INFORMATION OVERLOAD

As Druckman (Reference Druckman2019) observes in his article in this symposium, politically relevant information available to voters has gone from garden-hose to fire-hydrant proportions. Information scarcity is not a problem today, if it ever was.

Many scholars, including Kevin Esterling and Dmitri Landa in this symposium, think imaginatively about how we can capitalize as a society on unprecedented information flows between voters and policy makers. They lay out circumstances in which information can generate better collective choices: policy makers can gain a rich knowledge of voter preferences in virtual town hall meetings, for example, to arrive at better decisions.

However, radically decentralized democracy could run into problems. Condorcet’s jury theorem—essentially that “more heads are better than one”—relies on a process of aggregating the independent judgments of multiple people rather than “democratic deliberation” that could skew decisions toward the views of more persuasive, more influential, and more resourced actors (Waddington Reference Waddington2008). The key to accurate bull-weight guessing at the county fair, in the famous 1907 experiment, seems to rest on independent judgment rather than a meeting of the minds.^{Footnote 1}

As Lupia and McCubbins remind us, modern politics often takes place in environments unconducive to independent judgments. People may rely on flawed shortcuts or on information providers who may deliberately or unintentionally promote bias (Druckman, Fein, and Leeper Reference Druckman, Fein and Leeper2012, 441; Lord, Ross, and Lepper Reference Lord, Ross and Lepper1979; Schwartzberg Reference Schwartzberg2015; Sunstein Reference Sunstein2009; Tversky and Kahneman Reference Tversky and Kahneman1974). Glaeser and Sunstein (Reference Glaeser and Sunstein2014) pointed out that a profusion of balanced information—contrary to the optimistic view that information is democracy’s “disinfectant”—may be insufficient to overcome voters’ inclination to believe what confirms their previous biases and/or that comports with endorsements of people they trust. For these reasons, information technologies are unlikely to be a panacea for the democratic dilemma.

Massive information also can undermine verification, one of the key institutional fixes identified by Lupia and McCubbins (Reference Lupia and McCubbins1998). To put this in context, consider recent scholarship by Bullock and others (Bullock Reference Bullock2011; Bullock et al. Reference Bullock, Gerber, Hill and Huber2015) that suggests American voters use various types of information in addition to party cues in forming their judgments. This should not be surprising given that primaries and other forms of intraparty competition complicate party messaging. Seen in this light, the 2016 victory of Donald Trump could reflect voters’ ability to see past establishment cues to grasp their own economic insecurity and act on it (Morgan Reference Morgan2018; Mutz Reference Mutz2018). But this is hardly cause for celebrating the informational foundations of democratic accountability in America, given the precarious truth claims underlying Trump’s promises to redress their reasonable fears.

Lupia and McCubbins’s framework of strategically embedded voter choice should return to the conversation. They pointed us toward an analysis of conditions under which information can deceive (Lupia and McCubbins Reference Lupia and McCubbins1998, 98). People are more likely to believe Trump’s promises to protect American workers from foreign competition if they think that the naysayers are politically motivated to downplay the effectiveness of trade protectionism. The problem of Trump’s potent brand of populism indeed may owe to many voters’ genuine and well-grounded feelings of economic insecurity. But populism gains ground because coordination failures within political parties fail to rebut false claims, robbing voters of the opportunity to consider better choices. When individual politicians, even within the same party, contradict one another, voters find themselves in a more difficult information environment in which to evaluate competing claims. Stated in terms of Lupia and McCubbins’s (Reference Lupia and McCubbins1998) framework, the external availability of credible verification may be diminished today and the result might be increased deception and failed delegation. This is not an assertion per se but rather a possibility in need of study (Iyengar and Massey Reference Iyengar and Massey2018).

STRATEGY AND INFORMATION: MEASURING PARTY CUES

Lupia and McCubbins (1998, 207) drew attention to the way political institutions shape and anchor the value of the party cue to voters seeking reliable information shortcuts. The authors (1998, 206) stated: “How voters make these choices is inherent in the varying value and usage of party cues.” But party cues lose value for at least two reasons: (1) when they are internally divided; and (2) when there are so many parties that their connection to the process of legislative policy making is unclear. Druckman, Fein, and Leeper (2012, 447) corroborated this point in their consideration of the stability (and usefulness to voters) of information frames over time: competition among dueling factions complicates the understanding of which cues to trust and what they mean.

Normatively, fostering parties to be reliable cue-givers comports with Schattsneider’s (Reference Schattsneider1942) suggestion that the essence of democratic accountability is the ability of parties to offer and deliver on competing visions of the public good. Voters lack the time and knowledge to investigate the costs and benefits of every policy, much less think about how their own interests must weigh against those of other citizens. Downsian parties aiming at the political middle foster democratic accountability by considering the long-term consequences of each policy for every other goal or interest.

In practice, a difficulty with Schattsneider’s prescription for better cues is that only disciplined parties in a two-party system can credibly promise to deliver proposed policies if elected. The US system was designed with checks and balances to block “tyranny of the majority,” creating the avenue for competition along multiple policy dimensions. Within the legislative branch alone, primaries—particularly those that take place in safe districts—further fragment party messaging. Party leaders are hardpressed to insist on tight policy platforms when well-mobilized groups in one electoral district are at odds with those in other districts, disinclining legislators to submit to party platforms that aim at the overall best interests of the country as a whole. Southern Democrats’ withdrawal from the Democratic Party in the 1960s over differences about civil rights made the parties internally more coherent. However, primaries for legislative seats, which tend to be low-turnout elections, fragment the parties in new ways. Primaries pull politics away from the political middle by creating opportunities for Tea Party and other well-organized groups to gain a footing. Open-list proportional-representation electoral rules in a number of Latin American and former Eastern European countries are—if anything—even more damaging to party labels: candidates must compete against their copartisans for votes in the general election (Carey and Shugart Reference Carey and Shugart1995; Stein and Tommasi Reference Stein and Tommasi2008).

Another problem is when there are too many parties from which to choose. Parties in proportional-representation systems, even when they are closed-list, confront voters with complicated informational cues because they typically form governing coalitions after elections. In this case, voters cannot know in advance what voting for a particular party will mean for likely policy outcomes (Bawn and Rosenbluth Reference Bawn and Rosenbluth2006; Persson, Roland, and Tabellini Reference Persson, Roland and Tabellini2007).

The question then becomes how to evaluate the quality of democratic accountability in our (second-best) systems. What is needed, in the spirit of Lupia and McCubbins, is to measure how politicians’ strategic behavior in different institutional settings affects voters’ choices. Kernell (2018; forthcoming) found, for example, that allowing a wider set of voices to shape party platforms pulls policies away from the interests of the median voter. Another empirical strategy would be to enumerate vignette-based or conjoint survey experiments, in which voters are asked to choose between policies with randomly scrambled attributes. We might want to know, for example, the conditions under which voters favor broadly aimed policies with long-run payoffs—for example, education, health, and infrastructure—when they are traded off against the promise of low taxes (e.g., California’s Proposition 13). How do voters view tradeoffs across policies when they are made aware of the costs of one set of preferences over their other preferences? Do voters adjust their learning styles?

As Lupia and McCubbins (Reference Lupia and McCubbins1998) pointed out, people ascertain information from various sources (i.e., traditional media, parties/partisanship, neighbors, and friends), and they choose which sources to rely on based on reliability and alignment. How do different institutional environments determine which cues people use? Might individuals have more trust in parties with centralized nomination procedures? If so, do voters rely more on party cues in centralized systems and more on the media in decentralized systems? How do voters differentiate between party cues and candidate appeals? (See, e.g., Bawn et al. Reference Bawn, Cohen, Karol, Masket, Noel and Zaller2012; Hall and Thompson Reference Hall and Thompson2018; Hirano, Snyder, and Ting Reference Hirano, Snyder and Ting2009; Snyder, Hirano, and Ting Reference Snyder, Hirano and Ting2018.) Do individuals find new heuristics when institutions change? Lupia and McCubbins acknowledged heterogeneity among responses based on individual-level attributes, but this variation may reflect systematic, measurable differences in institutional settings.

Whatever we learn about voter preferences in different contexts, politicians in most institutional contexts cannot commit to policies that are good for most voters in the long run. A single-dimensional policy space is likely to be stable only in single-member district systems with two disciplined parties competing for the median voter (Ferejohn Reference Ferejohn and Grofman1993; Grofman Reference Grofman, Winer, Voigt and Grofman2017; Hinich and Munger Reference Hinich and Munger1996). In almost any other setting imaginable, politicians have incentives to introduce additional policy dimensions that cater to narrower groups of well-organized interests (Riker Reference Riker1986). However, if voters’ preferences are more closely approximated in single-issue–dimension settings, the implications for institutional reform could be substantial. Such a finding would argue in favor of eliminating primaries and redistricting in favor of competitive, demographically heterogeneous districts such that the median voter in each district is closer to the national median voter. This setup would push in the direction of endogenous party strengthening because representatives of each district would share with party leaders an incentive to be strategically moderate in favor of majority-benefiting policies (Cox Reference Cox1986; Rosenbluth and Shapiro Reference Rosenbluth and Shapiro2018).