Could knowledge about Central banks impact households’ expectations?

Emma Rockall

Should central banks care if people understand them? Whereas once Alan Greenspan famously declared: “If I seem unduly clear to you, you must have misunderstood what I said”, central bankers now dedicate considerable time and thought to transparency and communications. While transparency initiatives have value in their own right in improving accountability, results from the Bank’s Inflation Attitudes Survey suggest that they could have potentially far-reaching effects on the economy through their impact on households’ expectations. If they improve households’ knowledge of central banks, they may produce inflation expectations that are more stable and closer to the inflation target in the medium term – that is, ‘better-anchored’ expectations.

How might institutional knowledge affect expectations?

‘Institutional knowledge’ captures households’ awareness and understanding of central banks, rather than the economy more generally. While a large body of work has examined the impact of central bank transparency and communications on the expectations of financial markets and professional forecasters (see, for example, Blinder et al. for a survey of the theory and empirical evidence), much less has been done to look at the impact on the expectations of households. From a theoretical perspective, the expectations of households are no less important than those of financial markets.  It is reasonable to suppose that the more someone knows about a central bank and how it conducts policy, the more confidence they will have that the central bank will act to bring inflation back to target.  This appears to be borne out by the IAS, which shows that higher knowledge scores are associated with a statistically significant improvement in households’ view of the Bank of England’s credibility, and their confidence that the MPC will return inflation to target. Greater understanding of a central bank may also help households understand how the central bank might adjust interest rates to meet the inflation target, thereby improving the accuracy of their interest rate expectations (with associated financial planning benefits).

There have also recently been some attempts to try and quantify the impact on households’ expectations empirically. Binder shows that following the Fed’s announcement of a 2% inflation target, households’ inflation expectations were better anchored, but that anchoring increased more for more well-informed households compared to less well-informed households. And Haldane and McMahon, using the institutional knowledge score discussed above, show that for the UK, higher knowledge corresponds to greater satisfaction with the Bank, and inflation expectations closer to 2% at all horizons.

…and why should central banks care?

If institutional knowledge does effect households’ expectations, this could have far-reaching implications for the economy more broadly. Woodford goes as far as to argue that, ‘Not only do expectations about policy matter, but, at least under current conditions, very little else matters.’ In theory, households’ inflation expectations feed through into a range of real economy variables, by affecting wage setting, the timing of consumption, and saving and borrowing (by changing perceptions of real rates). And empirically, a number of papers (e.g. Crump et al., D’Acunto et al. and Duca et al.) use household surveys to show that households with higher inflation expectations are more likely to move forward consumption or increase spending on consumer durables. And 16.5% of respondents to the IAS in 2018 said they would try and increase their income (e.g. by asking for a pay rise) as a result of their 1 year inflation expectations. Only 3.3% said they would take no action. How well-anchored households’ inflation expectations are can also have important implications for the persistence of inflation, and the trade-off between inflation and output that central banks face when trying to bring inflation back to target.

How can households institutional knowledge be measured?

I construct a measure of individual households’ knowledge about the Bank of England specifically (their ‘institutional knowledge score’) based on three questions from the Bank/TNS Inflation Attitudes Survey (IAS) (see the technical annex for details on how the score is constructed).

From this score, one can see how households’ institutional knowledge of the Bank has evolved (Figure 1). Peaking in 2005 at 4.3, the mean knowledge score (for all households) fell post-crisis to a low of 3.9 (although it has recovered slightly since). The aggregate picture marks a more dramatic fall for households with below GCSE level qualifications, whose average knowledge score fell from 4.2 in 2004 to a low of 3.4 in 2016. For some households the level of institutional knowledge is very low – in 2018, 24% of households answered none of the questions correctly.

Figure 1: Households’ mean knoweldge scores, by qualifications held

Evidence from the Inflation Attitudes Survey

Using, microdata from the IAS I explore the relationship between households’ knowledge scores and the differences between their inflation expectations and the MPC’s forecast for inflation one and two years ahead (with the 2% target taken as the MPC’s forecast 5 years ahead).  The survey also asks households what they expect Bank Rate to be over the same horizons. Although the MPC doesn’t publish a forecast for Bank Rate, I can compare households’ expectations with those of financial markets (captured in the yield curve at the relevant horizon, which the MPC conditions their forecasts on). The heatmap below shows the number of households in each ‘bucket’ (comprising the possible combinations of knowledge scores and deviations in expectations) relative to what you would expect if expectations and institutional knowledge were uncorrelated. The ‘hot’ diagonal (running from top right to bottom left) suggests that on average, higher knowledge scores correspond to smaller deviations in households’ inflation expectations from the MPC’s forecast (Figure 2); households’ Bank Rate expectations show a similar pattern.

Figure 2: Heatmap of households’ knowledge scores and inflation expectation deviations 5 years ahead

Regression analysis shows that these relationships are robust to the inclusion of household-specific controls. Regressions (1a) and (3a) (Table 1) show that even with controls, households with higher institutional knowledge scores tend to have inflation expectations closer to the Bank’s forecasts at both short and long horizons (see the technical annex for results at the 2 year horizon, which are similar to those at the 5 year horizon). And regressions (4a) and (6a) show that the same is true of households’ Bank Rate expectations.

Not only are all coefficients on institutional knowledge statistically significant at the 1% level, comparison to the coefficients on the controls would suggest they are also economically significant. For example, all else equal, moving from the lowest knowledge score to the highest would imply five year inflation expectations 0.56pp closer to the 2% target, and five year Bank Rate expectations 0.34pp closer to those of financial markets. This is larger than the impact of any of the controls (including age, often thought to be an important determinant of inflation expectations). See the technical annex for the full results of all estimations. 

Table 1: Regressions of households’ expectations on their knowledge scores

However, these results do not necessarily imply that there is a causal relationship from institutional knowledge to expectations. While reverse causality isn’t a major concern (it’s unlikely that better anchored expectations produce a better understanding of the Bank), it’s possible that external factors, not captured in the controls, produce both higher institutional knowledge and better anchored expectations. One plausible candidate would be knowledge about the economy more generally. In this case, institutional knowledge would be proxying for economic knowledge in the regressions above, and the impact of institutional knowledge alone would be overstated.

From the perspective of motivating greater central bank transparency, this might not actually matter. Central bank communication efforts are likely to have a positive impact on both institutional knowledge and economic knowledge, rather than only one in isolation. In this case, even if the results above overstate the marginal impact of institutional knowledge, they may still provide support for the usefulness of greater transparency.

However, I can also use the IAS to try and tease out the impact of institutional knowledge specifically. From questions on the transmission mechanism of monetary policy, I  construct an ‘economic knowledge score’ to try and control for broader economic knowledge (see the technical annex for details on how it’s constructed). Table 2 shows the results of the previous regressions controlling for economic knowledge. These demonstrate a role for economic knowledge – a higher score has a statistically significant negative impact on deviations in inflation expectations from the MPC’s forecast. At the same time, the coefficients on institutional knowledge remain statically significant (and largely unchanged), suggesting the results above are robust to controlling (albeit imperfectly) for broader economic knowledge.

Table 2: Regressions of households’ expectations on their knowledge scores, controlling for economic knowledge

Conclusion

These results add weight to the need for and desirability of central bank transparency. Above and beyond the important role communication efforts play in bolstering democratic accountability, the IAS would suggest that, to the extent that they improve knowledge about central banks, they could help to better anchor expectations, thereby improving the efficacy of monetary policy. While there is still much to learn about how households form their expectations and how those expectations feed through into the real economy, this evidence provides support for recent initiatives undertaken by central banks (such as at the ECB, the BOI, and at the Bank of England – see here and here) to make their communications more accessible to more people.

View Technical Annex

Emma Rockall works in the Bank’s Structural Economics Division.

If you want to get in touch, please email us at bankunderground@bankofengland.co.uk or leave a comment below.

Comments will only appear once approved by a moderator, and are only published where a full name is supplied. Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.

1 Comment

Filed under Macroeconomics, Monetary Policy

One response to “Could knowledge about Central banks impact households’ expectations?

  1. Love this!! I was doing some work on this but I’m still quite inexperienced so its great to have an example of how to create a knowledge index and something to reference my own analysis against. My pet bugbear is people who moan and post vitriol on the Bank Facebook page but actually have no clue about how monetary policy actually works. The danger is that if people don’t understand when rates will rise, they could change their spending habits and ‘undo’ the intended effects of policy. Its interesting that the personal financial expectations index dropped sharply after the Brexit vote but quickly bounced back after QE was announced which suggests that consumers were more worried about a rate rise that Brexit itself*. (Although there could also be some effect where they believed that the Bank action would work.)
    *Even though there was talk of rate cuts, some feared they would rise to support sterling.