Finance and Economics Discussion Series Divisions of Research & Statistics and Monetary Aﬀairs
Federal Reserve Board, Washington, D.C.
The Hidden Heterogeneity of Inﬂation Expectations and its
Lena Dr¨ager, Michael J. Lamla, and Damjan Pfajfar
Please cite this paper as:
Dr¨ager, Lena, Michael J. Lamla, and Damjan Pfajfar (2020). "The Hidden Hetero-geneity of Inﬂation Expectations and its Implications," Finance and Economics Discus-sion Series 2020-054. Washington: Board of Governors of the Federal Reserve System,https://doi.org/10.17016/FEDS.2020.054.
NOTE: Staﬀ working papers in the Finance and Economics Discussion Series (FEDS) are preliminary materials circulated to stimulate discussion and critical comment. The analysis and conclusions set forth are those of the authors and do not indicate concurrence by other members of the research staﬀ or the Board of Governors. References in publications to the Finance and Economics Discussion Series (other than acknowledgement) should be cleared with the author(s) to protect the tentative character of these papers.
The Hidden Heterogeneity of Inﬂation Expectations and its
Lena Dra¨ger‡Michael J. Lamla† Damjan Pfajfar§
July 2, 2020
Using a new consumer survey dataset, we document a new dimension of heterogeneity in inﬂa-tion expectations that has implications for consumption and saving decisions as well as monetary policy transmission. We show that German households with the same inﬂation expectations dif-ferently assess whether the level of expected inﬂation and of nominal interest rates is appropriate or too high/too low. The 'hidden heterogeneity' in expectations stemming from these opinions is related to demographic characteristics and aﬀects current and planned spending in addition to the Euler equation eﬀect of the perceived real interest rate. Furthermore, these diﬀerences in opinions aﬀect German households diﬀerently depending on whether they are renters or home-owners.
Keywords: Macroeconomic expectations, monetary policy perceptions, survey microdata.
JEL classiﬁcation: E31, E52, E58, D84.
∗We would like to thank participants at the American Economic Association Meeting, the TU Dresden and Joint Deutsche Bundesbank and Bank de France Conference on Household Expectations for their comments. The views expressed in this paper are those of the authors and do not necessarily reﬂect those of the Federal Reserve Board.
‡Leibniz University Hannover. Email:email@example.com.
† Leuphana University of Lu¨neburg and ETH Zurich, KOF Swiss Economic Institute. Email:firstname.lastname@example.org.
§Board of Governors of the Federal Reserve System. Email:email@example.com.
In recent years, a large literature has developed that focuses on the formation of consumers' macroe-conomic expectations. Phenomena such as the tendency of consumers to overestimate actual inﬂa-tion or patterns in forecast accuracy across socio-demographic groups have been widely observed across diﬀerent surveys and countries (Jonung,1981;Bryan and Venkatu,2001;Coibion and Gorod-nichenko,2015). Several papers have shown that inﬂation expectations are formed heterogeneously and document that these heterogeneous expectations have implications for consumption and saving decisions (Bachmann et al.,2015;Duca et al.,2018;Dra¨ger and Nghiem,2020).1
In this paper, we present new evidence regarding the relevance of opinions on expected future economic developments. Speciﬁcally, we show that even households with the same inﬂation expec-tations can have very diﬀerent opinions about the appropriate level of inﬂation and interest rates, and thereby about the right stance of monetary policy. This heterogeneity has implications for the transmission channel of monetary policy. While in the economics profession opinions have so far been somewhat neglected as a source of heterogeneity, in the social psychology literature - speciﬁcly in attribution theory - it has been long established how people form opinions and how they justify them.Jones and Nisbett(1972) andTversky and Kahneman(1973) report ﬁndings that people tend to view their own behavior as reﬂecting the changing demands of their environment.2This provides some understanding how opinions could also matter for economic decisions. It would not be surprising to see that consumers in our survey 'act' on their opinions and actually make spending decisions based on their opinions.
We ﬁrst detail the 'hidden heterogeneity' in expectations using the Bundesbank Online Pilot Survey on Consumer Expectations. Overall, the majority of consumers believe that expected inﬂa-tion is too high and expected interest rates are too low. Remarkably, even consumers with inﬂation expectations that are well within the ECB's target inﬂation rate of close to, but under 2%, diﬀer substantially in their opinions of whether this is an appropriate level of inﬂation. Speciﬁcally, for consumers with inﬂation expectations between 1.5% and 2%, about 49% believe that expected in-ﬂation is appropriate, 46% think it should be lower and 5% think it should be higher. Strikingly, even among consumers who expect deﬂation in the next year, about 30% would still prefer lower inﬂation. Generally, a large share of German households believes that inﬂation is too high. We observe similar heterogeneity also for consumers' opinions regarding the stance of monetary policy, i.e., future interest rates.
We further document that these diﬀerences in attitudes result in some heterogeneity in house-holds' consumption and savings proﬁles, even for consumers who share similar inﬂation expectations.
This implies an additional channel of monetary policy transmission via the attitudes of households in addition to the eﬀect of inﬂation expectations on spending via the real interest rate. We ﬁnd that when households perceive higher real rates, they postpone part of their spending on durable goods. This eﬀect is in line with the theory, namely the intertemporal substitution eﬀect in the
1For a recent survey on the formation of inﬂation expectations and their eﬀect on economic decisions seeCoibion et al.(2020).
2They also report that people think that the behavior of others is trait dominated.
consumption Euler equation. Interestingly, the negative eﬀect of perceived rates is only signiﬁcant for consumers with inﬂation expectations in line with the ECB target, i.e., between 1.5% and 2%. Furthermore, those who believe interest rates should be lower in the future, are de facto acting as if (nominal) interest rates - and thus real interest rates - are already lower, as they have sig-niﬁcantly higher durable goods spending and also a higher negative elasticity with respect to real rates. Interestingly, the eﬀect of attitudes is not only relevant for current spending decisions, but also aﬀects future planned spending on durable goods in a way which is consistent with theory. Distinguishing further between homeowners and renters, we ﬁnd opposing eﬀects of preferring lower inﬂation on current durable goods spending: Homeowners report lower current spending when they think inﬂation should be lower (in line with an Euler equation), while renters report higher durable goods spending.
Our paper relates to the literature explaining the heterogeneity of expectations across socio-demographic groups. Earlier contributions byJonung(1981),Bryan and Venkatu(2001) andPfajfar and Santoro(2009) demonstrate higher levels of both perceived and expected inﬂation for women, low education and low income groups, with a u-shaped eﬀect of age where young and old respon-dents have higher expectations than middle age respondents. This pattern is highly prevalent in many diﬀerent surveys across both diﬀerent countries and time spans. More recent approaches byD'Acunto et al.(2019) andD'Acunto et al.(2019) demonstrate that the gender diﬀerences in inﬂa-tion expectations can be traced to diﬀerences in daily grocery shopping experiences (as hypothesized inJonung,1981) and that they spill over into gender diﬀerences in expectations on other macroe-conomic variables. Moreover,Ehrmann et al.(2017) demonstrate that consumers' attitudes like optimism or pessimism regarding the economic outlook inﬂuence also the level of inﬂation expec-tations, whileD'Acunto et al.(2019) show that cognitive abilities play an important role. Finally, personal inﬂation experience can explain some of the diﬀerences in inﬂation expectations across age cohorts (Malmendier and Nagel,2016) and across diﬀerent political systems, e.g., the Western part of Germany and the former German Democratic Republic (GDR) in the East of Germany before 1989 (Goldfayn-Frank and Wohlfahrt,2019).
Our paper is related also to a growing literature evaluating the link between survey inﬂation expectations and household spending decisions. Assuming consumers are following an Euler equa-tion, one would expect a positive eﬀect from higher inﬂation expectations on current spending via its eﬀect on the real rate, which could become particularly important when nominal interest rates are at the zero lower bound. WhileBachmann et al.(2015) andBurke and Ozdagli(2013) ﬁnd little evidence of a signiﬁcant link between inﬂation expectations and consumers' reported readiness to spend (or actual spending) on durables in the US,Crump et al.(2015) report a positive relation be-tween consumption growth and inﬂation expectations of US consumers in the Survey of Consumer Expectations (SCE) conducted at the New York Fed. Other studies on European and Japanese households ﬁnd signiﬁcantly positive links between household inﬂation expectations and (intended or actual) spending on both durables and non-durables (Ichiue and Nishiguchi,2015;D'Acunto
et al.,2016;Duca et al.,2018;Vellekoop and Wiederholt,2018;Dra¨ger and Nghiem,2020).
The remainder of the paper is organized as follows: Section 2 explains the data we use, while Section 3 discusses our results. Section 4 concludes.
Our research question is evaluated using a new survey dataset coming from the Bundesbank Online Pilot Survey on Consumer Expectations, which was ﬁelded on a representative sample of German households in three waves from April 2019 to June 2019. Overall, the dataset includes 6653 ob-servations, with 2009 participants in the ﬁrst wave, 2052 in the second wave and 2592 in the third wave. In addition, the survey includes a panel component, as about 500 respondents participated in all three waves, 500 in wave 1 and 2, 500 in wave 2 and 3 and 500 in wave 1 and 3. For our analysis, we use mainly the ﬁrst and second wave of the dataset and thus have about 1000 participants with responses in both waves.
The Bundesbank Online Pilot Survey on Consumer Expectations core questionnaire asks about consumers' macroeconomic expectations, housing market expectations and housing choices, current and planned spending and saving choices, as well as a large range of socio-demographic character-istics. We add the following questions to the core questionnaire. First, after the question on point estimates for inﬂation 12 months ahead, we ask about opinions on the expected level of inﬂation (included in the ﬁrst wave):
1. Do you think the average level of inﬂation you expect for the next 12 months will be more or less appropriate, or do you think a higher or lower inﬂation rate would be better?
(a) Higher inﬂation than expected would be better (d inﬂ highbetter)
(b) Inﬂation will be more or less appropriate (d inﬂ reason)
(c) Lower inﬂation than expected would be better (d inﬂ lowbetter)
Similarly, we ask about opinions on the expected level of nominal interest rates after the question on point estimates for expected saving rates in the next 12 months (included in the second wave):
2. Do you think the average level of interest rates you expect for the next 12 months will be more or less appropriate, or do you think a higher or lower interest rate would be better?
(a) Higher interest rate than expected would be better (d int highbetter)
(b) The interest rate will be more or less appropriate (d int reason)
(c) Lower interest rate than expected would be better (d int lowbetter)
In our analysis, we further control for quantitative point forecasts for the next 12 months regard-ing consumer price inﬂation,πe, the average savings rate,iesavingsand the average mortgage rate,imeortgage. In order to avoid an eﬀect from extreme outliers, inﬂation expectations are truncated in the range between -5% and +25% and interest rate expectations are truncated to be below or equal to 25%.