Opening Remarks at the Economic Press Briefing on the Survey of Consumer Expectations
November 18, 2016
William C. Dudley, President and Chief Executive Officer
Remarks at the Economic Press Briefing on the Survey of Consumer Expectations, Federal Reserve Bank of New York, New York City As prepared for delivery

Good morning, and welcome to the New York Fed’s Economic Press Briefing.  I am pleased to have the opportunity to speak with you today about our Survey of Consumer Expectations.  This is an important source of information that we use in our analysis of current and future economic conditions to help inform our monetary policy decisions.  As always, what I have to say reflects my own views and not necessarily those of the Federal Open Market Committee or the Federal Reserve System.1

In pursuing its monetary policy objectives, the Federal Reserve relies on a large number of data sources that provide information about recent developments and the current state of the economy.  Most of these data sources are backward looking.  That is, they provide data about past decisions and outcomes.  In contrast, we know relatively little about the forward-looking subjective expectations and intentions of economic agents, which are a key determinant of current and future outcomes.  This information, though, is essential for formulating policy and evaluating its effectiveness.  The macroeconomic outcomes we observe reflect preferences and constraints, as well as beliefs of individual consumers and firms.  Therefore, a fuller understanding of such outcomes requires reliable data on the beliefs and intentions held by economic agents and how they evolve over time.  In particular, consumer expectations influence a variety of considerations, including spending and saving behavior, wage dynamics, job search, housing demand and the demand for credit.  Given their importance in determining U.S. economic activity, consumer expectations are of particular value in assessing the economic outlook and the related risks.

Over the last decade, we have devoted significant resources to improving the measurement of the economic expectations of households.  This work culminated in the creation of the Survey of Consumer Expectations, or SCE, in 2013.  The expectations data collected in our survey give us direct insight into households’ decision-making and consumers’ financial and economic outlooks.  While we monitor expectations on a number of topics, of particular importance are households’ medium-term inflation expectations.  Having reliable measures of inflation expectations is crucial to assess the effectiveness of monetary policy and central bank communications.  Such an assessment focuses on analyzing the extent to which longer-run inflation expectations remain well-anchored, and on learning more about how agents update and revise their expectations in response to new information. The SCE allows us to carry out this sort of analysis.

Besides their obvious policy relevance, the SCE data are also highly valuable for research.  Our economists have shown that the SCE’s expectations data are informative and predictive of respondents’ current and future behavior.  They have also documented the substantial heterogeneity in the formation and updating of expectations across demographic groups.  To facilitate further research, we make the survey data available in the Center for Microeconomic Data section of our website.  Each month we also illustrate the trends in our expectations series through easy-to-use, interactive charts.  Further, a number of our Liberty Street Economics blog posts describe the data and the analysis we have conducted.  We do this because we want researchers, policymakers, commentators and the general public to understand the data and to follow our updates.

Let me now describe our Survey of Consumer Expectations in more detail.

The Survey of Consumer Expectations

Following extensive design and testing, we launched the Survey of Consumer Expectations in June 2013.  The survey has a number of innovative and unique features, three of which I would like to highlight today.  First and foremost, as I mentioned earlier, it collects households’ subjective expectations and intentions on a wide range of economic outcomes and behaviors.  In addition to short- and medium-term inflation expectations, it collects data on households’ expectations regarding the housing, labor and credit markets, as well as households’ financial situations and a host of other macroeconomic variables of interest.  For example, the survey elicits beliefs about future wage growth, house price appreciation, job search and labor market transitions, households’ spending and income growth, and credit applications.  By collecting data on a broad range of expectations, it is possible to analyze how they interact with each other and jointly influence households’ decisions.  We also study how expectations vary with demographics and other household characteristics.

The broad scope of the SCE allows us to study many questions relevant to economic policy—such as understanding housing market decisions, disentangling credit demand from supply, analyzing job search behavior and measuring wage growth as people transition between jobs.  In addition, the survey has been used to help answer other policy-relevant questions as they have come up in our discussions.  Two examples are households’ spending response to the decrease in gas prices in early 2014, and households’ early experiences and perceptions of the Affordable Care Act following its enactment.  Some of this analysis relies on the ability of our survey to shed light on whether consumers expect a specific change in economic conditions to be temporary or permanent.

A second important feature of the survey is that—in addition to simple point forecasts—for some variables, such as inflation, we get respondents’ entire forecast distributions by asking them to assign probabilities to various ranges of future outcomes.  These so-called density forecasts in turn allow us to measure households’ uncertainty about future events by looking at the spread in their subjective distributions.  Density forecasts also allow us to monitor the perceived likelihood that respondents assign to specific events, such as deflation or a large increase in earnings.

A third important characteristic of the SCE is that it is a rotating panel.  Each month about 1,300 household heads participate in our survey.  This is a nationally representative sample. Respondents remain on the panel for up to 12 months, with a constant fraction entering and rotating out each month.  This allows us to monitor the expectations of the same respondents over time, which provides important benefits.  First, it gives us greater confidence that month-to-month changes in our headline numbers reflect true changes in beliefs, and not just changes in the sample composition.  Second, observing repeated responses by the same individuals allows us to examine how expectations are formed and revised over time.  Third, by following the same people over time, we are able to study how expectations are related to contemporaneous and future economic decisions and outcomes for these respondents.

Today the SCE team will provide further background about the survey, discuss newly released findings for October and recent trends in some of its main indicators.  The team will then present findings from two special components of the SCE: one that covers households’ credit experiences and expectations, and one that deals with labor market behavior and expectations.  Before I pass this on to my colleagues, I would like to comment briefly on these two special surveys.

New findings on Credit Access

The SCE Credit Access Survey supplements the information collected in the Federal Reserve Board’s Senior Loan Officer Opinion Survey on Bank Lending Practices, also known as SLOOS.  The SLOOS is a primary source of information on credit standards and credit demand as perceived and reported by banks.  In contrast, the SCE Credit Access Survey, which has been conducted at a four-month frequency since October 2013, measures experiences in credit markets as perceived by households.  The combination of the two data sources provides a more complete picture of household credit access and demand.  Today, we are releasing the October 2016 findings as well as introducing new data from the survey.

A unique feature of the SCE is its measure of latent credit demand—that is, demand that does not lead to an application for credit.  We measure this latent credit demand as the proportion of households who report that, while needing credit, they did not apply because they expected their application to be rejected.  In addition to current credit experiences, we also elicit households’ expectations of their future credit applications and the likelihood of these applications being approved.  As we will show later, the survey provides useful insights into the extent and variation in met and unmet credit needs across households.  Moreover, our analysis indicates that reported expectations are indeed predictive of credit outcomes reported four months later.

New Findings on Job Search

While the SCE Credit Access Survey allows us to track credit market experiences and expectations, the SCE Labor Market Survey gathers information on the job search behavior and expectations of employed and unemployed respondents.  Aggregate labor market statistics on earnings, employment and labor market transitions are readily available, yet relatively little is known about the expectations and search behaviors of labor market participants and how these change over time.  Our SCE Labor Market Survey collects rich information on respondents’ actual and expected wage offers, as well as their reservation wage, which is the lowest wage or salary they say they would accept for a new job.  We also elicit expectations about future labor market transitions, such as the likelihood of staying with the same employer, switching employers, finding a job or being laid off.

Over time, changes in responses to these questions can shed light on how workers’ expectations evolve, and how these changes relate to observed trends in aggregate earnings and employment dynamics. Our preliminary analyses of the data suggest that these expectations measures are predictive of respondents’ subsequent labor market outcomes. In coming years, as we generate a longer time series for the survey, we will be able to explore whether the expectations we collect can be used to construct leading indicators of the job market.

Now, I'd like to turn things over to our economists, beginning with Olivier Armantier, who will provide further background and specifics about our survey.

1 Olivier Armantier, Giorgio Topa, Joseph Tracy, Wilbert van der Klaauw and Basit Zafar assisted in preparing these remarks.