AdobeStock_330211396.jpeg
VoxEU Column COVID-19 Macroeconomic policy

Policy communication during COVID

The COVID-19 pandemic has resulted in some of the largest monetary and fiscal policy responses around the world. This column uses a large-scale survey of US households during the pandemic to study how new information about the coronavirus and associated policy responses affect households’ expectations. It finds that such information treatments have little effect on both households’ economic beliefs and future spending plans. This result is a fundamental challenge to workhorse models used by macroeconomists in which the rapid and endogenous adjustment of household expectations is a key driver of macroeconomic outcomes.

The COVID-19 pandemic has resulted in some of the largest monetary and fiscal policy responses around the world. This column uses a large-scale survey of US households during the pandemic to study how new information about the coronavirus and associated policy responses affect households’ expectations. It finds that such information treatments have little effect on both households’ economic beliefs and future spending plans. This result is a fundamental challenge to workhorse models used by macroeconomists in which the rapid and endogenous adjustment of household expectations is a key driver of macroeconomic outcomes.

Monetary and fiscal policies affect the economy (Romer and Romer 2004, 2010) but how they operate remains a point of contention. A common thread across many macroeconomic models is the role of expectations: policies have powerful effects in modern mainstream models in large part because firms and households incorporate these announcements into their decision plans. In real business cycle models, for example, an announcement of higher government spending should make households feel poorer (since they will have to pay for this spending via higher taxes now or in the future) which induces them to work more. Forward guidance on the part of monetary policymakers is predicted to have large effects in New Keynesian models because the promise of future lower interest rates by the central bank should induce households to anticipate higher inflation in the future which in turn should lead them to consume more today before those price increases materialise.

How powerful are these mechanisms in practice? Recent research should give one pause: there is a growing body of evidence documenting that, in advanced economies, inattention to macroeconomic policy and the broader economic environment is pervasive among households and firms. Announcements by monetary and fiscal policymakers are rarely found to have large effects on the expectations of economic agents other than those participating directly in financial markets, suggesting that these expectational forces may in fact be quite weak. Still, one might expect a strengthening of these forces in a crisis, as a worried population turns its attention to its leaders for guidance and support.

Using a large-scale survey of US households during the COVID-19 pandemic, we study how new information about policy responses affects the expectations and decisions of respondents. Specifically, we provide random subsets of participants with different combinations of information about the severity of the pandemic, recent actions by the Federal Reserve, stimulus measures implemented by Congress, as well as recommendations from the US Center for Disease Control (CDC). We then characterise how their economic expectations and spending plans respond to these information treatments. This allows us to assess to what extent these policy announcements alter the beliefs and plans of economic agents.

By and large, we find very little effect of these information treatments on the economic expectations of agents for income, mortgage rates, inflation or the unemployment rate nor do we find an effect on their planned decisions, contrary to the powerful effects they have in standard macroeconomic models. Why might agents’ economic beliefs not respond to this information? One possible explanation is that they were already aware of the information provided in the treatments. While we do not have the prior beliefs of agents for all information treatments, those for which we do suggest that this is not a likely explanation. For example, households’ prior beliefs about the transmission rate of COVID-19 or its recovery rate were wildly misinformed prior to the information treatments (see Figure 1). Furthermore, previous work has documented how uninformed households tend to be about most monetary and fiscal policies and how even large policy announcements do not make their way into households’ aggregate expectations, even in the midst of a crisis (e.g. Coibion et al. 2020). Furthermore, Binder (2020) documents that even after the historic policy actions of the Federal Reserve in response to the COVID-19 crisis, only a third of US households had heard about these policy actions. A second possible explanation is if households are sceptical of the information that we provide. Again, we view this as very unlikely because other information treatments in identical settings have previously been found to lead to dramatic revisions in households’ views about the economy (e.g. Coibion et al. 2019). A third possible explanation rests on the idea that, because of cognitive constraints, many households might not directly understand the implications of complex policies for their optimal savings and consumption decisions (e.g. D’Acunto et al. 2020a,2020b). The fourth, and in our view most likely, explanation is that households do not believe that the policy responses described in the treatments are effective: i.e. the multipliers they associate with the described policy responses are close to zero. Note that zero multipliers may be observed because so-called information effects (i.e. policy actions reveal a bad state of the economy) offset any positive effects of a policy action.

Figure 1 Distribution of beliefs about how contagious and fatal the COVID-19 virus is

Panel A Infection rate

Panel B Recovery rate

Notes: Panel A: Infection rate is measured as the response to the following question, “Think of a person who has the coronavirus. How many non-infected people do you think will catch the virus from this person?”. The response is winsorised at 100. Panel B: the recovery rate is measured as the response to the following question, “If a person contracts coronavirus, what do you think is the probability that this person recovers from the virus?  Please enter a number between 0 (Do not recover) and 100 (Recover for sure)”. In each panel, the red, vertical line shows the estimates provided by the World Health Organization.

Previous work has documented extensively how inattentive households (and firms) tend to be to macroeconomic conditions (Bachmann et al. 2015, Coibion et al. 2018, Coibion et al. 2019, D’Acunto et al. 2019). We find the same qualitative patterns hold during the COVID crisis but also document that this lack of understanding extends to information about the coronavirus. For example, when we ask households what they think the recovery rate is once infected with COVID-19, they report an average answer of 73%, far lower than the 97% reported by the World Health Organization (WHO). Similarly, when we ask them how many people tend to be infected by someone carrying COVID-19, their average answer is 21, far higher than the actual rate of around two estimated by the WHO. This suggests that information treatments that provide factual information about transmission and recovery rates could potentially have important effects on households’ expectations about the economy.

Despite this, we find very small effects of providing information about the deadliness and ease of spread of the disease on households’ expectations. When respondents are treated with information that, on average, the disease is harder to spread and less deadly than they had original thought, their views about future inflation, mortgage rates, and unemployment are effectively unchanged. They reduce their reported expected future income on average but the change is economically insignificant. Their perceptions about whether now is a good or bad time to buy durables are also effectively unchanged. The one exception is for unemployed workers who are asked about the likelihood of finding a job: those who are treated with information about the disease raise their likelihood of finding a job by about 20 percentage points. These results suggest that the large changes in expectations during the COVID-19 pandemic for income, the stock market, or mortgage rates are less likely driven by direct concerns about the virus but more likely a response to the lockdowns imposed by local authorities in line with findings in Coibion et al. (2020).  

Information treatments about fiscal, monetary or health policies similarly do very little to the expectations of households, both about the aggregate economy or about their own income. And when they do, those effects are not necessarily positive. For example, among the unemployed who become more optimistic about their future job prospects when they are told that COVID-19 spreads less easily and is less deadly than they thought, providing additional information about the responses of policymakers fully offsets the effect of the information about the disease. This is consistent with the presence of an information effect to policies: finding out that fiscal, monetary or health policymakers are implementing large policy changes makes the unemployed less optimistic about their job prospects, but only when done in conjunction with information about the disease. Information treatments that are only about policy changes have effectively no effect on most agents’ macroeconomic or individual expectations. These results are consistent with recent findings documenting an information effects of monetary policy which suggest that large policy moves might reveal information about the state of the economy which is called Delphic in the context of forward guidance (see, for example, Campbell et al. 2012)

By studying the effect of policy actions on households’ macroeconomic expectations through RCTs, our paper is closest to Andre et al. (2019) who present respondents with hypothetical exogenous shocks to either fiscal or monetary policy whereas we present households with information about clearly endogenous policy responses. Our results therefore speak directly to the effects of systematic policy changes. Our findings suggest that these systematic policy responses have little effect on households’ expectations, either because they believe they are ineffective or because policy responses induce an information effect (in which households interpret the sheer fact of a policy response as indicative of a weaker economy) that effectively offsets the effect of the policy change.

Understanding the way in which policy actions affect the economy has long been a challenge for macroeconomists. Standard models imply that households’ expectations play a large role in driving these effects, as households incorporate the announcements into their expectations and their decisions. Our results challenge this key mechanism: we find little evidence that even large policy decisions have much of an effect on households’ economic expectations or their planned actions. We view this as a fundamental challenge to workhorse models used by macroeconomists in which the rapid and endogenous adjustment of household expectations is a key driver of macroeconomic outcomes.

References

Andre, P, C Pizzinelli, C Roth and J Wohlfart (2019), “Subjective Models of the Macroeconomy: Evidence from Experts and a Representative Sample”, Manuscript.

Bachmann, R, T Berg and E Sims (2015), “Inflation Expectations and Readiness to Spend: Cross-Sectional Evidence”, American Economic Journal: Macroeconomics 7(1): 1-35.

Binder, C (2020), “Coronavirus Fears and Macroeconomic Expectations”, Review of Economics and Statistics, forthcoming.

Campbell, J, C Evans, J Fisher and A Justiniano (2012), “Macroeconomic Effects of Federal Reserve Forward Guidance”, Brookings Papers on Economic Activity 44(1): 1-80.

Coibion, O, Y Gorodnichenko and S Kumar (2018), “How Do Firms Form Their Expectations? New Survey Evidence”, American Economic Review 108(9): 2671-2713.

Coibion, O, Y Gorodnichenko and M Weber (2019), “Monetary Policy Communications and their Effects on Household Inflation Expectations”, NBER Working Paper 25482.

Coibion, O, D Georgarakos, Y Gorodnichenko and M Weber (2020), “Forward Guidance and Household Expectations”, NBER Working Paper 26778.

D’Acunto, F, D Hoang and M Weber (2020a), “Managing Households' Expectations with Unconventional Policies”, Manuscript.

D’Acunto, F, D Hoang, M Paloviita and M Weber (2020b), “Human Frictions in the Transmission of Economic Policy”, Manuscript.

D’Acunto, F, D Hoang, M Paloviita and M Weber (2019), “Cognitive Abilities and Inflation Expectations”, AEA Papers and Proceedings 109: 562-66.

Romer, C and D Romer (2004), “A new measure of monetary shocks: Derivation and implications”, American Economic Review 94(4): 1055-1084.

Romer, C and D Romer (2010), “The macroeconomic effects of tax changes: estimates based on a new measure of fiscal shocks”, American Economic Review 100(3): 763-801.

420 Reads