Abstract

The Federal Reserve cut interest rates on March 3, 2020, in response to COVID-19. On March 5 and 6, I surveyed over 500 consumers about their concerns about COVID-19, awareness of the Fed's announcement, and macroeconomic expectations. Most consumers were concerned about effects of COVID-19 on the economy, their health, and their personal finances. About 38% were aware that the Fed had cut interest rates. Greater concern is associated with higher inflation expectations and more pessimistic unemployment expectations. I informed respondents about the Fed's announcement, which led some consumers to become more optimistic about unemployment and revise inflation expectations downward.

I. Introduction

ON March 3, 2020, the Federal Reserve lowered the federal funds rate target by 50 basis points to a range of 1% to 1.25%. This was the first rate cut made outside of a regularly scheduled Federal Open Markets Committee (FOMC) meeting since 2008. The FOMC statement noted that “the fundamentals of the U.S. economy remain strong. However, the coronavirus poses evolving risks to economic activity… . The Committee is closely monitoring developments and their implications for the economic outlook and will use its tools and act as appropriate to support the economy.”

In the press conference associated with the rate cut, chairman Jerome Powell added that the policy move would “help boost household and business confidence.” The spread of COVID-19 (coronavirus) is not neatly classified as either a demand or a supply shock (Cochrane, 2020), but it may result in both “practical and psychological” demand shocks, if consumers are prevented from getting to stores or postpone purchases in the face of huge uncertainty (Baldwin & di Mauro, 2020).

On March 5 and 6, 2020, I conducted an online survey using Amazon Mechanical Turk, a Web service that allows requesters to post small tasks in exchange for a posted monetary payment.1 I surveyed US consumers ages 18 and over about their attention to and concerns about the coronavirus, the news they heard about the Fed, and their expectations of inflation and unemployment. I then provided the respondents with the March 3 FOMC statement and information about the rate cut and resolicited their expectations of inflation and unemployment. I also collected information about respondents' demographics, numeracy, news sources, attention to the stock market, and confidence in the Fed and the president. Half of the consumers were provided with very brief information about coronavirus at the start of the survey. This treatment slightly increased health-related concerns but had no discernible effects on other outcomes.

Consumers were generally attentive to and concerned about the coronavirus; moreover, 28% had cancelled or postponed travel, and 40% had purchased food or supplies in response to these concerns. Concerns and responses vary with consumer characteristics. For example, respondents who own stocks or follow news about the stock market seem to be attentive to coronavirus news, more concerned, and more likely to have responded, and newspaper readers are also more concerned. However, much of the variation in consumer concern and response seems idiosyncratic, or not explained by basic demographic characteristics, numeracy, or even confidence in the president. Consumer characteristics also help predict awareness of the Fed's March 3 rate cut. Around 52% of consumers had heard news about the Fed in the past week, and 38% knew that the Fed had cut rates. Numerate consumers, stock owners, and print news readers were significantly more aware of the rate cut.

Note that when I conducted my survey, the effects of COVID-19 had not yet spread widely in the United States, and major business and school closures and stay-at-home orders had not yet happened. Thus, there was room for a good deal of heterogeneity in awareness of and concern about the virus. This heterogeneity is useful for allowing me to study the relationship of pessimism, information, and macroeconomic expectations. Greater concern about coronavirus is associated with higher inflation expectations and more pessimistic unemployment expectations. This is consistent with recent research showing that many consumers equate “bad times” with “high inflation” (Kamdar, 2019).2 Provision of information about the Fed announcement leads some consumers to become more optimistic about unemployment and revise inflation expectations downward. Consumers who were not already aware of the rate cut are more likely to revise their expectations in response to information about the rate cut. But overall, information about the announcement did not reduce disagreement, as consumers reacted heterogeneously to the information.

Since early March, awareness of and concern about the virus has grown, as far more people have experienced health and economic consequences. In late March and early April, for example, around 6 million new unemployment claims were filed per week (Coibion, Gorodnichenko, & Weber, 2020). Meanwhile, a survey-based literature on the COVID-19 outbreak and consumer beliefs, expectations, experiences, and preferences is rapidly emerging. Bu et al. (2020) conduct a repeated survey of a panel of graduate students in Wuhan, China, and find that exposure to strict quarantine led to more pessimistic beliefs about the economy, lower risk tolerance, and lower trust in others. Fetzer et al. (2020) conduct an online survey of US consumers on March 5 and 16 and find that concern about the virus grew from the earlier to the later survey date. They elicit respondents' subjective mental models of infectious disease spread and find that cognitive limitations (e.g., underestimation of the nonlinear nature of disease spread) affect individuals' economic anxieties associated with the COVID-19 pandemic. More recent surveys reveal strong partisan differences in social distancing and self-quarantining behavior and beliefs about the virus (Gadarian et al., 2020; Barrios & Hochberg, 2020; Allcott et al., 2020). Hanspal, Weber, and Wohlfart (2020) survey in April find that younger and poorer households face larger income shocks related to the pandemic and that households exposed to larger income losses are more likely to report plans to decrease total expenditures. Early April survey evidence from Coibion, Gorodnichenko, and Weber (2020) indicates that the COVID-19 crisis may be driving a wave of earlier-than-planned retirements.

This paper is also related to a broader recent literature that uses online experiments or surveys to study the formation of consumer expectations and response to information or Fed communication (Armantier et al., 2016; Binder & Rodrigue, 2018; Binder, 2020b; Coibion, Gorodnichenko et al., 2020). For example, Lamla and Vinogradov (2019) conduct a series of online surveys a few days before and after FOMC announcements and find that consumers are more likely to hear news about the Fed following an FOMC announcement, but the news does not appear to change their inflation and interest rate expectations. The announcement I study was made outside a regularly scheduled FOMC meeting, and therefore potentially more newsworthy: 35% of respondents in their surveys and 52% in mine had heard recent news about the Fed.

II. Survey Design

I ran the survey in several batches on March 5 and 6, to reach people in different time zones or with different work schedules. Following Allcott and Gentzkow (2017) and Binder and Rodrigue (2018), I allowed respondents to take the survey only if they answered the following question affirmatively:

We care about the quality of our data. In order for us to get the most accurate measures of your knowledge and opinions, it is important that you thoughtfully provide your best answers to each question in this survey. Do you commit to thoughtfully provide your best answers to each question in this survey?

A total of 520 respondents answered affirmatively and went on to complete the survey. I dropped 18 respondents who completed the survey in less than 2 minutes, leaving 502 respondents, who took the survey in 7.3 minutes on average.

The survey begins with questions about age, gender, education, household income, and stock market participation.3 Online appendix table A.1 summarizes basic demographic information of respondents. One-third of the sample is female, 21% have household income below $30,000 per year, and 26% have household income above $75,000 per year. I constructed survey weights to match the gender and income distribution of the national population, which I use in all of the analysis.

Next, respondents select their primary source(s) of news about the economy from social media, print sources or newspaper, online sources, television, and radio. They are asked, “On a scale from 1 to 7, how well would you say you understand what ‘inflation’ means?” They are also asked if they know the Fed's inflation target and, if so, to provide the number.

Respondents then answer a series of questions about their attention to and concerns about the coronavirus and news about the stock market and the Fed. Half of respondents, selected randomly, receive the following information about the coronavirus before answering these questions:

The World Health Organization (WHO) recently upgraded the global risk from the coronavirus outbreak to “very high.”

In the United States, cases have been confirmed in Arizona, California, Florida, Georgia, Illinois, Massachusetts, New Hampshire, New York, New Jersey, North Carolina, Oregon, Rhode Island, Texas, Washington and Wisconsin, according to researchers at Johns Hopkins University.

The other half receive no information. Questions and possible responses are as follows:

  • How closely have you been following the news about the coronavirus (Covid-19) outbreak? (Not closely at all, somewhat closely, very closely)

  • How concerned are you about the effects that the coronavirus might have on the US economy? (Not at all concerned, somewhat concerned, very concerned)

  • How concerned are you about the effects that the coronavirus might have on your health or the health of members of your household? (Not at all concerned, somewhat concerned, very concerned)

  • How concerned are you about the effects that the coronavirus might have on the financial situation of your household? (Not at all concerned, somewhat concerned, very concerned)

  • Have you cancelled or postponed any travel plans due to coronavirus concerns? (Yes, no)

  • Have you purchased food or supplies due to coronavirus concerns? (Yes, no)

  • How closely do you follow news about the stock market? (Not closely at all, somewhat closely, very closely)

  • In the past week, have you heard or read any news about the Federal Reserve? (Yes, no)

  • (If “yes” to previous question) What news did you hear or read about the Federal Reserve? (The Fed raised interest rates, the Fed cut interest rates, other news [please describe])

  • As to the economic policy of the government—I mean steps taken to fight inflation or unemployment—would you say the government is doing a good job, only fair, or a poor job? (Good job, only fair, poor job)4

Next, respondents provide their expectations of unemployment and inflation in the next twelve months, following the elicitation procedure of the Michigan Survey of Consumers.

  • How about people out of work during the coming twelve months–do you think that there will be more unemployment than now, about the same, or less? (More unemployment, about the same, less)

  • During the next twelve months, do you think that prices in general will go up, or go down, or stay where they are now? (Stay the same, lower, don't know)5

  • By about what percent per year do you expect prices to go (up/down) on the average during the next twelve months?

I next provide respondents with the information about the Fed's rate cut. The Federal Reserve issued the following statement on March 3, 2020:

The fundamentals of the U.S. economy remain strong. However, the coronavirus poses evolving risks to economic activity. In light of these risks and in support of achieving its maximum employment and price stability goals, the Federal Open Market Committee decided today to lower the target range for the federal funds rate by 1/2 percentage point, to 1 to 11/4 percent. The Committee is closely monitoring developments and their implications for the economic outlook and will use its tools and act as appropriate to support the economy.

In a press conference following the Federal Reserve's rate cut on March 3, the Federal Reserve chair said the following:

Monetary policy can be an effective tool to support overall economic activity. We do recognize that a rate cut will not reduce the rate of infection. It won't fix a broken supply chain. We get that. We don't think we have all the answers. But we do believe that our action will provide a meaningful boost to the economy. More specifically, it will support accommodative financial conditions and avoid a tightening of financial conditions which can weigh on activity, and it will help boost household and business confidence. That's why you're seeing central banks around the world responding as they see appropriate in their particular institutional context.

I reelicit unemployment and inflation expectations exactly as before. Then I ask respondents to report “how much confidence you have in each of the following to do or to recommend the right thing for the economy” for President Donald Trump and the Federal Reserve. Choices were “almost no confidence,” “a little confidence,” “a fair amount of confidence,” and “a great deal of confidence.” Respondents are also asked to identify the Fed chair, with the possible options of “Jerome Powell,” “Alan Blinder,” and “Alan Greenspan.” Respondents answer two numeracy test questions from the Federal Reserve Bank of New York's Survey of Consumer Expectations:

  • If the chance of getting a disease is 10 percent, how many people out of 1,000 would be expected to get the disease?

  • Imagine the interest rate on your savings account was 1% per year and inflation was 2% per year. After one year, how much would you be able to buy with the money in this account? (More than today, exactly the same, less than today)

I classify respondents as numerate if they answer both questions correctly. Finally, respondents could provide an open-ended response about “anything at all that you would like to add or to tell us about this survey.”

III. Concern about Coronavirus

Figure 1 summarizes respondents' attention to, concerns about, and responses to the coronavirus outbreak. Nearly all participants follow news about coronavirus—50% somewhat closely and 43% very closely. Consumers vary in how concerned they are about the effects of coronavirus on the national economy, their household's health, and their personal finances. Concerns about economic effects are most prevalent, with 52% somewhat concerned and 38% highly concerned. Consumers who follow news about the coronavirus more closely tend to be more concerned about the effects of the virus.
Figure 1.

Coronavirus Attention, Concern, and Response

Panel A shows the percent of consumers who report following news about coronavirus not closely at all, somewhat closely, or very closely. Panels B, C, and D show the percent who are not at all concerned, somewhat concerned, or very concerned about effects of coronavirus on the US economy, their household's health, and their personal finances. Panels E and F show the percent who have cancelled or postponed travel or purchased food or supplies in response to coronavirus concerns.

Figure 1.

Coronavirus Attention, Concern, and Response

Panel A shows the percent of consumers who report following news about coronavirus not closely at all, somewhat closely, or very closely. Panels B, C, and D show the percent who are not at all concerned, somewhat concerned, or very concerned about effects of coronavirus on the US economy, their household's health, and their personal finances. Panels E and F show the percent who have cancelled or postponed travel or purchased food or supplies in response to coronavirus concerns.

The bottom panels of figure 1 show how consumers have actually responded to their concerns. When asked, “Have you cancelled or postponed any travel plans due to coronavirus concerns?” 28% say yes, and 40% say they have “purchased food or supplies due to coronavirus concerns.” For consumers with greater concern about the effects of coronavirus, canceling travel or making purchases is more prevalent; for example, 45% of consumers with high concerns about effects on their household health have cancelled travel and 56% have purchased food or supplies.6

A. Predictors of Concern

Table 1 displays ordered probit regressions of the coronavirus attention, concern, and response variables on respondent characteristics. Women are more likely than men to follow coronavirus news, to be concerned about economic and personal financial effects, and to have made purchases in response to concerns; each of these marginal effect sizes is around 10 percentage points. D'Acunto, Malmendier, and Weber (2020) document a “gender expectations gap” for a range of macroeconomic and financial expectations and argue that this is attributable to gendered differences in grocery shopping and exposure to price signals. More generally, women tend to be more pessimistic than men in a variety of contexts (Dawson, 2017; Bjuggren & Elert, 2019).

Table 1.
Consumer Characteristics and Coronavirus Attention, Concerns, and Responses
 (1) News (2) Economy (3) Health (4) Finances (5) Travel (6) Purchases 
age −0.05 0.00 −0.01 0.02 −0.02 −0.03 
 (0.04) (0.04) (0.04) (0.04) (0.06) (0.05) 
ageSq 0.00 0.00 0.00 −0.00 0.00 0.00 
 (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) 
female 0.28** 0.30** 0.07 0.26** 0.04 0.33** 
 (0.14) (0.14) (0.14) (0.13) (0.18) (0.16) 
numerate −0.01 −0.30* −0.63*** −0.74*** −0.96*** −0.39** 
 (0.16) (0.16) (0.18) (0.14) (0.19) (0.18) 
stockowner 0.11 0.26 0.15 0.36** 0.79*** 0.66*** 
 (0.19) (0.18) (0.16) (0.17) (0.24) (0.19) 
stocknews 1.15*** 0.98*** 0.40* 0.89*** 0.28 −0.13 
 (0.21) (0.22) (0.23) (0.22) (0.25) (0.23) 
collegedegree 0.18 0.19 0.12 0.09 0.27 0.11 
 (0.18) (0.18) (0.17) (0.18) (0.23) (0.18) 
highincome −0.15 −0.16 −0.11 −0.31** −0.43** −0.24 
 (0.16) (0.15) (0.15) (0.14) (0.19) (0.18) 
lowincome −0.20 0.00 −0.16 0.08 −0.08 −0.18 
 (0.19) (0.17) (0.18) (0.17) (0.22) (0.20) 
socialmedia −0.08 0.05 0.02 0.10 0.66*** 0.25 
 (0.15) (0.14) (0.14) (0.13) (0.19) (0.19) 
print 0.38** 0.44** 0.19 0.40** 0.19 0.04 
 (0.18) (0.19) (0.21) (0.16) (0.25) (0.23) 
online 0.17 0.26* −0.14 0.02 0.09 0.05 
 (0.15) (0.15) (0.15) (0.14) (0.20) (0.19) 
radio 0.10 −0.21 0.06 −0.52** −0.28 0.15 
 (0.18) (0.19) (0.25) (0.23) (0.29) (0.25) 
tv 0.10 0.04 0.19 0.15 0.29 0.52*** 
 (0.14) (0.14) (0.16) (0.13) (0.20) (0.18) 
govgoodjob −0.04 −0.18 −0.14 −0.23 0.32 0.20 
 (0.16) (0.17) (0.18) (0.15) (0.20) (0.20) 
govpoorjob 0.28 0.12 0.41** 0.03 0.65** 0.33 
 (0.21) (0.19) (0.20) (0.20) (0.27) (0.22) 
N 498 498 497 499 499 499 
R2 pseudo 0.10 0.11 0.06 0.12 0.27 0.13 
 (1) News (2) Economy (3) Health (4) Finances (5) Travel (6) Purchases 
age −0.05 0.00 −0.01 0.02 −0.02 −0.03 
 (0.04) (0.04) (0.04) (0.04) (0.06) (0.05) 
ageSq 0.00 0.00 0.00 −0.00 0.00 0.00 
 (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) 
female 0.28** 0.30** 0.07 0.26** 0.04 0.33** 
 (0.14) (0.14) (0.14) (0.13) (0.18) (0.16) 
numerate −0.01 −0.30* −0.63*** −0.74*** −0.96*** −0.39** 
 (0.16) (0.16) (0.18) (0.14) (0.19) (0.18) 
stockowner 0.11 0.26 0.15 0.36** 0.79*** 0.66*** 
 (0.19) (0.18) (0.16) (0.17) (0.24) (0.19) 
stocknews 1.15*** 0.98*** 0.40* 0.89*** 0.28 −0.13 
 (0.21) (0.22) (0.23) (0.22) (0.25) (0.23) 
collegedegree 0.18 0.19 0.12 0.09 0.27 0.11 
 (0.18) (0.18) (0.17) (0.18) (0.23) (0.18) 
highincome −0.15 −0.16 −0.11 −0.31** −0.43** −0.24 
 (0.16) (0.15) (0.15) (0.14) (0.19) (0.18) 
lowincome −0.20 0.00 −0.16 0.08 −0.08 −0.18 
 (0.19) (0.17) (0.18) (0.17) (0.22) (0.20) 
socialmedia −0.08 0.05 0.02 0.10 0.66*** 0.25 
 (0.15) (0.14) (0.14) (0.13) (0.19) (0.19) 
print 0.38** 0.44** 0.19 0.40** 0.19 0.04 
 (0.18) (0.19) (0.21) (0.16) (0.25) (0.23) 
online 0.17 0.26* −0.14 0.02 0.09 0.05 
 (0.15) (0.15) (0.15) (0.14) (0.20) (0.19) 
radio 0.10 −0.21 0.06 −0.52** −0.28 0.15 
 (0.18) (0.19) (0.25) (0.23) (0.29) (0.25) 
tv 0.10 0.04 0.19 0.15 0.29 0.52*** 
 (0.14) (0.14) (0.16) (0.13) (0.20) (0.18) 
govgoodjob −0.04 −0.18 −0.14 −0.23 0.32 0.20 
 (0.16) (0.17) (0.18) (0.15) (0.20) (0.20) 
govpoorjob 0.28 0.12 0.41** 0.03 0.65** 0.33 
 (0.21) (0.19) (0.20) (0.20) (0.27) (0.22) 
N 498 498 497 499 499 499 
R2 pseudo 0.10 0.11 0.06 0.12 0.27 0.13 

Robust standard errors in parentheses. ***p<0.01, **p<0.05, and *p<0.10. Ordered probit (columns 1–4) and probit (columns 5–6) regressions. Dependent variables in columns 1 to 4 are categorical variables describing respondent's attention to news about coronavirus, concerns about effects of coronavirus on national economy, household health, and personal finances. Dependent variables in columns 5 and 6 are dummy variables indicating that the respondent has cancelled or postponed travel due to coronavirus concerns or has purchased food or supplies due to coronavirus concerns.

Respondents who own stocks or follow news about the stock market seem more attentive to coronavirus news, more concerned, and more likely to have responded. For example, owning stocks is associated with a 10 percentage point greater likelihood of being highly concerned about personal finances and following news about stocks with a 26 percentage point greater likelihood. Stock prices began falling in late February, so by March 5, stock owners could have already experienced substantial losses of wealth. High-income respondents are 9 percentage points less likely to be highly concerned about effects on their personal finances, after controlling for stock market participation. These respondents likely have greater job security and ability to work from home. They may be salaried rather than hourly workers and thus are less likely to face a major loss of wages.

Respondents' level of concern also depends on where they get their news. Readers of print news (including newspapers) are more attentive to and concerned about coronavirus, perhaps because newspapers cover the economy more than other media platforms and frequently exhibit “negativity bias” (Soroka & McAdams, 2015; Binder, 2017b). Social media news consumers are more likely to have cancelled travel plans. Social media users are more likely to share travel experiences and recommendations and collaborate on travel planning in online communities, which might make them more aware of health risks in their travel destinations (Cahyanto et al., 2011).

Consumers with a poor opinion of the government's economic policies are more concerned about the coronavirus (though only the coefficient on health concern is statistically significant) and more likely to have cancelled travel. Opinion of government economic policy may be a proxy for political party, as many consumers blame or credit the president for the state of the economy (Binder, 2017a).7 Poor opinion of government economic policy may go hand in hand with low confidence in the government's ability to manage a public health crisis. Fetzer et al. (2020) find that Democrats are more concerned about the COVID-19 crisis than Republicans. Finally, note that the pseudo-R2 of the regressions is low. Thus, concern about coronavirus seems to be largely idiosyncratic.

B. Information Treatment and Concern

Recall that I randomly provided half of the respondents with information from the World Health Organization and John Hopkins University about the coronavirus. Table 2 shows ordered probit regressions of coronavirus-related concern on the treatment dummy and respondent characteristics.8 The treatment is associated with a statistically insignificant increase in health and personal finance–related concern. Columns 4 to 6 of the table restrict the sample to respondents who follow coronavirus news somewhat closely or not at all (excluding respondents who follow the news very closely). These respondents should be more susceptible to the information treatment, since they are less likely to be already aware of the information. Indeed, the treatment effect on health concerns is larger and statistically significant. The marginal effect implies that a respondent who receives the information treatment is 11 percentage points more likely to be somewhat or very concerned about the effects of coronavirus on household health.9

Table 2.
Response of Coronavirus Concerns to Information Treatment
 (1) Economy (2) Health (3) Finances (4) Economy (5) Health (6) Finances 
Treated −0.05 0.23 0.15 −0.06 0.40** 0.23 
 (0.14) (0.14) (0.13) (0.18) (0.19) (0.16) 
age 0.01 −0.01 0.02 0.10* 0.11** 0.09* 
 (0.04) (0.04) (0.04) (0.05) (0.05) (0.05) 
ageSq 0.00 0.00 −0.00 −0.00 −0.00* −0.00 
 (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) 
female 0.29** 0.07 0.27** 0.11 0.03 0.22 
 (0.14) (0.14) (0.14) (0.17) (0.18) (0.17) 
numerate −0.26* −0.55*** −0.70*** −0.36* −1.01*** −0.78*** 
 (0.15) (0.17) (0.14) (0.20) (0.20) (0.19) 
stockowner 0.22 0.10 0.30* 0.23 0.23 0.50** 
 (0.18) (0.16) (0.17) (0.21) (0.20) (0.20) 
stocknews 0.93*** 0.32 0.83*** 0.77* −0.43 0.86*** 
 (0.22) (0.22) (0.22) (0.45) (0.31) (0.29) 
collegedegree 0.14 0.10 0.09 −0.09 −0.05 −0.04 
 (0.18) (0.17) (0.18) (0.23) (0.20) (0.23) 
highincome −0.12 −0.10 −0.28** −0.35* −0.09 −0.24 
 (0.15) (0.14) (0.14) (0.19) (0.20) (0.18) 
lowincome 0.05 −0.08 0.11 −0.35* −0.16 0.00 
 (0.18) (0.17) (0.18) (0.20) (0.24) (0.21) 
socialmedia 0.04 0.01 0.10 0.21 0.19 0.38** 
 (0.13) (0.14) (0.13) (0.16) (0.19) (0.16) 
print 0.45** 0.18 0.36** 0.23 −0.05 0.44** 
 (0.18) (0.21) (0.16) (0.18) (0.22) (0.22) 
online 0.29** −0.06 0.06 0.24 −0.04 0.08 
 (0.14) (0.14) (0.14) (0.17) (0.17) (0.17) 
radio −0.20 0.05 −0.52** 0.21 0.25 −0.34 
 (0.19) (0.26) (0.24) (0.23) (0.25) (0.28) 
tv 0.03 0.18 0.14 0.14 0.28 0.20 
 (0.14) (0.15) (0.13) (0.17) (0.18) (0.16) 
N 498 497 499 298 297 299 
Pseudo-R2 0.10 0.06 0.12 0.10 0.12 0.15 
Sample All All All Low news Low news Low news 
 (1) Economy (2) Health (3) Finances (4) Economy (5) Health (6) Finances 
Treated −0.05 0.23 0.15 −0.06 0.40** 0.23 
 (0.14) (0.14) (0.13) (0.18) (0.19) (0.16) 
age 0.01 −0.01 0.02 0.10* 0.11** 0.09* 
 (0.04) (0.04) (0.04) (0.05) (0.05) (0.05) 
ageSq 0.00 0.00 −0.00 −0.00 −0.00* −0.00 
 (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) 
female 0.29** 0.07 0.27** 0.11 0.03 0.22 
 (0.14) (0.14) (0.14) (0.17) (0.18) (0.17) 
numerate −0.26* −0.55*** −0.70*** −0.36* −1.01*** −0.78*** 
 (0.15) (0.17) (0.14) (0.20) (0.20) (0.19) 
stockowner 0.22 0.10 0.30* 0.23 0.23 0.50** 
 (0.18) (0.16) (0.17) (0.21) (0.20) (0.20) 
stocknews 0.93*** 0.32 0.83*** 0.77* −0.43 0.86*** 
 (0.22) (0.22) (0.22) (0.45) (0.31) (0.29) 
collegedegree 0.14 0.10 0.09 −0.09 −0.05 −0.04 
 (0.18) (0.17) (0.18) (0.23) (0.20) (0.23) 
highincome −0.12 −0.10 −0.28** −0.35* −0.09 −0.24 
 (0.15) (0.14) (0.14) (0.19) (0.20) (0.18) 
lowincome 0.05 −0.08 0.11 −0.35* −0.16 0.00 
 (0.18) (0.17) (0.18) (0.20) (0.24) (0.21) 
socialmedia 0.04 0.01 0.10 0.21 0.19 0.38** 
 (0.13) (0.14) (0.13) (0.16) (0.19) (0.16) 
print 0.45** 0.18 0.36** 0.23 −0.05 0.44** 
 (0.18) (0.21) (0.16) (0.18) (0.22) (0.22) 
online 0.29** −0.06 0.06 0.24 −0.04 0.08 
 (0.14) (0.14) (0.14) (0.17) (0.17) (0.17) 
radio −0.20 0.05 −0.52** 0.21 0.25 −0.34 
 (0.19) (0.26) (0.24) (0.23) (0.25) (0.28) 
tv 0.03 0.18 0.14 0.14 0.28 0.20 
 (0.14) (0.15) (0.13) (0.17) (0.18) (0.16) 
N 498 497 499 298 297 299 
Pseudo-R2 0.10 0.06 0.12 0.10 0.12 0.15 
Sample All All All Low news Low news Low news 

Robust standard errors in parentheses. ***p<0.01, **p<0.05, and *p<0.10. Ordered probit regressions. Dependent variables are categorical variables describing respondent's attention to news about coronavirus, concerns about effects of coronavirus on national economy, household health, and personal finances. “Treated” indicates that the respondent received information about the coronavirus prior to reporting her concerns. In columns 4 to 6, sample is restricted to respondents who follow coronavirus news somewhat closely or not at all.

IV. Awareness of Fed Policy

Previous literature documents that consumer knowledge of the Fed, monetary policy, interest rates, and inflation is quite limited and heterogeneous (Kumar et al., 2015; Binder, 2017a; Coibion et al., 2019; Coibion, Gorodnichenko et al., 2020), and that neither households' nor firms' expectations respond much to monetary policy announcements in low-inflation economies (Coibion, Gorodnichenko et al., 2020). Consistent with this literature, I find incomplete and heterogeneous consumer knowledge of the March 3 rate cut, the Fed chair, and the inflation target.

Of the 52% of respondents who had heard news about the Federal Reserve in the past week, most (73%) knew that the Fed had cut interest rates, though 25% thought that the Fed had raised interest rates. The remaining five respondents described other news that they had heard about the Fed; one mentioned “repo operations are continuing” and the others were vague or uncertain (e.g., “Not sure, as I didn't read the article. Something was going on.”) Overall, 38% of respondents knew that the Fed had cut rates.

I find that 79% of respondents select the correct Fed chair, compared to 70% in 2019 (Binder, 2020b) and 67% in 2017 (Binder & Rodrigue, 2018). The share who knew that the Fed's inflation target is 2% increased to 44%, versus 32% in 2019 (Binder, 2020b) and 26% in 2017 (Binder & Rodrigue, 2018).10

Knowledge of the rate cut, the inflation target, and the Fed chair are moderately positively correlated with each other and depend on certain consumer characteristics (online appendix tables A.6 and A.7). For example, a numerate consumer is 22 percentage points more likely, and a stock owner 15 percentage points more likely, to be aware of the rate cut.11 Print news readers are more likely to have heard of the rate cut, while TV news consumers are less likely. Binder (2017b) finds that newspapers are more likely to cover the Federal Reserve than are cable and network television. Neither attention to nor concerns about coronavirus are associated with greater awareness of the rate cut. Knowledge of the inflation target and Fed chair is associated with similar characteristics but also positively associated with income.

V. Macroeconomic Expectations

In the first round of expectations elicitation, mean and median inflation expectations are, respectively, 3.12% and 2%, with an interquartile range of 0% to 4%.12 “Don't know” responses are given by 23.2%, and 44.7% of forecasts are multiples of 5%—indicative of high uncertainty (see Binder, 2017c). Only 6.5% of respondents expect lower unemployment in twelve months; 57.4% expect around the same amount of unemployment, and 36.1% expect more unemployment.

As shown in table 3, concerns about coronavirus are associated with more pessimistic unemployment expectations and higher inflation expectations. For example, mean inflation expectations are 2.0% for unconcerned consumers and 3.4% for very concerned consumers. This is consistent with recent research showing that consumer pessimism is associated with higher inflation expectations.

Table 3.
Macroeconomic Expectations by Concerns about Coronavirus Effects on the Economy
 Not Concerned Somewhat Concerned Very Concerned 
Less unemployment 0.17 0.035 0.079 
More unemployment 0.19 0.33 0.45 
Expected inflation 2.04 3.16 3.35 
Multiple of 5% inflation 0.4 0.44 0.47 
 Not Concerned Somewhat Concerned Very Concerned 
Less unemployment 0.17 0.035 0.079 
More unemployment 0.19 0.33 0.45 
Expected inflation 2.04 3.16 3.35 
Multiple of 5% inflation 0.4 0.44 0.47 

The table summarizes twelve-month-ahead unemployment and inflation expectations for consumers reporting that they are not at all concerned, somewhat concerned, or very concerned about the effects of coronavirus on the US economy.

A. Effects of Information about Monetary Policy

The Federal Reserve's rate cut and announcement on March 3 could have affected consumer expectations in various ways. On one hand, the Fed's response may have signaled that the economic outlook was worse than consumers previously knew (the information channel of monetary policy) (Nakamura & Steinsson, 2018). On the other hand, the rate cut itself, reassurance that “the fundamentals of the U.S. economy remain strong,” and the promise to “act as appropriate to support the economy” may have reassured consumers, making them more optimistic. The statement that “our action … will help boost consumer confidence” may also have persuaded respondents to report more optimistic expectations. Yet another possibility is that the length and complexity of the statement may have discouraged respondents from reading it carefully or inhibited their comprehension of the treatment. However, Coibion et al. (2019) found that providing consumers with the March or May 2018 FOMC statement—“written in the dense language that is typical of central bank communications”—had similar effects on subjects' inflation expectations as simpler information treatments, such as telling respondents the inflation target or recent inflation. Moreover, even respondents with lower income and education levels revised their expectations in response to those FOMC statements.

I examine possible effects of the rate cut and announcement in several ways. First, I compare the expectations of consumers with and without prior knowledge of the Fed's rate cut. As discussed earlier, prior knowledge of the rate cut is associated with certain respondent characteristics, such as numeracy and stock ownership, which may also affect macroeconomic expectations. Macroeconomic expectations may also directly influence awareness of the rate cut; for example, consumers who were especially pessimistic about future unemployment might have deliberately sought out news about Federal Reserve policy. Second, I examine how respondents' expectations change when I provide them with information about the rate cut. This information treatment is provided to all respondents, regardless of their observable or unobservable characteristics.

Panels A and B of figure 2 show that consumers with prior knowledge of the rate cut are more pessimistic about future unemployment.13 They also have higher inflation expectations: the mean is 3.7% for consumers who know about the rate cut versus 2.8% for consumers who do not. When controlling for respondents' observable characteristics, however, the differences in macroeconomic expectations by prior knowledge of the rate cut are not statistically significant. Panels C and D summarize revisions to unemployment expectations following the treatment for respondents with and without prior knowledge of the rate cut. About 80% of respondents with prior knowledge and 71% without prior knowledge make no revision to their unemployment expectations. Conditional on revising, 60% of respondents with prior knowledge and 47% without prior knowledge become more optimistic about future unemployment.
Figure 2.

Expectations and Revisions by Knowledge of Rate Cut

Panels A and B show initial expectations of twelve-month-ahead unemployment for consumers who do not know or know that the Fed cut rates on March 3. Panels C through F show revisions of expectations of twelve-month-ahead inflation and unemployment after provision of information about the Fed rate cut on March 3, for consumers with or without prior knowledge of the rate cut.

Figure 2.

Expectations and Revisions by Knowledge of Rate Cut

Panels A and B show initial expectations of twelve-month-ahead unemployment for consumers who do not know or know that the Fed cut rates on March 3. Panels C through F show revisions of expectations of twelve-month-ahead inflation and unemployment after provision of information about the Fed rate cut on March 3, for consumers with or without prior knowledge of the rate cut.

Panels E and F of figure 2 summarize qualitative revisions to inflation expectations. About 64% of respondents with prior knowledge of the rate cut and 40% without prior knowledge make no revision. Conditional on revising, 73% of respondents with prior knowledge and 58% of respondents without prior knowledge revised down. Quantitative inflation expectations, on aggregate, did not respond to the treatment: the pretreatment mean and median were 3.2% and 2%, while the posttreatment mean and median were 2.9% and 2%. However, this lack of aggregate response reflects the heterogeneity of individuals' responses. Online appendix figure A.2 shows kernel density estimates of quantitative revisions to inflation expectations for respondents with and without prior knowledge of the rate cut. Revisions are centered around 0 but with mass on either side, including extra mass at 5% and -5% from uncertain consumers who select “round” forecasts both before and after the treatment (Binder, 2017c).

Online appendix table A.8 summarizes inflation expectations revisions based on unemployment expectations revisions. Among consumers who became more optimistic about unemployment following information about the Fed's response, 59% revised their inflation expectations down and only 11% revised their inflation expectations up. In contrast, for consumers who became more pessimistic about unemployment—who interpreted the Fed's response as a signal of the poor state of the economy—28% revised their inflation expectations down and 47% revised up. This is consistent with what Andre et al. (2019) call the “good-bad-heuristic”: many consumers consider both inflation and unemployment “bad” and therefore expect them to comove (also see Kamdar, 2019, and Coibion, Gorodnichenko et al., 2020). Andre et al. find that consumers' predictions about how unemployment will respond to various shocks, including interest rate shocks, are generally in line with experts' predictions, but consumers' predictions about inflation are not.

It is interesting to note that consumer disagreement about inflation did not decline from the first to the second elicitation. Pretreatment and posttreatment disagreement about inflation was identical, with an interquartile range of 0% to 4% in both cases.14 The signal about the Fed's March 3 announcement and rate cut appears to be an exception to the finding in Coibion et al. (2019) that “there is in general little variation in terms of how different types of consumers respond to most signals: conditional on their initial beliefs (which do differ across groups), the way they respond to a common signal is broadly similar. This pattern in updating yields declines in disagreement across agents after each treatment.” The COVID-19 crisis and related policy responses are rare events, associated with high uncertainty, making it difficult for consumers to understand how to update their beliefs about the future (Gallagher, 2014; Mackowiak & Wiederholt, 2018). As a result, there is more variation in how consumers respond to the Fed's March 3 announcement compared to the information treatments in Coibion et al. (2019).

VI. Discussion and Conclusion

As of March 5 and 6, 2020, many consumers were concerned about the potential effects of coronavirus on their health and finances and on the US economy. These concerns were associated with more pessimistic unemployment expectations and higher inflation expectations. My results suggest that possible increases in consumer inflation expectations in the next few months of the COVID-19 crisis might best be interpreted as increases in pessimism rather than as improved expecations of aggregate demand.

The Federal Reserve's emergency rate cut on March 3 was relatively newsworthy, and 38% of consumers became aware of the cut. But consumers had mixed responses to learning about the Fed's announcement, and disagreement about inflation expectations did not decline. The fact that fewer than half of consumers were aware of such a major policy move and may have had trouble interpreting it points to large challenges in the central bank's efforts to communicate with the general public. These challenges will become especially important to address with the federal funds rate at the zero lower bound.

Larger-scale surveys will be helpful in the upcoming months as economic and public health conditions rapidly evolve, and especially as consumers experience more direct effects of the COVID-19 crisis on their lives. Panel surveys to facilitate analysis of the persistence of effects on beliefs and expectations will be especially valuable.15 Shortly after this survey, long lines in grocery stores and shortages of toilet paper and other items became widespread. This may exacerbate pessimism and high inflation expectations, as D'Acunto et al. (2019, 2020) find that many consumers extrapolate from their grocery shopping experiences to form inflation expectations. Job insecurity and job loss, illness, school and business closures, and future fiscal and monetary policy responses may have notable effects on expectations and beliefs.

Notes

1

Mechanical Turk allows for recruitment of subject pools that are more nationally representative than typical convenience samples, making it a popular choice for social science experiments (Berinsky, Huber, & Lenz, 2012; Casler, Bickel, & Hackett, 2013; Levay, Freese, & Druckman, 2016).

2

For example, a recent decline in inflation expectations on the Michigan Survey of Consumers reflected improved macroeconomic conditions and consumer confidence (Binder, 2020a). Many consumers also have a “1970s model” of the economy and interpret rising gas prices as both inflationary and a sign of low economic activity (Binder & Makridis, 2020).

3

Many of the survey questions follow Binder and Rodrigue (2018) and Binder (2020b). The household income question asks, “Which category represents the total combined pre-tax income of all members of your household (including you) during the past 12 months? Please include money from all jobs, net income from business, farm or rent, pensions, interest on savings or bonds, dividends, social security income, unemployment benefits, Food Stamps, workers compensation or disability benefits, child support, alimony, scholarships, fellowships, grants, inheritances and gifts, and any other money income received by members of your family who are 15 years of age or older.” The stock market participation question asks, “Do you (or any member of your family living there) have any investments in the stock market, including any publicly traded stock that is directly owned, stocks in mutual funds, stocks in any of your retirement accounts, including 401(K)s, IRAs, or Keogh accounts?” Wording is from the Michigan Survey of Consumers.

4

This question is from the Michigan Survey of Consumers.

5

“Stay the same” and “don't know” responses prompt further questioning. See “Survey of Consumers Questionnaire” (University of Michigan Survey Research Center, n.d.) codebook for details.

6

Online appendix table A.2 summarizes the correlations between each of these concerns and response-related variables.

7

Online appendix table A.3 summarizes the correlations between opinion of the government's economic policies, confidence in the president, and confidence in the Federal Reserve. All three are positively correlated. Opinion of government economic policy is more strongly correlated with confidence in the president.

8

Summary statistics of respondent characteristics for the treatment and control groups are in online appendix table A.4. I do not include the other coronavirus-related variables (news, travel cancellations, and purchases) as outcome variables because they should not plausibly respond to the treatment. I have verified that they do not respond to the treatment.

9

The information treatment mentions specific states that had reported cases at the time. I use the IP addresses of the users to construct a proxy for their state of residence; 59% of respondents live in the states mentioned in the information treatment. I construct a dummy variable S indicating that the respondent lives in a state mentioned in the information treatment. To test whether the effect of the information treatment is stronger for participants living in the mentioned states, in online appendix table A.5, I regress key survey responses (related to coronavirus concerns, macroeconomic expectations, and opinion of government policy) on S, the treatment dummy, and their interaction, along with the demographic control variables included in the other regressions. I do not find a statistically significant coefficient on S or the interaction in any case. It may be that the sample is too small or that the IP address-based proxy is too noisy.

10

Coibion, Gorodnichenko, and Weber (2019) find that less than 20% of consumers guess that the inflation target is 2%, and almost 40% guess that the target is 10% or greater. My results are not directly comparable to those of Coibion et al., since I first ask respondents if they know the Fed's target and ask for a guess only if they respond affirmatively. Among respondents who claim to know the Fed's target, 11% report a target that is larger than 10%. Knowledge of the inflation target may be gradually increasing over time but nonetheless remains low.

11

Among respondents who own stocks, 26% say they follow news about the stock market very closely. Following stock market news closely is also associated with greater awareness of the rate cut, though the effect is not statistically significant.

12

Inflation forecasts of 50% or more in absolute value were recoded as “don't know” responses. This accounted for 48 responses in the first round and 54 in the second round.

13

Online appendix figure A.1 replicates figure 2, omitting respondents who took the survey in less than four minutes, who might not have taken the time to read the full text of the treatment. The figure is nearly identical.

14

Coibion et al. (2019) note that higher moments of inflation expectations are sensitive to outliers and suggest using robust measures of disagreement in place of variance or standard deviation. The interquartile range is a robust measure of disagreement. Another robust measure, the median absolute deviation, is also identical pre- and posttreatment.

15

In more normal times, information treatments about inflation and Federal Reserve policy have only mildly persistent effects on consumer expectations (Coibion et al., 2019).

REFERENCES

Allcott
,
Hunt
,
Levi
Boxell
,
Jacob C.
Conway
,
Matthew
Gentzkow
,
Michael
Thaler
, and
David Y.
Yang
, “
Polarization and Public Health: Partisan Differences in Social Distancing during the Coronavirus Pandemic
,”
NBER working paper
26946
(
2020
).
Allcott
,
Hunt
, and
Mathew
Gentzkow
, “
Social Media and Fake News in the 2016 Election,
Journal of Economic Perspectives
31
(
2017
),
211
236
.
Andre
,
Peter
,
Carlo
Pizzinelli
,
Christopher
Roth
, and
Johannes
Wohlfart
, “
Subjective Models of the Macroeconomy: Evidence from Experts and a Representative Sample
,”
working paper
(
2019
), https://papers.srn.com/s013/papers.cfm?abstract_id=3355356.
Armantier
,
Olivier
,
Scott
Nelson
,
Giorgio
Topa
,
Wilbert
van der Klaauw
, and
Basit
Zafar
, “
The Price Is Right: Updating Inflation Expectations in a Randomized Price Information Experiment
,” this review 98 (
2016
),
503
523
.
Baldwin
,
Richard
, and
Beatrice Weder
di Mauro
, eds.,
Economics in the Time of COVID-19
(
London
:
CEPR Press
,
2020
).
Barrios
,
John M.
, and
Yael
Hochberg
, “
Risk Perception through the Lens of Politics in the Time of the COVID-19 Pandemic
,”
NBER working paper
27008
(
2020
).
Berinsky
,
Adam
,
Greg
Huber
, and
Gabe
Lenz
, “
Evaluating Online Labor Markets for Experimental Research: Amazon.com's Mechanical Turk,
Political Analysis
20
(
2012
),
351
368
.
Binder
,
Carola
, “
Fed Speak on Main Street: Central Bank Communication and Household Expectations,
Journal of Macroeconomics
52
(
2017a
),
238
251
.
Binder
,
Carola
Federal Reserve Communication and the Media,
Journal of Media Economics
30
(
2017b
),
191
214
.
Binder
,
Carola
Measuring Uncertainty Based on Rounding: New Method and Application to Inflation Expectations,
Journal of Monetary Economics
90
(
2017c
),
1
12
.
Binder
,
Carola
Long-Run Inflation Expectations in the Shrinking Upper Tail,
Economics Letters
186
(
2020a
),
1
4
.
Binder
,
Carola
Presidential Antagonism and Central Bank Credibility
,”
working paper
(
2020b
), https://papers.srn.com/s013/papers.cfm?abstract_id=3545379.
Binder
,
Carola
, and
Christos
Makridis
, “
Stuck in the Seventies: Gas Prices and Consumer Sentiment Stuck in the Seventies
,”
working paper
(
2020
).
Binder
,
Carola
, and
Alex
Rodrigue
, “
Household Informedness and Long-Run Inflation Expectations: Experimental Evidence,
Southern Economic Journal
85
(
2018
),
580
598
.
Bjuggren
,
Carl
Magnus
, and
Niklas
Elert
, “
Gender Differences in Optimism,
Applied Economics
51
(
2019
),
5160
5173
.
Bu
,
Di
,
Tobin
Hanspal
,
Yin
Liao
, and
Yong
Liu
, “
Economic Preferences during a Global Crisis: Evidence from Wuhan
,” working paper (
2020
), https://papers.srn.com/s013/papers.cfm?abstract_id=3559870.
Cahyanto
,
Ignatius
,
Lori
Pennington-Gray
,
Laura
Mandala
, and
Ashley
Schroeder
, “
The Effects of Social Media Usage on Travel Information Searching and Travel Experience Sharing
,”
Advancing Tourism Research Globally,
42
(
2011
).
Casler
,
K.
,
L.
Bickel
, and
E.
Hackett
, “
Separate But Equal? A Comparison of Participants and Data Gathered via Amazon's MTurk, Social Media, and Face-to-Face Behavioral Testing,
Computers in Human Behavior
29
(
2013
),
2156
2160
.
Cochrane
,
John H.
,
Coronavirus Monetary Policy
(
London
:
CEPR Press
,
2020
).
Coibion
,
Olivier
,
Yuriy
Gorodnichenko
,
Saten
Kumar
, and
Mathieu
Pedemonte
, “
Inflation Expectations as a Policy Tool?
Journal of International Economics
124:C
(
2020
).
Coibion
,
Olivier
,
Yuriy
Gorodnichenko
, and
Michael
Weber
, “
Monetary Policy Communications and Their Effects on Household Inflation Expectations
,”
NBER working paper 2020-07
(
2019
)
Coibion
,
Olivier
,
Yuriy
Gorodnichenko
, and
Michael
Weber
Labor Markets during the COVID-19 Crisis: A Preliminary View
,”
NBER working paper
27017
(
2020
).
D'Acunto
,
Francesco
,
Ulrike
Malmendier
,
Juan
Ospina
, and
Michael
Weber
, “
Exposure to Daily Price Changes and Inflation Expectations
,”
NBER working paper
26327
(
2019
).
D'Acunto
,
Francesco
,
Ulrike
Malmendier
, and
Michael
Weber
, “
Gender Roles and the Gender Expectations Gap
,”
NBER working paper
26837
(
2020
).
Dawson
,
Chris
, “
The Upside of Pessimism: Biased Beliefs and the Paradox of the Contented Female Worker,
Journal of Economic Behavior and Organization
135
(
2017
),
215
228
.
Fetzer
,
Thiemo
,
Lukas
Hensel
,
Johannes
Hermle
, and
Christopher
Roth
,
Perceptions of Pandemic Risk Factors and Economic Anxiety: Evidence from the Coronavirus Pandemic
(
2020
), https://arxiv.org/abs/2003.03846.
Gadarian
,
Shana Kushner
,
Sara Wallace
Goodman
, and
Thomas B.
Pepinsky
, “
Partisanship, Health Behavior, and Policy Attitudes in the Early Stages of the COVID-19 Pandemic
,”
working paper
(
2020
), https://papers.srn.com/s013/papers.cfm?abstract_id=3562796.
Gallagher
,
Justin
, “
Learning About an Infrequent Event: Evidence from Flood Insurance Take-Up in the United States,
American Economic Journal: Applied Economics
6
(
2014
),
206
233
.
Hanspal
,
Tobin
,
Annika
Weber
, and
Johannes
Wohlfart
, “
Income and Wealth Shocks and Expectations during the COVID-19 Pandemic
,”
CESito working paper
8244
(
2020
).
Kamdar
,
Rupal
, “
The Inattentive Consumer: Sentiment and Expectations
,”
meeting paper from Society for Economic Dynamics
(
2019
).
Kumar
,
Saten
,
Hassan
Afrouzi
,
Olivier
Coibion
, and
Yuriy
Gorodnichenko
, “
Inflation Targeting Does Not Anchor Inflation Expectations: Evidence from Firms in New Zealand
,”
Brookings Papers on Economic Activity
(
2015
),
151
206
.
Lamla
,
Michael J.
, and
Dmitri V.
Vinogradov
, “
Central Bank Announcements: Big News for Little People?
Journal of Monetary Economics
108
(
2019
),
21
38
.
Levay
,
Kevin E.
,
Jeremy
Freese
, and
James N.
Druckman
, “
The Demographic and Political Composition of Mechanical Turk Samples
,”
SAGE Open
(
2016
),
1
17
.
Mackowiak
,
Bartosz
, and
Mirko
Wiederholt
, “
Lack of Preparation for Rare Events,
Journal of Monetary Economics
100
(
2018
),
35
47
.
Nakamura
,
Emi
, and
Jon
Steinsson
, “
High-Frequency Identification of Monetary Non- Neutrality: The Information Effect,
Quarterly Journal of Economics
133
(
2018
),
1283
1330
.
Soroka
,
Stuart
, and
Stephen
McAdams
, “
News, Politics, and Negativity
,”
Political Communication
,
32
(
2015
),
1
22
.
University of Michigan Survey Research Center, Institute for Social Research
, ”
Survey of Consumers
” (
n.d.
), https://data.sca.isr.umich.edu/fetchdoc.php?docid=24776.

Author notes

This research received funding from the Haverford College faculty research fund.

A supplemental appendix is available online at http://www.mitpressjournals.org/doi/suppl/10.1162/rest_a_00931.

Supplementary data