June 2nd, 2020

Clouds, Sun, Sea

UNREST AND POLICY

This week has seen policymakers, scholars, and the public debate the meaning of collective violence. While political and media discourse often fails to examine the long-term effects of social unrest, a vast literature grapples with the mechanisms that link protests and uprisings with institutional change.

A 1978 book by JAMES W. BUTTON integrates a vast amount of interviews, archival sources, and statistical data to analyze the public response to the US urban uprisings of the 1960s. Focusing the analysis on three federal agencies—the (now-dismantled) Office of Economic Opportunity, HUD, and the DOJ—the book suggests that the 1960s riots were understood by policymakers as political demands.

From the introduction:

"Although domestic collective violence has played a prominent role in American history, few other episodes of urban violence in this country's history have been as dramatic as the black riots of the 1960s. As a result, the causes, precipitating events, and participants of the outbursts have been thoroughly studied over the past several years. Yet what is remarkable about this extensive analysis is the almost complete neglect of the political effects or consequences of these pervasive disorders. By concentrating instead on the factors that may have caused the riots, most investigators have implicitly reflected a normative bias concerning the disutility of domestic violence for affecting social and political change.

The fundamental purpose of this study is to evaluate some of the political consequences of the urban black riots of the 1960s and early 1970s, focusing on responses of the executive branch of the federal government. In fulfilling such a task, it asks: did the riots affect executive officials' decisions and ultimately federal public policy? Did the federal executive branch respond differently to the initial, less intense riots (1963-1966) than it did to the later, more severe disorders (1967-1968) and, if so, why? how have national executive responses to urban rioting been affected by the local political and environmental context and by local reactions to such violence? And how do public officials tend to view the role of violence in American society?"

Link to the book page.

  • A new article by Omar Wasow examines the relationship between violent and nonviolent protests, media, public opinion, and policy alignment from the Civil Rights Era, and in particular on Nixon's election in 1968. Link. And a 2018 paper by Shom Mazumder looks at the persistent effects of Civil Rights protects on political attitudes. Link.
  • A 1978 paper by sociologist Charles Tilly on collective violence: "Historically, collective violence has flowed regularly out of the central political processes of western countries. People seeking to seize, hold, or realign the levers of power have continually engaged in collective violence as part of their struggles." Link.
  • In a 2007 article, historian Michael Kazin asks: "Many of the conditions thought to have precipitated the eruption of civil violence in the 1960s either persist or have grown worse. What accounts for the absence of civil violence on American streets?" Link. And a new book by Mike Davis and Jon Wiener looks at the 1960s in Los Angeles. Link.
 Full Article

May 28th, 2020

Digital Scab, Digital Snitch

On automation and worker surveillance

Before Covid-19 hit, we'd become used to reports about Amazon's robotics innovations and the impending large-scale automation of warehouse jobs. But recent strikes and protests by Amazon's very human workers have exposed how far we are from robotic warehouses. In fact, as part of its effort to keep its warehouses fully staffed during the crisis, Amazon recently announced that it is ending its recently-instituted sick leave and base pay expansions, replacing both with increased overtime pay. While higher pay encourages more workers to apply for jobs, overtime incentivizes existing workers to work longer hours. Amazon’s strategy for increasing output in the pandemic seems to be getting its human employees to work harder.

In late-February, I took what must have been one of the last public tours of an Amazon warehouse in Edison, New Jersey. Prepared to witness vast and impression automation, I was met instead by a traffic jam of workers exiting the facility. Inside, there were of course robots—shelving units known as "pods," whizzing stuff from one end of the warehouse to the other—performing tasks previously done by people with forklifts. (According to Amazon, these robots have raised productivity of the remaining workers by orders of magnitude.) But there were other innovative technologies on view, at the workstations of human “stowers,” who distribute incoming products to the pods, “pickers,” who take items off the pods to fulfil orders, and “packers,” who put the orders in boxes and tape them shut. Over their shoulders, there were clocks counting down how much time each worker had to complete each task. The technology tracks which workers fall behind, and ‘learns’ how hard it can push them. In place of a human foreman, Amazon's timers dictate the pace of work and mete out automated discipline.

 Full Article

December 23rd, 2019

The Road in the Forest

Thank you for reading the JFI letter this year. As we prepare for another year of research and link sharing, here's some of what we sent in 2019.
We'll see you in 2020.

OVER ILLUMINATION

Highlights from a year of JFI Letters

+ In our first newsletter of 2019, we looked at a report by Yale School of Management's Evidence in Practice project, which considered the relationship between research and policymaking: "The most successful examples of evidence integration lessen the distinction between evidence generation and application, and focus on designing approaches that simultaneously generate (different types of) rigorous evidence and develop an iterative process for integrating evidence into practice." Link to the archived letter. + A July letter features work by Jonas Hjort et. al on how research evidence shaped the decisions of policymakers in Brazil. Link to the archived letter.

+ Recurring debates on the future of work: Brishen Rogers argues that labor precarity is the result of politics, not the outcome of any force of automation outside of our control: "Hotel work, food services, janitorial work, and retail work have become precarious over the past twenty years because companies in those sectors forcibly de-unionized and/or 'fissured' away their workers to subcontractors or franchisors, thereby denying them effective access to many legal rights." Link to the April letter. + In the first of a two-part series, Aaron Benanav historicizes automation debates in order to shed light on their significance for the present. Link to the October letter.

+ On education and the labor market: Alicia Sasser Modestino et. al criticize the "skills gap" theory, which suggests that labor standards are declining because American workers lack the training to enter high paid industries. They argue instead that it's due to an abundance of skilled workers that employers have raised the credentials required for entry level jobs. And Marshall Steinbaum and Julie Morgan show the "skills gap" theory to be inconsistent with the student debt crisis. Link. + In March, we looked at the proliferation of certificate programs, an understudied development in the higher education landscape. Link to the archived letter.

+ On the nuts and bolts policy history and implementation: A recent letter discussed Pamela Herd and Donald Moynihan's book, Administrative Burden, which argues that, "ultimately, burdens are the fine print in the social contract between citizens and their government." Link to the archived letter. + A look into the history of the EITC, from May. Link.

 Full Article

October 21st, 2019

Sunset Red and Gold-- The Gondolier

DESCRIBED FUTURES

Automation fears and realities

Of the many justifications for introducing a universal basic income, automation is among the most popular. Over the past years, a slew of reports and endless media coverage has raised the specter of mass "technological unemployment"—a possible future that has been taken up by basic income proponents across the political spectrum. It was even a point of argument in this week's Democratic presidential debate.

In the first of a two-part series, historian AARON BENANAV (whose work on the history of unemployment categories we shared in a previous letter) critiques and situates the automation debates within long-term global trends. Framed as a response to what Benanav terms the "automation theorists," who maintain a sense of inevitability about the robot takeover, the paper pursues alternate explanations: declining labor demand, global deindustrialization, and manufacturing overcapacity.

From the paper:

“Automation turns out to be a constant feature of the history of capitalism. By contrast, the discourse around automation, which extrapolates from instances of technological change to a broader social theory, is not constant; it periodically recurs in modern history.

The return of automation discourse is a symptom of our era, as it was in times past: it arises when the global economy’s failure to create enough jobs causes people to question its fundamental viability. The breakdown of this market mechanism today is more extreme than at any time in the past. This is because a greater share of the world’s population than ever before depends on selling its labour or the simple products of its labour to survive, in the context of weakening global economic growth.”

Link to the paper, and link to an ungated version on the author's website.

  • David Autor's 2016 paper "Paradox of Abundance" examines the problem of its title: "technological changes threatens social welfare not because it intensifies scarcity but because it augments abundance." Link.
  • A previous newsletter highlights a paper by legal scholar Brishen Rogers, which critiques automation fears in the US context by pointing to labor law and the "fissuring" of the workforce as more consequential for stagnating wages and declining job security. Link. Along the same lines, but in the European context, Zachary Parolin's recent work for the OECD measures the effects of collective bargaining agreements on wages in automatable occupations. Link.
  • Three post-debate accounts of the issue: Paul Krugman in the Times; Matt Yglesias in Vox; and Jordan Weisman in Slate, featuring the following quote from David Autor: "If we talk about the economic trauma of the 2000s, that’s not primarily due to automation. Nobody can tell you what great invention happened in 1999 that wiped out 20 percent of manufacturing jobs."
  • For another broad view of macro trends and low-demand problems, see JW Mason's "Macroeconomic Lessons from the Past Decade." Link.
 Full Article

May 13th, 2019

Reality Slays Art

PREMATURE PROGRESS

New patterns in deindustrialization

As economies across Europe and in the United States have become more knowledge-based, urban-centered, and tech-driven, people in manufacturing reliant regions have seen declining life expectancies, stagnating real incomes, and minimal job growth.

In recent years, social scientists have been grappling with the interconnected political, economic, and social effects of deindustrialization. But this literature is almost entirely confined to Europe and the U.S. In a new paper, DAVID KUNST broadens the scope of this research using a novel dataset on manufacturing employment by occupation in developing countries. He studies the labor market effects of 'premature deindustrialization,' finding a general decline in the hiring capacity of manufacturing sectors and a genuine risk from automation in emerging markets. The study comes to four conclusions:

"First, it is mostly unskilled jobs that have disappeared, and also the wage premium of workers with little formal education in manufacturing relative to other industries has declined. Second, the disappearing jobs have been among the most formal both relative to other industries, and to the manufacturing average. Third, premature deindustrialization has been driven by occupations which are intensive in tasks that are vulnerable to an increasing adoption of ICT. Fourth, the phenomenon pertains most clearly to middle income countries, as low income countries have been spared from premature job losses.

250 years after the beginning of the Industrial Revolution, it appears that manufacturing is losing its ability to employ unskilled workers more productively than other industries. Developing countries, abundant in unskilled labor, lose their comparative advantage in producing an increasing range of manufactured goods. Hence, future growth in developing countries may have to rely more on improvements in 'fundamentals' such as education and governance, and policy makers need to focus on a broader range of sectoral policies than in the past."

Link to the full paper.

  • The notion of 'premature deindustrialization' was developed by Dani Rodrik in 2015. In that paper, Rodrik argued that "countries are running out of industrialization opportunities sooner and at much lower levels of income compared to the experience of early industrializers" and suggested that "early deindustrialization could well remove the main channel through which rapid growth has taken place in the past." Link.
  • In a 2017 report, Carol Graham, Sergio Pinto, and John Juneau II map the "geography of desperation" in the United States: "In general, minorities scored worse on all of the variables in states where there are proportionately fewer minorities, such as Washington State and Kansas. These include Maine, Wisconsin, North Dakota, and Florida. Poor whites, meanwhile, tended to score lower across the board in the Appalachian states, and then poorly in many of the Midwestern and Western heartland states." Link. Two more reports from Brookings offer suggestions for place-based policies in the U.S. to counter these effects. Link, link.
  • David Clingingsmith and Jeffrey G. Williamson study the causes behind Indian deindustrialization from 1750-1860. Unlike literature which attributes the decline to growing competition in textile production from Britain, the authors find that the dissolution of Mughal hegemony and deteriorating climate conditions better account for the shift. Link.
 Full Article

April 13th, 2019

Assemblies

FISSURED CHURN

Reexamining claims about automation and labor displacement

Current UBI discussions emerged out of concerns over the role of human beings in a machine-dominated labor market. In 2013, a paper by Oxford University professors Carl Benedikt Frey and Michael Osborne claimed that 47% of US jobs were at risk of long term automation. The statistic circulated widely, prompting fears of widespread unemployment. The debate over these predictions is complex: those who deny any threat from automation often point to near-full employment, and risk overlooking the proliferation of low-paying and precarious jobs; while those who forecast mass unemployment risk assuming that technological development necessarily leads to labor displacement.

In a 2018 paper, legal scholar BRISHEN ROGERS argues that fears of a robot takeover misapprehend the real dynamics in the labor market:

"In a period of technological upswing, with companies rapidly installing robotics and other automation devices, we would also see significant increases in labor productivity. In fact, productivity growth has recently been the slowest as at any time since World War II. What’s more, productivity change in the manufacturing sector—where automation is easiest—has been especially tepid lately, at 0.7 percent over the last decade. On a related note, levels of 'occupational churn,' or the net creation of jobs in growing occupations and loss of jobs in declining occupations, are also at historic lows.

Even more striking, if firms expected artificial intelligence to be a major source of productivity in the near future, they would surely be investing in information technology and intellectual property. But they aren’t. Computers and software constituted 13.5 percent of the value of companies’ investments from 2000-2007, as the internet was coming into wide use. Over the last decade, that rate declined to 4.8 percent. These differences strongly suggest that there is nothing inevitable about precarious work or economic inequality. Hotel work, food services, janitorial work, and retail work have become precarious over the past twenty years because companies in those sectors forcibly de-unionized and/or 'fissured' away their workers to subcontractors or franchisors, thereby denying them effective access to many legal rights."

Link to the paper.

  • An MIT Technology Review from 2018 surveyed the predictions of every paper published on job losses due to automation. The results: "There is really only one meaningful conclusion: we have no idea how many jobs will actually be lost to the march of technological progress." Link.
  • "...even those occupations which are contracting due to technological change will continue to provide plenty of job openings over the next two decades. The challenge lies in improving the quality of these jobs going forward." Paul Osterman anticipates Rogers' arguments in a column from 2017. Link.
  • Another recent paper by Brishen Rogers (to which we previously linked) continues the thread: "Based on a detailed review of the capacities of existing technologies, automation is not a major threat to workers today, and it will not likely be a major threat anytime soon." Link.
  • Daron Acemoglu and Pascual Restrepo published two papers on automation and employment: the first uses industry level data to observe changes in the task content of production. The second argues that automation has been primarily concerned with reducing the need for labor, with insufficient attention being paid to socially productive investment. Link to the first, link to the second.
  • Frank Levy on the relationship between automation-induced job losses and the rebirth of populist politics. Link.
  • From EconFIP, a research brief on automation, AI, and the labor share. Link.
 Full Article

June 23rd, 2018

Yielding Stone

VISIBLE CONSTRAINT

Including protected variables can make algorithmic decision-making more fair 

A recent paper co-authored by JON KLEINBERG, JENS LUDWIG, SENDHIL MULLAINATHAN, and ASHESH RAMBACHAN addresses algorithmic bias, countering the "large literature that tries to 'blind' the algorithm to race to avoid exacerbating existing unfairness in society":  

"This perspective about how to promote algorithmic fairness, while intuitive, is misleading and in fact may do more harm than good. We develop a simple conceptual framework that models how a social planner who cares about equity should form predictions from data that may have potential racial biases. Our primary result is exceedingly simple, yet often overlooked: a preference for fairness should not change the choice of estimator. Equity preferences can change how the estimated prediction function is used (such as setting a different threshold for different groups) but the estimated prediction function itself should not change. Absent legal constraints, one should include variables such as gender and race for fairness reasons.

Our argument collects together and builds on existing insights to contribute to how we should think about algorithmic fairness.… We empirically illustrate this point for the case of using predictions of college success to make admissions decisions. Using nationally representative data on college students, we underline how the inclusion of a protected variable—race in our application—not only improves predicted GPAs of admitted students (efficiency), but also can improve outcomes such as the fraction of admitted students who are black (equity).

Across a wide range of estimation approaches, objective functions, and definitions of fairness, the strategy of blinding the algorithm to race inadvertently detracts from fairness."

Read the full paper here popup: yes.

 Full Article

January 27th, 2018

A Ha?

DISCONTINUOUS ADVANCE

A flurry of articles in December and January assess the state of artificial intelligence

From Erik Brynjolfsson et al, optimism about productivity growth:

“Economic value lags technological advances.

“To be clear, we are optimistic about the ultimate productivity growth fueled by AI and complementary technologies. The real issue is that it takes time to implement changes in processes, skills and organizational structure to fully harness AI’s potential as a general-purpose technology (GPT). Previous GPTs include the steam engine, electricity, the internal combustion engine and computers.

“In other words, as important as specific applications of AI may be, the broader economic effects of AI, machine learning and associated new technologies stem from their characteristics as GPTs: They are pervasive, improved over time and able to spawn complementary innovations.”

Full post at The Hill here.

 Full Article

January 13th, 2018

Landing Place

THE WAGE EFFECT

Higher minimum wages and the EITC may reduce recidivism

“Using administrative prison release records from nearly six million offenders released between 2000 and 2014, we use a difference-in-differences strategy to identify the effect of over two hundred state and federal minimum wage increases, as well as 21 state EITC programs, on recidivism. We find that the average minimum wage increase of 8% reduces the probability that men and women return to prison within 1 year by 2%. This implies that on average the wage effect, drawing at least some ex-offenders into the legal labor market, dominates any reduced employment in this population due to the minimum wage.”

Full paper by AMANDA Y. AGAN and MICHAEL D. MAKOWSKY here.

  • Jennifer Doleac responds, “The results in this new paper…definitely surprised me—my prior was that raising the min wage would increase recidivism.” She explains: “Those coming out of prison are very likely to be on the margin of employment (last hired, first fired). Given some disemployment effects, marginal workers are the ones who are going to be hurt. Amanda and Mike find that the positive effects of pulling some (higher-skilled?) offenders into the legal labor market outweigh those negative effects.” Link to Doleac’s Twitter thread.
  • One of the co-authors, Makowsky, adds, “The EITC, dollar for household dollar, generates larger effects [than minimum wage increases], but is hampered by its contingency on dependent children. This is one more reason to remove the contingency, extending it to everyone.” Link to Makowsky’s Twitter thread.
  • Another piece sourced from Makowsky’s thread: “The Impact of Living-Wage Ordinances on Urban Crime.” “Using data on annual crime rates for large cities in the United States, we find that living-wage ordinances are associated with notable reductions in property-related crime and no discernable impact on nonproperty crimes.” Link.
  • Noah Smith rounds up recent studies on increasing the minimum wage, many of which come to contradictory conclusions. "At this point, anyone following the research debate will be tempted to throw up their hands. What can we learn from a bunch of contradictory studies, each with its own potential weaknesses and flaws?" Link.
 Full Article

December 2nd, 2017

gesture/data

ARTIFICIAL AGENCY AND EXPLANATION

The gray box of XAI

A recent longform piece in the New York Times identifies the problem of explaining artificial intelligence. The stakes are high because of the European Union’s controversial and unclear “right-to-explanation” law, which will become active in May 2018.

“Instead of certainty and cause, A.I. works off probability and correlation. And yet A.I. must nonetheless conform to the society we’ve built — one in which decisions require explanations, whether in a court of law, in the way a business is run or in the advice our doctors give us. The disconnect between how we make decisions and how machines make them, and the fact that machines are making more and more decisions for us, has birthed a new push for transparency and a field of research called explainable A.I., or X.A.I. Its goal is to make machines able to account for the things they learn, in ways that we can understand. But that goal, of course, raises the fundamental question of whether the world a machine sees can be made to match our own.”

Full article by CLIFF KUANG here. This page provides a short overview of DARPA's XAI (Explainable Artificial Intelligence) program.

An interdisciplinary group addresses the problem:

"Contrary to popular wisdom of AI systems as indecipherable black boxes, we find that this level of explanation should often be technically feasible but may sometimes be practically onerous—there are certain aspects of explanation that may be simple for humans to provide but challenging for AI systems, and vice versa. As an interdisciplinary team of legal scholars, computer scientists, and cognitive scientists, we recommend that for the present, AI systems can and should be held to a similar standard of explanation as humans currently are; in the future we may wish to hold an AI to a different standard."

Full article by FINALE DOSHI-VELEZ et al. here. ht Margarita For the layperson, the most interesting part of the article may be its general overview of societal norms around explanation and explanation in the law.

Michael comments: Human cognitive systems have generated similar questions in vastly different contexts. The problem of chick-sexing (see Part 3) gave rise to a mini-literature within epistemology.

From Michael S. Moore’s book Law and Society: Rethinking the Relationship: “A full explanation in terms of reasons for action requires two premises: the major premise, specifying the agent’s desires (goals, objectives, moral beliefs, purposes, aims, wants, etc.), and the minor premise, specifying the agent’s factual beliefs about the situation he is in and his ability to achieve, through some particular action, the object of his desires.” Link. ht Margarita

  • A Medium post with an illustrated summary of some XAI techniques. Link.
 Full Article