU XIONG—a Phd candidate at NORTHWESTERN and newly appointed professor at CASE WESTERN RESERVE UNIVERSITY—studied the political consequences of television celebrity. He used the career of Ronald Reagan as a case study and exploited quasi-experimental variation in television reception to estimate the effects of celebrity media exposure on political outcomes, finding that
support for Reagan based on non-political factors extended nearly two decades after his television career—an effect more pronounced in areas in which Reagan was not a known political entity. The findings suggest that elections hinge considerably more on non-political media exposure and personal characteristics than previously assumed.

From the abstract:

"My results contribute to our knowledge of the vote decision process. Understanding what candidate information is pertinent and how that information is processed is key to understanding the selection of elected officials and, subsequently, the policies those elected officials enact. The economic theory of electoral competition is traditionally situated in the framework of the policy oriented voter. Even without the assertion of rationality, voters are, at the very least, presumed to be voting in order to advance a policy position or to express a political preference. While this preoccupation is not misplaced, the results suggest that candidates' personal characteristics constitute a significant, if substandard, criterion for vote choice."

Link to the paper.

⤷ Full Article

February 23rd, 2019

Grievous Plans

NO SHORTAGE

New evidence on the relationship between skills and labor supply

More than a decade after the financial crisis of 2008, median household incomes have stagnated at their pre-2008 levels, and global economic growth is expected to decline further from what is already a historic low. While the unemployment rate has rebounded, part time, service, and temporary work remain the principal drivers behind labor market growth. Weak recovery from the crisis has been widely attributed to the “skills gap”; commentators and policymakers alike hold that quality jobs are there, but Americans are simply not qualified to perform them.

At the American Economic Association’s most recent conference, ALICIA SASSER MODESTINO, DANIEL SHOAG, and JOSHUA BALLANCE provide evidence against this view. Using a proprietary database of more than 36 million online job postings, they show that employers increased skill requirements in states and occupations which experienced larger increases in the unemployment rate. Their findings suggest that it wasn’t a shortage of skills which weakened labor markets, but rather the ubiquity of qualified applicants which drove employers to raise hiring standards. By testing employer responses to an influx of veterans from Iraq and Afghanistan, the authors are able to confirm this mechanism:

"As a source of exogenous variation in the availability of skilled workers, we make use of a natural experiment resulting from the large increase in the post-9/11 veteran labor force following troop withdrawals from Iraq and Afghanistan... Panel A of Table 5 demonstrates that there is a strong, significant, and positive relationship between the sharp increase in the supply of returning veterans and the rise in employer skill requirements for both education and experience."

This is among the first pieces of empirical evidence which suggests that employer skill requirements are driven, in part, by labor supply. Link to the conference webpage, where a full version of the paper is available for download.

  • As early as 2011, Lawrence Mishel argued against analysts who asserted that the unemployment crisis was structural, proposing instead that the economy was experiencing a crisis of demand. Link.
  • In his most recent book, LSE anthropologist David Graeber examines the relationship between skill and value, questioning why jobs which produce the most social value tend to be categorized as unskilled, consequently earning lower wages. Link to Graeber's widely acclaimed essay from 2013 that first outlined his argument, and link to the Google preview of his new book.
  • In a report for the Roosevelt Institute, Marshall Steinbaum and Julie Margetta Morgan argue that the 'skills gap' narrative is inconsistent with student debt crisis: "Although the country’s populace is becoming more educated, each educational group is becoming less well paid." Link.
  • Paul Osterman wrote an accessible overview of the debate for The Atlantic in 2014: “The claim that a shortage of skilled workers has exacerbated inequality has gained traction but it is not supported by the data… For instance, while 38 percent of manufacturing firms require math beyond simple addition, subtraction, and multiplication, the type of math employees need to be able to handle are standard features of a good high school education and part of the curriculum for most community-college students…Nearly 65 percent of businesses report they have no vacancies whatsoever, and another 76.3 percent report they have no long-term vacancies…” Link.
⤷ Full Article