We have seen enough to know that seemingly small amounts of sludge can have large consequences—a clear demonstration of the effects of “choice architecture” in determining outcomes.1 Choice architecture is the background against which decisions are made. Consider the layout of a grocery store (what’s at eye level? what do you see when you enter?); the design of a website (what’s in large font?); applications for permits and loans. Sludge is part of choice architecture. A store can impose some sludge before you buy cigarettes or e-cigarettes; perhaps you have to show identification and fill out some forms. A website can create some sludge to discourage you from selecting certain options. Applications can be so long and complicated that some people, or many people, will not bother.
Choice architecture can be pervaded by sludge or can be nearly free of it. Companies are well aware of that point. When they want you to choose an option, they make it really easy. They take away the sludge. With one click, you do what they want you to do. Dark patterns online—understood as manipulative practices designed to trick people into parting with their money—consist in large part of selective decreases and increases in sludge.
I have noted that in many domains, participation rates can be dramatically increased with a mere shift from requiring people to apply (opt in) to automatically enrolling them (opt out) and thus eliminating sludge. In an especially dramatic study, Professors Peter Bergman of Columbia University and Todd Rogers of Harvard University found that if parents are asked whether they want to sign up to receive text message alerts about the academic progress of their children, participation rates are tiny—around 1 percent.2 If the sign-up process is simplified, participation rates increase significantly, to about 8 percent.3 But if parents are automatically signed up, participation rates jump to 96 percent!4 An opt-out process is essentially an elimination of sludge, and with an opt-out approach, we can often achieve important social goals, whether we are interested in education, reduction of poverty, consumer protection, or climate change.
To be sure, most changes in choice architecture do not have effects of that magnitude.5 If you switch from opt in to opt out, the standard increase in participation has been found to be around 26 percent. But that is a lot. More generally, simplification and sludge reduction do not merely reduce frustration; they change people’s lives.
The right to vote may be the most fundamental of all, and it is often compromised by sludge. Here is a small example of how new or different architecture can help. The example has everything to do with overcoming the bandwidth problem.6
For millions of Americans, the steps needed to understand the process of where and how to become registered, and then following those steps, operate as serious barriers to voting. One result is that the nation ends up with stark demographic differences with respect to voter registration. In 2016, for example, a disproportionately large share of eligible Americans who were not registered were low-income citizens and people of color.7
Among eligible voters, 31 percent of African Americans, 43 percent of Hispanics, and just over 43 percent of low-income Americans were not registered to vote in the 2016 presidential election. When asked why they had not registered to vote, more than 60 percent of eligible voting-age adults responded that they had simply never been offered the chance to register.8 And more than one-third of those not registered said that they intended to do so, but had not gotten around to it or found the process inconvenient.
Hospitals can play a role in closing this gap by offering citizens a chance to register to vote in a place where poor people, and people of color, disproportionately show up for care: the emergency room. In that way, hospitals can reduce sludge. Emergency rooms care for higher rates of low-income, minority, and uninsured Americans than the average population. In 2016, for example, the annual visit rate among the entire population was 45.8 ER visits per one hundred persons.9 But when stratified among African American patients, the ER visit rate was much higher: eighty visits per one hundred persons. In addition, people without a college education and with lower incomes are especially reliant on the ER for nonemergency care.
While this illustrates a problem in our current medical system, it also introduces an opportunity to increase voter registration. Patients coming in for low-acuity complaints are more likely to wait while ER hospital staff care for other patients who have more urgent problems. The waits can be for four to six hours. If someone is having acute chest pain, obviously it is not the time to think about voter registration. But if people are asking for a strep test and would likely have to wait for several hours to be seen, that time could be leveraged. Why not ask patients if they want to use approximately ninety seconds of their waiting time to register to vote?
There is precedent for doing something like that. In 2008, the National Association of Community Health Centers ran a voter registration drive in health centers where patients received care. The result was that more than eighteen thousand low-income and middle-income citizens were added to the official rolls. Another program, conducted in 2012 at two Federally Qualified Health Centers in the Bronx, New York, showed that a large number of voters could be registered easily without requiring significant physician effort, creating undue political influence, or compromising patient-doctor relationships.
Many unregistered voters are highly receptive to the prospect of being offered a chance to register, with one study finding an 89 percent agreement rate among those approached in a hospital waiting room and asked whether or not they would be open to registering to vote.10 Hospitals have been initiating efforts to increase voter registration, such as MGH Votes, a campaign launched at Massachusetts General Hospital.11 MGH also adopted a program called VotER,12 which essentially eliminates sludge for voter registration, with a simple kiosk that allows registration in a very short time. That program has been duplicated in many other hospitals. Extending the VotER program to emergency rooms in general has extraordinary promise, and for one reason above all: it engages Americans who are not registered to vote by meeting them exactly where they are.
There are plenty of other ways to protect people’s voting rights by reducing sludge. Voting by mail eliminates what some people consider to be the burdens of getting to the polls and standing in line. So long as the risk of fraud can be contained (and current evidence suggests that it can), voting by mail is an excellent way to cut sludge with respect to a right that is central to self-government.
Federal law requires states to send mail-in forms (return cards) before purging voters from electoral rolls on change-of-residence grounds (if a voter has not already confirmed a move).13 That’s good; it reduces the risk of unjustified purges. At the same time, each state is allowed to choose its own trigger for sending the return card. Some states use change-of-address information provided by the US Postal Service, but others use methods that can very foreseeably flag voters who have in fact not moved and thus remain eligible.14 For example, a qualified voter can be struck for failing to mail the return card back and not voting for four years.15 Voters—along with Congress and the Supreme Court—may be optimistic that they will do that, but their optimism is almost certainly misplaced.16 Having to mail that card back is a form of sludge, and some people might forget to do so.
I will return to voting rights and sludge in chapter 4. The only point here is that the architecture for voting and voting registration can contain a lot of sludge or a little, and for the democratic process, the amount greatly matters.
In many contexts, sludge is produced by architects who know exactly what they are doing. But in other contexts, sludge has a significant impact that architects do not foresee. Many of them are unrealistically optimistic about the likelihood that people will overcome inertia. Even specialists are surprised at the extent to which apparently promising strategies for defeating sludge turn out to fail. For example, people might be reminded about what they have to do, but they might ignore reminders. In the private sector, sludge is used opportunistically by clever marketers who seek to give consumers the impression that they will receive an excellent deal but who know that people will not read the fine print or take advantage of an available opportunity.17
In government, sludge might have a damaging effect that public officials do not anticipate. In particular, officials might not understand the extent to which sludge will adversely affect a population that they are seeking to help. I can report, from personal experience (and as a lawyer), that lawyers are often the problem here. They impose sludge with the belief that it really is not much of a problem. Surely people can fill out a four-page form? Surely they can answer questions about their employment history and their places of residence over the past fifteen years? Maybe so, but maybe not, and maybe it is a lot harder than we lawyers think. Something similar can be said of the private sector—which is a reason that businesses, nonprofits, and educational institutions can benefit from Sludge Audits.
But no one should doubt that sometimes public officials impose sludge on purpose. They know exactly what they are doing. They are self-conscious choice architects, fully cognizant of the impact of sludge. They might agree to some costly program, designed to help poor people. But they might not love that program. They use sludge as a way of reducing costs or decreasing the real-world impact of the program. They are entirely aware that they are doing that. Sludge might be part of the price of a program designed to protect a vulnerable group that does not have a lot of political power. The program itself might seem exceedingly generous. But in the fine print, you can find the sludge.
In many cases, government officials are responding to political values and commitments, and they are not hiding anything. They impose sludge to reduce costs, to ensure that licensees are properly qualified, or to prevent benefits from going to people who do not really need or deserve them—points to which I will return.
In the private sector, similar things happen. Sludge might be a way of ensuring that people really qualify for something—say, a mortgage. Or it might be a self-serving effort to make relevant terms incomprehensible; if you have to wade through lengthy and complex materials, you might not understand key provisions.18 In some of the worst cases, people are automatically subjected to certain terms or conditions, or to certain economic obligations, without consenting in any meaningful sense; sludge is the basic problem. There is a good argument in favor of a new right: the right not to be manipulated. Sludge can be the mechanism by which that right is violated. And when the public sector uses sludge to manipulate people, something has gone badly wrong.
I have referred to the idea of dark patterns, which have been defined as “user interface design choices that benefit an online service by coercing, steering, or deceiving users into making unintended and potentially harmful decisions.”19 Dark patterns come from both public and private institutions. The category goes well beyond sludge. Deception and steering need not involve sludge. People could, for example, be nudged into making harmful choices, as when a costly or harmful path or outcome is made especially easy (a kind of dark pattern). In such cases, sludge may not be involved. Shrouded attributes, such as add-on costs, might be self-consciously hidden from consumers, but it is fair to question whether they count as sludge.20 Probably we should say that sludge is involved insofar as consumers have to do real work to learn about those attributes.
However we answer that question, a lot of dark patterns involve sludge, and they are very dark.
1. See generally Richard H. Thaler & Cass R. Sunstein, Nudge 83–105 (2008) (describing choice architecture).
2. Peter Bergman & Todd Rogers, The Impact of Defaults on Technology Adoption 5 (Harvard Kennedy Sch. Faculty Research Working Paper Series, Working Paper No. RWP17-021, 2018), https://scholar.harvard.edu/files/todd_rogers/files/bergman_and_rogers_the_impace_of_defaults.pdf (https://perma.cc/N7GF-BCY9).
3. Id.
4. Id.
5. See, for example, Brigitte C. Madrian & Dennis F. Shea, The Power of Suggestion: Inertia in 401(k) Participation and Savings Behavior, 116 Q.J. Econ. 1149, 1184 (2001) (summarizing behavioral changes resulting from 401(k) participation and savings behavior as a result of changing default options). For a discussion of the effect of inertia on choice of travel modes, see Alessandro Innocenti, Patrizia Lattarulo & Maria Grazia Pazienza, Heuristics and Biases in Travel Mode Choice 20 (LabSi, Working Paper No. 27/2009, 2009), http://www.labsi.org/wp/labsi27.pdf (https://perma.cc/P23F-42UL).
6. I draw here on joint work with Alister Martin, who hss been a pioneer in this area. See Alister Martin & Cass R. Sunstein, In the ER? Sign Up to Vote, Bos. Globe (Jan. 13, 2020), https://www.bostonglobe.com/2020/01/13/opinion/er-sign-up-vote/.
7. Thom File, US Census Bureau, No. P20–577, Who Votes? Congressional Elections and the American Electorate: 1978–2014.
8. Why Are Millions of Citizens Not Registered to Vote?, Pew Charitable Trs. (June 21, 2017), https://www.pewtrusts.org/en/research-and-analysis/issue-briefs/2017/06/why-are-millions-of-citizens-not-registered-to-vote.
9. James J. Augustine, The Latest Emergency Department Utilization Numbers Are In, ACEP Now (Oct. 20, 2019), https://www.acepnow.com/article/the-latest-emergency-department-utilization-numbers-are-in/.
10. Alisha Liggett et al., Results of a Voter Registration Project at Two Family Medicine Residency Clinics in the Bronx, New York, 12 Annals Fam. Med. 466 (2014).
11. MGH Votes!, Mass. Gen. Hosp. (Sept. 7, 2018), https://www.massgeneral.org/news/article/mgh-votes.
12. VotER, https://vot-er.org/ (last visited June 15, 2020).
13. National Voter Registration Act of 1993, 52 U.S.C. § 20507(d) (2012). This provision of the National Voter Registration Act, among other purposes, is aimed to “ensure that accurate and current voter registration rolls are maintained.” 52 U.S.C. § 20501(b)(4).
14. This is the practice suggested by federal law. See 52 U.S.C. § 20507(c)(1). Thirty-six states do at least this. See Nat’l Assn. of Sec’ys of State, NASS Report: Maintenance of State Voter Registration Lists 5–6 (2017) (Dec. 2017), https://www.nass.org/sites/default/files/reports/nass-report-voter-reg-maintenance-final-dec17.pdf (https://perma.cc/FXJ6-RPXK). See, for example, Iowa Code § 48A.28.3 (2018) (permitting the sending of notice each year); Ga. Code Ann. § 21-2-234(a)(1)–(2) (2018) (notice sent to registrants with whom there has been “no contact” for three years); Pa. Stat. Ann., tit. 25, § 1901(b)(3) (2018) (notice sent to voters who have not voted in five years); Ohio Rev. Code Ann. § 3503.21(B)(2) (2018) (notice sent to those who fail to vote in two consecutive federal elections). Note also that some states trigger notices based on dubious interstate databases. See, for example, Okla. Admin. Code § 230:15-11-19(a)(3) (2018) (notice sent to those who have not voted since the “second previous General Election” and those who fail references to interstate databases); Wis. Stat. Ann. § 6.50(1) (2018) (notice sent to voters who have not voted in four years). See also Jonathan Brater et al., Brennan Center for Justice at the New York University School of Law, Purges: A Growing Threat to the Right to Vote 7–8 (2018) (explaining how the system used by Oklahoma, “Crosscheck,” is unreliable and inaccurate).
15. See 52 U.S.C. § 20507(d)(1)(ii).
16. 52 U.S.C. § 20507(d) makes failure to send the return card back one of the two sufficient conditions for removing a registered voter from the rolls on change-of-address grounds. Husted v. A. Philip Randolph Inst., 138 S. Ct. 1833, 1845 (2018) (rejecting the argument that voters throw away return cards so often as to make them “worthless”).
17. For relevant discussion, see Petra Persson, Attention Manipulation and Information Overload, 2 Behav. Pub. Pol’y 78 (2018); Thomas Blake, Sarah Moshary, Kane Sweeney & Steven Tadelis, Price Salience and Product Choice (NBER, Working Paper No. 25186, 2018), https://www.nber.org/papers/w25186?sy=186 (https://perma.cc/Y54U-9K9S).
18. See Wendy Wagner, Incomprehensible! (2019).
19. Arunesh Mather et al., Dark Patterns at Scale: Findings from a Crawl of 11K Shopping Websites, 3 Proc. ACM Hum.-Comput. Interact. 81 (2019), https://arxiv.org/pdf/1907.07032.pdf.
20. See Xavier Gabaix & David Laibson, Shrouded Attributes, Consumer Myopia, and Information Suppression in Competitive Markets, 121 Q.J. Econ. 505 (2006).