Line of Inquiry: Michael Weber on How Central Banks Should Communicate

Michael Weber: The models policy makers typically use to determine the effectiveness of monetary and fiscal policy have a common assumption that there’s a representative agent—so you and I—that has all available information, then forms expectations rationally, and whenever there’s new information, would update his or her expectations, and then crucially also his or her consumption plans. 

Yet, in the data, we oftentimes see that policies that at least theoretically, in those types of models, should operate through the very same mechanism and channels are vastly differently effective in affecting consumption and savings decisions of broader populations like in Europe, or in the US. For example, two very stark differences are so-called forward guidance announcements, or unconventional fiscal policies. 

So imagine unconventional fiscal policy relies on the idea that the government preannounces a future increase in consumption taxes, and then it’s fairly straightforward for ordinary people to understand: before taxes go up, prices are more expensive. Maybe I should go out, buy a new car, buy new furniture, and so on. And indeed, in the data, we see that those types of policies, very simple and straightforward to communicate, are very effective for the broader population to raise their inflation expectations, but crucially also their consumption spending. 

Instead, a policy that also theoretically works through the same intertemporal substitution margin, forward guidance, is actually not effective at all in stimulating overall consumption and inflation expectations.

So the idea here is whenever the Federal Reserve lowers interest rates, this then lowers the mortgage payments individuals have to make, and through that mechanism, at the end of the month, individuals have more available income, and therefore can actually start spending more. However, at a very basic level, there are two different kinds of mortgages. On the one hand, we have adjustable-rate mortgages, and on the other hand, we have fixed-rate mortgages. So what does it mean for the transmission of policy? 

Whenever we see a decrease in the federal funds target rate—so the key policy instrument of the Federal Reserve—this typically maps one-to-one into a reduction in mortgage payments for all individuals that have an adjustable-rate mortgage. Theoretically, of course, you could also assume that fixed-rate mortgage holders might see a reduction in mortgage payments, but in this case, it’s not automatic, but instead individuals would have to go to their bank, start to refinance, and then actually after locking in a lower rate for, again, a fixed-rate mortgage, they would also get benefits of reduced mortgage payments because of an interest rate cut by the Federal Reserve. 

However, we oftentimes see in the data that individuals, despite being possibly able to save thousands of dollars per month in mortgage payments, that they will not go to their bank, that they do not refinance their mortgages. And so, therefore, if you compare the effectiveness of monetary policy through this refinancing, or maybe directly through this adjustable-rate-mortgage channel, we see that in countries like, let’s say, Finland or England, that largely have adjustable-rate mortgages, monetary policy is potentially way more powerful and effective compared to countries like the US, where a big chunk of outstanding mortgages are fixed-rate mortgages. 

And so, therefore, to have this power of monetary policy through mortgage payments, we do have to rely on individuals refinancing their mortgages, and so, therefore, this requires some additional communication, both from maybe directly banks, but crucially also the federal government and the Federal Reserve, to convince individuals that they actually should go out and refinance their mortgages when policy rates drop to take advantage of those lower mortgage payments. 

But we oftentimes see that those announcements by the government and the Federal Reserve that rely on this to raise inflation expectations oftentimes do not reach broader populations, which then of course also means that there might be the concern of an implicit redistribution from the part of the population that does adjust consumption plans to the part of the population that is actually not adjusting. And so, therefore, it raises concern that maybe there is a part of the population that might actually be disadvantaged by those policies. 

And to get an idea, actually, whether this is partially indeed what’s going on in the data, we were able to access data for all men in Finland. The reason we only have males in our population is the data availability that comes from the military, where we got from the military-entrance-test-type data on IQ for the whole male population in Finland, which then allowed us to study, on the one hand, whether there is indeed heterogeneity in the expectations formation by cognitive abilities, but crucially also in this intertemporal substitution channel, and the adjustment of mortgage payments to changes in interest rates. 

And what we really found really stark in the data is that the top 50 percent of the population by cognitive abilities, at least in Finland, they behaved like our textbook rational agent in the microeconomics course that typically policy institutions would actually base their inference on. So we see that when they raise their inflation expectations, they want to consume more today. When the European Central Bank lowers interest rates, they will actually go out, take out more loans. When the European Central Bank raises interest rates, they will actually lower their propensity to take out loans. 

Instead, the bottom 50 percent of the population—by cognitive abilities—they have completely insensitive consumption plans to changes in interest rates. They are worse at forming inflation expectations, and they do not adjust their outstanding mortgages to changes in interest rates. So, therefore, there is this big concern that there might be an implicit redistribution by common policies for maybe the less fortunate part of the population, at least by cognitive abilities, to the more fortunate part of the population. 

To understand whether there are ways to possibly strengthen the communication and effectiveness of those policies, we recently cooperated with the Kilts Center for Marketing at the University of Chicago Booth School of Business, and ran our own customized survey together with Nielsen on their Homescan Panel, to get an idea of whether individuals would adjust their inflation expectations to targeted communication, possibly directly coming from the Federal Reserve. 

And the very basic idea is: imagine you ask individuals, “What are your one-year-ahead inflation expectations?” Then you randomly provide different pieces of information to different parts of the survey population, and then subsequently you listen again what their inflation expectations are. 

What we found in this setting is that potentially, if you are able to directly communicate with individuals, very simple messages are very powerful in moving expectations. For example, just telling individuals: “Inflation currently is, let’s say, 1 percent.” “The Federal Reserve targets inflation over longer periods of time of 2 percentage points.” Or: “The forecast by the Federal Reserve is a certain number.” Those simple messages were actually very, very powerful in moving the expectations of a broad population. 

Policy makers instead often would argue they mainly focus on financial-markets participants, because by committing to a path of short-term interest rates, they would be able to possibly also affect long-term interest rates, which, of course, are directly important for firm investment, and household consumption decisions. And then, indirectly, the hope is that the media would pick up those communications, and then, let’s say, individuals would read about it in the New York Times, the Wall Street Journal, or any other of the big newspapers. 

So we tried to understand whether that indirect transmission of policy news through the media is actually effective in shaping expectations, similar to the direct communication we documented in this experiment. We then actually gave individuals in our survey just a recent coverage in the news media of the Federal Open Market Committee policy meeting. We saw that actually this text, which conveyed way more information than just those simple messages, was actually only half as effective in moving expectations. 

Now, of course you could say: simple message, clearly communicated—people understand it. Long text, lots of content—maybe it was just an information overflow. To get an idea whether it was just the length of the article, or the complexity, we gave another population in our survey just the most recent statement by the Federal Reserve to read, about twice in length as the newspaper article, written in way more complex English compared to the newspaper article. Instead we saw that the whole FOMC statement, written for a technical audience, was as effective as those simple messages. 

So one conclusion from all those findings, of course, could be, if only half of the population reacts to certain policies, I just actually have to maybe scale the policy by a factor of two to get the same outcome. But, of course, this totally ignores what we discussed earlier, that there is this concern for implicit redistribution from the part of the population that does not react to the policy to the part of the population that does react to the policy. 

And so, therefore, I think a more constructive solution would be to implement these direct forms of communication from the policy institution like the Federal Reserve to the broader population. Use simple messages instead of complex text. And actually also try to directly communicate and explain what those policy changes mean for our consumption and saving decisions.