A key baseline assumption of many economists and central bankers is that higher interest rates slow the economy while lower interest rates stimulate it. 

Another widely held assumption is that inflation is mostly caused by “excess demand” (spending that exceeds the economy’s productive capacity), and the two combine to create the view that the way to fight inflation is to reduce demand (slow the economy) by raising rates.

Question assumptions that are not supported by data

In my view, these widely held assumptions are unreliable and do not fit the data of recent decades well, but they are complicated topics. In this note I will focus on one element: do higher interest rates reliably restrain household income and spending, and thus economic growth? Not necessarily, and only somewhat indirectly.

The key point that is often overlooked is that households in aggregate have more income from interest-bearing assets than interest expenses. So higher rates increase both income and expenses (while lower rates may reduce both), and the net effect of the two depends on the details of what kinds of assets and liabilities household have. That is, the type and duration of their interest-bearing assets and interest-bearing liabilities (e.g. money market funds vs 30-year mortgages).

Relevant factoid: about 89% of US household debt is fixed rate, only 11% is floating rate.

Household Debt is Mainly Fixed Rate

So changes in short-term interest rates have limited impact on much of the household debt outstanding, and changes take a long time to flow through.

The other key point, however, is that the net interest income (income vs payments) for households is, in aggregate, a small proportion of total household income. While $600-800 billion of aggregate net interest income sounds like a lot, total personal income is currently running at about $23 trillion per year. This means net interest income has been perhaps 3% of total income in recent years.

Notably, as shown in the chart below, net personal interest income is only very loosely tied to the level of the Fed funds policy rate. The common assumption is that higher rates rapidly push up interest costs to households and thus reduce aggregate household net income, while lower rates do the opposite, but we do not see this in the chart. This is due to the fact that rate changes affect both interest income and expenses, and the level and types of debt held and owed changes over time.

Source: Mill Street Research, Factset, Bureau of Economic Analysis

To put it in perspective, over just the last 12 months, aggregate wages and salaries alone have increased by about $700 billion (from $11.3T to $12T), which is more than the entire amount of annual non-mortgage interest payments ($564B currently), and not far from the amount of all interest payments by households.  And aggregate dividend income to households of $1.8T (of similar magnitude to interest income) has been more than all household interest payments combined.

Clearly, higher interest rates have pushed up interest costs for households in absolute terms, particularly non-mortgage interest. But notably, total mortgage interest paid is only back to where it was in 2007-08, even though total mortgage debt outstanding is much higher now ($20T vs $14T). This is because the average mortgage rate households are paying is still much lower now than in 2007 despite current mortgage rates being above 2006-08 levels. The average mortgage rate has risen only modestly due to the fact that many homeowners still have low-rate mortgages from before 2022. It can take years for average mortgage rates to shift, particularly when people are not moving as often, and there is a disincentive to refinance.

Non-mortgage interest payments have clearly risen, as they include more variable-rate debt like credit cards or home equity lines of credit, along with things like auto loans that have much shorter average durations than mortgages. But even after a big increase, the aggregate amount of interest paid is still far below total interest received.

So the net amount of aggregate income changes to households caused directly by interest rate changes can be easily dwarfed by movements in wage income and other items. In other words, if you are trying to constrain consumer spending via the impact on aggregate personal income, simply changing interest rates is not going to do much.

Rate changes may affect spending more than income

Where rate changes may have more impact is from the distribution of income, and thus on spending. Of course, not all households have assets and liabilities in proportion to the aggregate figures. Some wealthier households have lots of interest-bearing assets and few liabilities, while less wealthy households may have proportionally larger liabilities and few interest-bearing assets. So higher interest rates will tend to increase the net income of those households with more assets and fewer liabilities, and decrease that of less wealthy households who must pay more on their liabilities.

Because economists know that a given household’s propensity to consume (spend) is highly correlated with income and wealth (less wealthy people regularly spend a higher percentage of their income than wealthier people), an interest rate change that shifts income (in aggregate) toward higher-income households and away from lower-income households may have more of a tendency to reduce consumption (rather than income) at the margin (all else held equal). That is, the hit to spending from paying higher rates is proportionally larger for lower-income households, since they do not have excess income (saving) to cut back and thus must reduce consumption, while higher-income households have more of a buffer on their spending in addition to getting more interest income on average.

So there is an argument that higher interest rates are, at the margin, regressive in that they hit lower income households (who tend to be net borrowers) harder than higher income households (who tend to be net savers). But again, the net impact of rate changes can easily be outweighed by larger factors like wage/employment growth and fiscal policy.

If higher rates do not reliably reduce total household income (or do so only very modestly), but could dampen spending because they are regressive and cause lower-income households to cut spending, is raising rates a good way to try to restrain the economy, assuming you would in fact want to? The same question would apply in reverse if policy makers wanted to stimulate the economy: would lower rates be the best way to do that?

Almost certainly, changing interest rates is not the most effective way to manage the economy, since by construction there are winners and losers within the private sector. That is, interest rates are a price (price of money) and so higher prices benefit suppliers (lenders/savers) and lower prices benefit buyers (borrowers). 

“Long and variable lags” = not reliable

We can see this in the data by the fact that monetary policy changes over the last 30+ years have had very mixed effects on the economy, along with the well-known maxim from economist Milton Friedman decades ago that the impacts of monetary policy changes have “long and variable lags”. Researchers at the St. Louis Fed wrote about this last year, noting that estimated lags can range from 6 to 24-30 months, quite a wide range given that business cycles may only last 3-4 years in some cases:

Examining Long and Variable Lags in Monetary Policy

So using monetary policy changes alone as a means of predicting the path of economic growth seems unreliable both empirically and theoretically. It also means that the effectiveness of changes in monetary policy depends heavily on the ability of central bankers to forecast economic trends well in advance, which is also not a reliable method empirically (the Fed’s economic forecasts, like consensus economists’ forecasts, are often wrong).

The difficulty comes from the fact that tighter monetary policy often coincides with other factors that can cause economic slowdowns: tighter fiscal policy, poor bank regulation, extreme stock market volatility, wars, pandemics, oil shocks, etc. (and vice versa), and so results get attributed to monetary policy when other factors are in fact more important (though monetary policy clearly does make things worse in some cases). This reflects the salience of monetary policy (the Fed has a press conference every 6-8 weeks nowadays, and testifies before Congress twice a year) and the implicit message from Congress essentially handing off economic policy to the Fed, most recently via the Federal Reserve Reform Act of 1977.

As noted at the outset, the impact of monetary policy is a complex topic, and we have not yet discussed the impact on corporate spending and investment, or the impact from government interest payments (spoiler: the US federal government is almost entirely a payer of interest and has the unique ability to create US dollars at will, so higher rates push more money into the economy via federal interest payments, and thus are in fact more stimulative than contractionary if all else is held equal). I will cover more on this topic in future commentary.

But looking more closely just at the impact (or lack thereof) of rates on household income should prompt a reconsideration of some of the widely held assumptions about the Fed’s near-term impact on the economy. Lots of other things (employment/wages, fiscal policy, etc.) can easily outweigh the impact of changes in short-term interest rates, at least in the short- to intermediate-term. This highlights the frequent “gap” between what drives stock and bond markets in the shorter-term and what actually drives the underlying economy and earnings. The Fed influences markets a lot (arguably too much), but influences the economy much less than most people assume.

This means focusing on every subtle shift in monetary policy may be a distraction from much more important (but less salient) economic drivers, which can lead even experts astray.

Sam Burns, CFA

Chief Strategist