Do consumers actually monitor the inflation rate? Evidence from New Zealand*

Abstract

In this note, we study whether consumers actually monitor the inflation rate, an assumption that is often made in studies on inflation perceptions and expectations as well as policy analyses. We analyse this question using unique representative survey data on New Zealand collected in 2016. In this case of an inflation targeting country and an environment of low inflation rate, we find that only about one third of the population says that it monitors the inflation rate. These are people characterised by a significantly higher degree of objective and subjective economic knowledge as well as interest in monetary issues.

KEYWORDS:

• Inflation monitoring
• New Zealand
• monetary policy
• household survey

JEL:

• E52
• E58
• Z1

1. Introduction

In many industrialised countries, the inflation rate has reached unprecedentedly low levels. Arguably, this implies that consumers have relatively little incentive to keep an eye on the development of prices. In the case of New Zealand, for which we ran a representative household survey in 2016, the inflation rates based on the consumer price index in 2015, 2016, and 2017 were roughly 0.3, 0.6, and 2 per cent, respectively. But in many macroeconomic household studies, especially those analysing inflation expectations, it is assumed that consumers pay close attention to the inflation rate and that their answers are based on constant monitoring. In fact, in the literature on forward guidance, it is often assumed that central banks can steer inflation expectations just by appropriately-worded communication with the public (e.g. Woodford, 2005). Empirical studies seem to support that conclusion. For instance, using data from the Michigan Survey, Dräger and Lamla (2012) report that while the general inflation outlook is rarely adjusted, individuals fine-tune their short-run expectations quite regularly. Binder (2017) finds that survey respondents adjust their inflation expectations roughly four times in six months. Dräger, Lamla, and Pfajfar (2016) even claim that a substantial share of households has expectations consistent not only with the Fisher equation, but also the Taylor rule as well as the Phillips curve. In the case of New Zealand and using a combination of different inflation expectation surveys, Lewis and McDermott (2016) report that numerical changes in the inflation target result in an immediate change in long-term inflation expectations.

All of these studies take as the starting point that households behave in line with standard economic theory and vigilantly monitor key economic variables such as the rate of inflation. But as the tests have low power if the data are not very informative, it is hard to reject specific hypotheses. Testing is more powerful when two or more competing hypotheses are tested against each other instead of only testing each of the hypotheses against the data (Hendry, 1993). Moreover, non-rejection of a hypothesis does not imply its acceptance. Finally, most of these studies have been undertaken on US data and it is not clear how representative these are for the rest of the world.

In this note, we would like to enter the discussion and address the question of whether in a low inflation environment laypersons really pay attention to the inflation rate. Given the complexity of the topic, obtaining knowledge about the inflation rate requires some effort. Without some alertness to and interest in the development of inflation, it seems unlikely that economic agents will be able to accurately report its rate. Thus, our study is a critical assessment of the typical ‘vigilance’ assumption made in household inflation studies by economists.

We study whether laypersons actually monitor the inflation rate in a low inflation environment with the help of data from a representative population survey from New Zealand. A summary of the period of inflation targeting within which our survey took place is provided by Buckle (2019). Preceding the introduction of inflation targeting, Silverstone (2014) discusses a small survey conducted in 1987 amongst members of the Reserve Bank of New Zealand and finds that, initially, there appeared to be a preference for a more flexible approach to inflation targeting. Using survey data on New Zealand firms, Kumar, Afrouzi, Coibion, and Gorodnichenko (2015) discover that inflation targeting does not appear to anchor expected inflation rates and Parker (2017) reports that the median number of price reviews is twice per year. Thus, even at the firm level, the notion of constant inflation vigilance may be more myth than fact. In light of low inflation rates in New Zealand, Karagedikli and McDermott (2018) identify a change in the behaviour of inflation expectations and argue that these have become more backward-looking.

The household survey data we are using in this note were collected on our behalf by Research New Zealand in May 2016 (see Hayo & Neumeier, 2016 for a description of the survey). The survey has been used to study other research questions. In particular, Hayo and Neumeier (2018) discuss households’ inflation perceptions and expectations, Hayo and Neumeier (2020) analyse trust in the Reserve Bank of New Zealand, and Hayo and Neumeier (2020) investigate central bank independence and the Policy Targets Agreement.

Methodologically, we combine descriptive statistics with an explorative multivariate regression approach. Given the nature of our data, we cannot properly test for causality. Hence, our results should be interpreted as multivariate correlations. Conceptually, our analysis takes place within an extension of a framework put forward by Ranyard, Del Missier, Bonini, Duxbury, and Summers (2008). However, to the best of our knowledge (see Hayo & Neumeier, 2018 for an extensive summary of the literature), our research question has not been studied before, definitely not in a similar framework and with a focus on New Zealand.

We believe that New Zealand is an excellent case to study our question of interest. As the first country to adopt an explicit inflation target, New Zealanders should be used to inflation as the prime monetary policy target.¹ Thus, we would expect them to consciously monitor the inflation rate. Moreover, by using this case, we have a combination of inflation targeting regime and low inflation environment, which should be conducive for our empirical investigation.

In Section 2, we are interested in finding out whether New Zealanders make a conscious effort to learn the inflation rate. Section 3 concludes.

2. Do people actively monitor the inflation rate?

Studies of household inflation expectations are commonly based on the premise that persons actually monitor the inflation rate (e.g. Thomas, 1999). However, as discussed by Sims (2003) and Caplin and Dean (2015), in a situation of imperfect information it can be rational for economic actors not to engage in extensive information gathering as information search and learning is costly. In our context, we would argue that monitoring inflation takes time and (mental) effort. Thus, during a period of low inflation, constant vigilance may simply be too costly for households.

To see whether households keep updated about the inflation rate, we asked our respondents about whether they actually monitor the rate of inflation. Table 1 shows that only 35 per cent of the population keeps an eye on the inflation rate, which does not bode well for assumptions of rational expectation formation based on the idea that people collect all available, or at least all easily available, information before making decisions.

Table 1. Do you monitor the rate of inflation? (absolute and relative number of respondents).

Yes	No	Don’t know
352 (35%)	605 (61%)	43 (4%)

To learn more about the characteristics of respondents who either do or do not monitor the inflation rate, we compute conditional associations in a multivariate setting. As a dependent variable in our logit regression, we use a dummy variable equal to 1 if a person monitors the inflation rate and 0 otherwise. As building blocks for a general model, we include explanatory variables covering various dimensions, namely:

1. ‘Economic Situation’ (measured by: Income, Net personal wealth, Saver, Debtor, Satisfaction with financial situation, Self-employed full time, Self-employed part time, Employed full time, Employed part time, Homemaker, Student, Retired, Unemployed, Beneficiary);

2. ‘Economic Knowledge’ (measured by: Macroeconomic knowledge, Feels informed about RBNZ, Feels informed about inflation, Feels informed about OCR, Heard of PTA);

3. ‘Information Search’ (measured by: Desire to be informed about RBNZ, Information through newspaper, Information through radio, Information through TV, Information through Internet, Information through friends, Information through colleagues, Information through own bank, Information through financial sector, Does not keep up with RBNZ);

4. ‘Attitudes and Values’ (measured by: Institutional trust, General trust, Politicians act in public’s best interest, Politicians long-term oriented, Politicians fiscally competent, Confidence in politicians, Egalitarian attitude, National Party, Labour Party, Green Party, New Zealand First).

5. We also include socio-demographic and psychological indicators, which control for a number of other influences (Female, Age, Children, Married, Secondary school qualification, Polytechnic qualification or trade certificate, Bachelor’s degree or higher, Town, Rural, North Island, Auckland, NZ European, Maori, Asian, Time spent on survey, Risk propensity, Future-oriented time preference, Short-run impatience).

Descriptive information about these variables can be found in the Appendix. Starting with a model containing these 59 potentially relevant variables (see Equation 1(1) $\text{Inflation Monitoring Dummy}=\alpha +\sum _1^{59}{\beta }_i{X}_i+\epsilon ,$(1) ), we use general-to-specific modelling to simplify the model. (1) $\text{Inflation Monitoring Dummy}=\alpha +\sum _1^{59}{\beta }_i{X}_i+\epsilon ,$(1) where $X_i$ reflects the 59 explanatory variables, ${\beta }_i$ contains their corresponding parameters, α is a constant, ϵ a random error, and Inflation Monitoring Dummy is a dummy variable that takes the value 1 when respondents state that they monitor the inflation rate.

Using logit estimation, Table 2 presents the estimation results for the reduced model.² Seven variables survive the testing-down process and are significant at either the 1 or 5 per cent level of significance, the majority of which relate to subjective or objective economic knowledge. Note that a model selection process conditional on previous tests leads to an underestimation of the actual level of significance α* (Lovell, 1983). However, general-to-specific modelling, which is based on a sequence of tests within nested models, allows precisely computing α* (see Maddala, 1988, p. 425). In our case, two rounds of testing were needed to arrive at the final model, which implies that α* is either 2 per cent or 10 per cent, depending on whether we apply a nominal significance level of 1 per cent or 5 per cent.³

Table 2. Explaining ‘Monitoring the inflation rate’.

	Coefficient	SE	Average marginal effects (AME)	AME by standard deviation
Feels informed about inflation	0.84***	0.11	0.13	0.15
Feels informed about RBNZ	0.41***	0.14	0.06	0.06
Desire to be informed about RBNZ	0.31***	0.11	0.05	0.05
Does not keep up with RBNZ	−1.18**	0.50	−0.18	n.a.
Macroeconomic knowledge	0.32***	0.07	0.05	0.09
Egalitarian attitude	−0.17**	0.08	−0.03	−0.04
Risk propensity	0.40***	0.15	0.06	0.04
Constant	−5.95***	0.60
Pseudo-R^2 = 0.32; Chi^2(7) = 219***; log pseudolikelihood = −366; testing-down restriction:
F(52, 4.5E+07) = 1.14; number of observations: 808.

Notes: Estimated using a logit model. White (1980) robust SEs are used. For dummy variable reference values, see list of variables in the Appendix. *, **, and *** indicate nominal significance at a 10%, 5%, and 1% level, respectively.

Regarding subjective knowledge, we discover that if people feel informed about inflation or RBNZ, then it is more likely that they monitor the inflation rate. Of course, causality may run the other way, but this is not the issue here. To get an idea about the magnitude of the estimated relationships, we compute average marginal effects and, for those variables that are not dummies, multiply these by the variables’ standard deviation. The result can be interpreted as the impact of a one standard deviation change of an explanatory variable on the likelihood that the dependent variable is equal to unity.

For subjective knowledge, we find a notable impact of 15 percentage points (pp) on the likelihood of monitoring the inflation rate. At 6 pp, the positive association between subjective knowledge about RBNZ and inflation monitoring is less than half as large. Respondents who desire to obtain information about the inflation rate are more likely to monitor it, whereas those who do not bother keeping up with the RBNZ are also less inclined to follow the development of inflation. The impact of a standard deviation change is about 5 pp in the case of the variable measuring information desire. Since the ‘Does not keep up with RBNZ’ variable is a dummy, we just look at a change from 0 to 1 and find that the likelihood of monitoring the inflation rate decreases by 18 pp.

The estimates for the reduced model shown in Table 2 are based on the same number of observations used in estimating the general model. Due to including fewer variables in the reduced model, we now have additional observations available for estimation. After increasing the sample size to 893, that is, extending it by more than 10 per cent, and re-estimating the reduced model, our results remain robust.⁴

The likelihood of ‘Monitoring the inflation rate’ is not only influenced by the various dimensions of subjective knowledge. A standard deviation change in our indicator for macroeconomic knowledge raises the likelihood of monitoring inflation by almost 10 pp. In addition, we find that respondents who are more concerned about equality are less interested in monitoring the inflation rate, whereas those who are less risk averse are significantly more interested. In both cases, the absolute effect of a standard deviation change is relatively small (4 pp).

Overall, the results are consistent with the notion that monitoring is a precondition for acquiring information about inflation. On average, respondents either make a conscious effort to collect information about inflation or they have sketchy objective and subjective knowledge. Put differently, there is a group of citizens who consciously and actively think about inflation and monetary policy and this group, at least to some extent, fulfils the rationality assumption often made by macroeconomists. However, in our sample, this group makes up only slightly more than 30 per cent of the population.

Many of our variables that one might expect to influence the likelihood of watching the inflation rate are not significant. For instance, after controlling for the remaining variables in the reduced model given in Table 2, it does not matter whether the respondent is a debtor/saver or rich/poor. When regressing these variables individually on ‘Monitoring the inflation rate’, we find that savers and the rich are significantly more likely to monitor the inflation rate. This suggests that models that include these economic variables, but do not control for the other variables discussed above, likely suffer from biased estimates.

3. Conclusion

Using representative population survey data from New Zealand collected in 2016, we study whether laypersons actively monitor the inflation rate in a low inflation environment. We find that only 35 per cent of the population explicitly follows the inflation rate. Those who do possess a high level of subjective and objective macroeconomic knowledge as well as an interest in the RBNZ. Thus, the results are consistent with the notion that actively monitoring inflation is a precondition for having a relatively precise idea of the inflation rate and stand in contrast to the notion that people unconsciously acquire this information.

The assumption typically made in studies on inflation perceptions and expectations, namely that consumers pay close attention to the inflation rate, receives little support in our case study in a situation of low inflation. It also suggests that central banks are unlikely going to be in a position to steer inflation expectations of laypersons just by communicating with the public. This would explain why Lewis and McDermott’s (2016) finding that numerical changes in the inflation target result in an immediate change in long-term inflation expectations was not significant. Moreover, in light of the evidence presented here, Karagedikli and McDermott’s (2018) conclusion of an increased importance of backward-looking expectation formation could be re-interpreted as reflecting little price monitoring by households during times of low inflation rates. In contrast, our finding is in line with results reported in Hayo and Neumeier (2020), which suggest that New Zealand’s population has little interest in and knowledge of monetary policy.

Finally, we would like to emphasise that it is not clear whether these are externally valid findings or restricted to New Zealand’s low inflation environment at the time of the survey. However, additional research on other countries and different time periods should help answer that question.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Note that since our survey was conducted, there was a change in New Zealand’s monetary policy regime (see Hayo & Neumeier, 2020).

2 The estimates for the general model are available on request.

3 The computations are: 5 per cent level: 1-(1-0.05)(1-0.05) = 0.1; 1 per cent level: 1-(1-0.01)(1-0.01) = 0.02.

4 Results available on request.

Appendices

Online Appendix

Variable Descriptions. See Hayo and Neumeier (2016) for more information about the survey and the questionnaire.

Explained Variables

Table

Variable	Coding and Comments	Mean	Std. dev.	Min	Max
Monitoring the inflation rate	Dummy. Coded 1 if respondents monitor the inflation rate.	0.35	0.48	0	1

Explanatory Variables

Table

Variable	Coding and Comments	Mean	Std. dev.	Min	Max
(i) ‘Economic Situation’
Income	Per capita household income in NZD1,000. We added 184 observations through 10 rounds of imputations using: Age, Age squared, education dummies, Saver, Future-oriented time preference, Self-employed full time, Employed full time, Employed part time, Retired, Student, Unemployed, Beneficiary. Descriptive statistics for imputation 10.	34.0	27.1	2.7	240
Net personal wealth	In NZD1,000. We added 224 observations through 10 rounds of imputations using: Age, Age squared, education dummies, Saver, Future-oriented time preference, Self-employed full time, Employed full time, Employed part time, Retired, Student, Unemployed, Beneficiary. Descriptive statistics for imputation 10.	35.2	88.0	−375	500
Saver	Dummy	0.63	0.48	0	1
Debtor	Dummy	0.30	0.46	0	1
Satisfaction with financial situation	Very dissatisfied (coded 1)	3.31	1.12	1	5
	Dissatisfied (coded 2)
	Neither satisfied nor dissatisfied (coded 3)
	Satisfied (coded 4)
	Very satisfied (coded 5)
	Don’t know (coded 3)
Employed full time	Dummy	0.38	0.49	0	1
Employed part time	Dummy	0.11	0.32	0	1
Homemaker	Dummy	0.06	0.24	0	1
Student	Dummy	0.08	0.27	0	1
Retired	Dummy	0.12	0.33	0	1
Unemployed	Dummy	0.05	0.21	0	1
Beneficiary	Dummy	0.04	0.20	0	1
(ii) ‘Economic Knowledge’
Macroeconomic knowledge	The sum of correct answers to questions on the bond rate, the goal set in the Fiscal Strategy Report, debt-to-GDP ratio, Official Cash Rate, main objective of RBNZ, independence of RBNZ with regard to interest rate setting, interest rate reaction to an expected increase in the inflation rate, inflation target as agreed in PTA.	2	1.75	0	8
Feels informed about RBNZ	Very poor (coded 1), Poor (coded 2), Neither poor nor good (coded 3), Good (coded 4), Very good (coded 5)	2.72	0.96	1	5
Feels informed about inflation	Very poor (coded 1), Poor (coded 2), Neither poor nor good (coded 3), Good (coded 4), Very good (coded 5)	3.42	1.17	1	5
Feels informed about OCR	Very poor (coded 1), Poor (coded 2), Neither poor nor good (coded 3), Good (coded 4), Very good (coded 5)	3.10	1.34	1	5
Heard of PTA	Dummy. Coded 1 if respondent has heard of the Policy Targets Agreement.	0.15	0.36	0	1
(iii) ‘Information Search’
Desire to be informed about RBNZ	Not at all important (coded 1), Unimportant (coded 2), Neither important nor unimportant (coded 3), Important (coded 4), Very important (coded 5), Don’t know (coded 3)	3.18	1.06	1	5
Information through newspaper	Dummy	0.11	0.31	0	1
Information through radio	Dummy	0.08	0.27	0	1
Information through TV	Dummy	0.18	0.39	0	1
Information through Internet	Dummy	0.22	0.42	0	1
Information through friends	Dummy	0.12	0.32	0	1
Information through colleagues	Dummy	0.07	0.26	0	1
Information through own bank	Dummy	0.06	0.24	0	1
Information through financial sector	Dummy	0.06	0.24	0	1
Does not keep up with RBNZ	Dummy	0.12	0.32	0	1
(iv) ‘Attitudes and Values’
Trust in RBNZ	5-point Likert scale ranging from (1) ‘No trust and confidence at all’ to (5) ‘Complete trust and confidence’; Don’t know (coded 3).	3.30	0.96	1	5
Institutional trust	Principal component based on trust in government, trust in parliament, trust in United Nations, and trust in International Monetary Fund.	−3e−09	1.55	−3.50	4.38
General trust	Dummy	0.34	0.47	0	1
Politicians act in public’s best interest	5-point Likert scale ranging from (1) ‘Most politicians in New Zealand serve the interests of particular groups’ to (5) ‘Most politicians in New Zealand act with the general public’s best interests in mind’.	3.02	0.93	1	5
Politicians long-term oriented	5-point Likert scale ranging from (1) ‘Most politicians are only concerned about the next election’ to (5) ‘Most politicians are concerned about New Zealand’s long-term well-being’.	2.38	1.15	1	5
Politicians fiscally competent	5-point Likert scale ranging from (1) ‘The Government wastes the revenue it collects in taxes’ to (5) ‘The Government conscientiously manages the revenue it collects in taxes’.	2.73	1.11	1	5
Confidence in politicians	5-point Likert scale ranging from (1) ‘I do not have confidence in New Zealand politicians’ to (5) ‘Overall, I have confidence in New Zealand politicians’.	2.59	1.12	1	5
Egalitarian attitude	5-point Likert scale ranging from (1) ‘To encourage individual effort, the difference between people’s incomes should be greater’ to (5) ‘People’s incomes should be more equal’.	3.32	1.20	1	5
National Party	Dummy	0.29	0.45	0	1
Labour Party	Dummy	0.23	0.42	0	1
Green Party	Dummy	0.14	0.34	0	1
New Zealand First	Dummy	0.08	0.28	0	1
(v) ‘Socio-Demographic and Psychological indicators’
Female	Dummy	0.52	0.50	0	1
Age	5-year intervals starting from 18 years	6.58	3.33	1	13
Children	Dummy	0.31	0.46	0	1
Married	Dummy	0.62	0.48	0	1
Secondary school qualification	Dummy	0.26	0.44	0	1
Polytechnic qualification or trade certificate	Dummy	0.20	0.40	0	1
Bachelor’s degree or higher	Dummy	0.41	0.49	0	1
Self-employed full time	Dummy	0.06	0.24	0	1
Self-employed part time	Dummy	0.05	0.22	0	1
Town	Dummy	0.28	0.45	0	1
Rural	Dummy	0.20	0.40	0	1
North Island	Dummy	0.43	0.50	0	1
Auckland	Dummy	0.32	0.47	0	1
NZ European	Dummy	0.68	0.47	0	1
Maori	Dummy	0.04	0.19	0	1
Asian	Dummy	0.10	0.30	0	1
Risk propensity	Continuous variable that varies between −1 (maximum risk aversion) and +1 (maximum risk propensity). We assessed the interviewees’ risk preferences by confronting the interviewees with the choice of either receiving a safe payoff or taking part in a lottery.	0.03	0.65	−1	1
Future-oriented time preference	Continuous variable running from 0 (completely impatient) to 1 (completely patient). Two experiments are conducted to assess the respondents’ time preferences in order to account for the fact that many people are more patient in the long run than in the short run.	0.61	0.28	0.29	1
Short-run impatience	Continuous variable running from 0 (completely impatient) to 1 (completely patient). Two experiments are conducted to assess the respondents’ time preferences in order to account for the fact that many people are more patient in the long run than in the short run.	0.56	0.27	0.29	1
Time spent on survey	Time respondent needed to fill out the questionnaire (in hours)	1.62	11.3	0.06	194