Skip to main content
ABC News
Twitter Politics, Drunken Driving And Tax Deductions

inthepapersEvery Monday, the National Bureau of Economic Research, a nonprofit organization made up of some of North America’s most respected economists, releases its latest batch of working papers. The papers aren’t peer-reviewed, so their conclusions are preliminary (and occasionally flat-out wrong). But they offer an early peek into some of the research that will shape economic thinking in the years ahead. Here are a few of this week’s most interesting papers:


Title: “Homophily, Group Size, and the Diffusion of Political Information in Social Networks: Evidence from Twitter”

Authors: Yosh Halberstam, Brian Knight

What they found: Twitter users — both liberal and conservative — are politically cocooned; they tend to encounter only like-minded political information. Furthermore, users affiliated with the majority group in a state have more connections per capita and receive their political information more quickly.

Why it matters: Homophily is the tendency to associate with like-minded people. Such group inbreeding has been documented in numerous social networks, including high school friendships (no, really!). But some research had claimed that the Internet was less segmented than traditional social networks. This paper, however, concludes that we can add Twitter to the list of ideological echo chambers. The authors of this study examined millions of Twitters users who followed a candidate for the U.S. House of Representatives during the 2012 election, and then categorized the users’ ideology based on the political party of the candidate. They then tested whether these users’ political insulation by scanning who else they followed, what they retweeted, whom they linked to, etc. Not only are these Twitter users closed off from different points of view, the effect worsens if their group is in the majority. To demonstrate this, the authors clustered users by state, and showed how Twitter users in the state’s larger group have more connections and how information propagates more quickly within their circle.

Key quote: “Taken together, these results suggest that social networks in general, and social media in particular, may be a force for increasing differences in exposure to information between majority and minority groups and may also increase exposure to like-minded information for all groups.”

Data they used: 2.2 million Twitter users


Title: “Subjective Beliefs, Deterrence, and the Propensity to Drive While Intoxicated”

Authors: Yiqun Chen, Frank Sloan

What they found: The perceived risk of drunken driving is itself a deterrent to drunken driving, but only for those who are addicted to alcohol or those with more impulsive behavior generally.

Why it matters: Driving while intoxicated (DWI) is a huge problem in the United States. It caused more than 10,000 deaths in 2010. To combat drunken driving, the government has various sanctions for those caught, including jail time, license suspension and heavy fines. These explicit risks would seem to be a deterrent, but the authors of this study find they aren’t. Individuals don’t seem fazed by such “objective” risks. Instead, their “subjective” beliefs — the perceived risk of getting pulled over while driving drunk and of getting in an accident — are what matter. But the deterrent effects from higher perceived risks are limited only to those individuals with an addiction to alcohol or who have impulsive behaviors generally. The authors suggest a campaign touting the increased accident risk when driving intoxicated as a way to reduce DWI violations.

Key quote: “Harsher sanctions to drinking and driving may generate no marginal deterrent effect if drivers’ perceived risks of receiving the punishments are low. Therefore, rather than escalating sanction severity, increasing the probability of detection and/or enforcement may be more effective in reducing drinking and driving.”

Data they used: Survey of Alcohol and Driving


Title: “Raising Revenue by Limiting Tax Expenditures”

Authors: Martin S. Feldstein

What they found: Capping the amount a person can deduct or exclude from his tax bill can be an effective way for the government to raise revenue and reduce the national debt, without cutting spending directly or raising marginal tax rates. Specifically, limiting so-called tax expenditures to only 2 percent of an individual’s income would shave off at least $1.8 trillion from the national debt over the next decade.

Why it matters: The national debt is projected to grow in the long run. The bulk of this growth comes from entitlement programs, such as Medicare, and from interest. The normal routes to reducing debt — raising tax rates or cutting spending — are highly contentious. Thus one avenue both parties might be more likely to agree on is limiting “tax expenditures,” which are the deductions and exclusions someone can use to lower his tax bill. Examples include the mortgage interest deduction, or the tax credit he can get for installing a solar panel on his house. They are essentially subsidies. By limiting them, the government takes in more tax revenue without cutting spending or raising marginal tax rates. Feldstein proposes a cap on tax expenditures of 2 percent on an individual’s adjusted gross income and calculates that it could reduce the national debt by $1.8 trillion over the next decade (without counting additional savings from reduced interest costs).

Key quote: “The ability to frame tax expenditures as either revenue increases or spending decreases should make limiting tax expenditures appeal to those Republicans who want to reduce government spending as well as to those Democrats who want to use additional revenue to help shrink fiscal deficits. ”

Data they used: Congressional Budget Office

Andrew Flowers wrote about economics and sports for FiveThirtyEight.