Joshua Brault and Hashmat Khan

Ed Comment: “I’m not sure I understand the paper on inequality and interest rates, which I would like to include. Four questions. First, is the regression/data US or worldwide? Second, how do they separate the fact that as the world, chiefly China, has gotten richer, it has also saved more? Similarly, as the US has gotten richer the world has also saved more, flooded the US with risk-averse savings, and the rate has declined. They are correlated, but is it causal? How do they stake their claim that it’s casual? Third, I’m not sure I understand the 2 graphs. Of the top graph, the text says: “Figure 3 quite clearly shows a minimal relationship between income inequality and trend growth rates of the natural rate of output.” I thought the (mian) theory is that richer people save more rather than consume. We lack uses for investing all the savings (in effect a Keynesian paradox of thrift). So the interest rate declines to equilibrate. Is so, then doesn’t (the growth in) demand have to decline to reduce the rate? If declining rates equilibrate the supply and demand for savings, why do we care? An increase in savings would increase investment, which would presumably accelerate growth/increase productivity/raise standards of living. It only seems like a problem if it produces a Keynesian paradox of thrift, i.e. if demand declines (is correlated with inequality). Help me understand their story. Fourth, how do they differentiate between risk-averse savings that results in debt and risk-bearing equity? let’s say I insist on holding more low-risk debt for every dollar of risky equity I invest. (risk-bearing investors hold about 70% equity, not 100%) Low rates might be a “necessary evil,” and the resulting slowdown from any paradox of thrift might be offset by an acceleration of growth (relative to the counterfactual). I’m not sure how to interpret the results.”

Ben/Steve Comment:” Ed: First, is the regression/data US or worldwide? Ben: The Brault/Khan note that was in the newsletter today is a comment on a Mian, Straub, and Suffi “Indebted Demand” paper (attached) The regression run in this comment is on international data (Australia, Canada, EuroZone, Norway, UK, and US). The indebted demand paper also uses international data but averages the data across countries. There is another Mian, Straub and Suffi paper called “The saving glut of the right and the rise in debt”. Steve did a write up of that paper before I joined. (Also attached) That paper focuses on the US. Ed: Second, how do they separate the fact that as the world, chiefly China, has gotten richer, it has also saved more? Similarly, as the US has gotten richer the world has also saved more, flooded the US with risk-averse savings, and the rate has declined. They are correlated, but is it causal? How do they stake their claim that it’s casual? Ben:The short answer here is: they don’t. First, China and other countries developing countries (India, etc) don’t even appear in the data. Secondly they use time fixed effects to control for global trends like those you’re describing. “Indebted Demand” uses “motivating facts” (not causal estimation) to motivate its model and then inequality drives interest rates within the model. This paper is not really an empirical exercise. Finally, in the savings glut of the rich, Mian, Straub, and Suffi have this explicit warning: "we acknowledge up front that there is no specific source of exogenous variation in the rise in inequality used in this study,” which is to say, these are not causal estimates.Ed: Third, I’m not sure I understand the 2 graphs. Of the top graph, the text says: “Figure 3 quite clearly shows a minimal relationship between income inequality and trend growth rates of the natural rate of output.” I thought the (mian) theory is that richer people save more rather than consume. We lack uses for investing all the savings (in effect a Keynesian paradox of thrift). So the interest rate declines to equilibrate. Is so, then doesn’t (the growth in) demand have to decline to reduce the rate? If declining rates equilibrate the supply and demand for savings, why do we care? An increase in savings would increase investment, which would presumably accelerate growth/increase productivity/raise standards of living. It only seems like a problem if it produces a Keynesian paradox of thrift, i.e. if demand declines (is correlated with inequality). Help me understand their story. Ben:For the graphs: Equation 1 in Brault/Khan paper says the natural rate of interest is the growth rate (g) plus other stuff (what they call z). The first panel shows the relationship between inequality and g and the second panel shows the relationship between inequality and z. The relationship between inequality and z is very strong and that’s what puts the downward pressure on the interest rate. The story really comes from the two Mian, Straub and Suffi papers previously mentioned. Their story is that, as the rich save more, they finance the consumption of the rest of the people in the economy. So the rich having lots of money is more of a supply story and less of a demand story in the lending markets (as you pointed out by describing the paradox of thrift). However, what Mian, Straub and Suffi are pointing out is that for every dollar saved by a rich guy, there is a dollar borrowed by a poorer guy (approximately, there are non-consumption-financing savings and equity etc, but to a first approximation). These borrowed dollars are exactly why demand (for goods and services) doesn’t go down. Ed: how do they differentiate between risk-averse savings that results in debt and risk-bearing equity? Ben: Today’s note doesn’t deal with anything like this. Both Mian, Straub and Suffi papers look at the net- debt position of rich households vs the net debt position of poorer households. This measure nets out equity - both intermediation and risk bearing equity - but doesn’t make distinctions finer than this. Steve: this all connects to research that Mian/Straub/Sufi presented at Jackson Hole (attached) that argues that the rise in inequality, rather than demographics has driven down the interest rates. The basic mechanism is the rich have a greater propensity to save and the resulting risk adverse capital drives down r*, note this study looks at the United States (so it’s not accounting for incoming hot money) Bottom line, "...Downward pressure on the natural rate of interest (r*) is often attributed to an increase in saving. This study uses microeconomic data from the SCF+ to explore the relative importance of demographic shifts versus rising income inequality on the evolution of saving behavior in the United States from 1950 to 2019. The evidence suggests that rising income inequality is the more important factor explaining the decline in r*. Saving rates are significantly higher for high income households within a given birth cohort relative to middle and low income households in the same birth cohort, and there has been a large rise in income shares for high income households since the 1980s. The result has been a large rise in saving by high income earners since the 1980s, which is the exact same time period during which r* has fallen. Differences in saving rates across the working age distribution are smaller, and there has not been a consistent monotonic shift in income toward any given age group. Both findings challenge the view that demographic shifts due to the aging of the baby boom generation explain the decline in r*...."
Ed Comment: “Two thoughts First what matters is at the margin. So the savings of the rich may increase, but if (productive) investment also rises then we don’t get the effects that concern M&S (increased consumption or Keynesian paradox of thrift). It only matters when savings are greater than investment. Savings could be greater than investment if we run trade deficits for example AND rich people save more. But it might be the case that, were it not for trade deficits savings would not be greater than investment. From my prior readings, Mian and Sufi, seemed eager/determined to attribute all the increase to a rise in domestic inequality. That seems distorted/overstated to me. I believe the rich save a greater share of their income. (Surprisingly, others have argued otherwise.) I think it matters greatly whether they are risk averse or risk underwriting. If savers are risk averse, they deposit money in the bank (or money markets) and it predominantly end up funding mortgages, which can increase consumption (now it increased government debt, which also increases consumption but not via the free market). Or if savers invest in risky equity, it can result in a speculative boom, which can accelerate innovation, as we read about recently, So, it could be the case that rich Americans are saving but their savings are fueling a speculative boom while (at the margin) risk-averse offshore savers fund mortgages, gov debt, and ultimately consumption. To make a credible case that the rich/growing domestic inequality are to “blame,” we have to show that 1) it’s Americans, not risk-averse offshore savers who are funding consumption and 2) that the value of any speculative boom is less than the problems stemming form an offsetting rise in domestic risk-averse savings. That is, speculators may logically hedge the risk of speculation by holding risk-averse savings. So the slow down from unborrowed savings may be less than the acceleration of the speculative boom. My hunch is that, were it not for trade deficits, the system would largely be in equilibrium. My sense of M&S is that they are either stupid (don’t see these issues) or don’t care because they have an ax to grind. For example, the correlate things to inequality while ignoring trade deficits. I’m not say that’s the case. I’m asking how they think through the issues I’m raising.”

Steve Comment: “In none of M&S stuff that I’ve read they have drawn or acknowledge a distinction btw risk adverse capital (ze Germans) and risk seeking equity (USA!) I thinkthis Mian quoteis helpful in understanding their worldview, “What is savings? We all know what it is at a personal level: I get 100$ check and might save 20$ out of it But what is savings in the aggregate? Say for the world as a whole? That's where it gets tricky. e.g. people talk about a "global saving glut". But can we measure it? Not really - at least not in a direct sense The reason is that when a person (say Adam) "saves", someone else (say Eve) uses that saving, i.e. "dissaves", by borrowing it and spending it for one purpose or another So, aggregate total saving is always exactly zero! Now you might say, that according to the world bank worldwide saving is about 25% of GDP historically. What is that number about? Well it is just an accounting convention! Someone somewhere decided (for some good reasons at the time), that if Eve spends Adam's savings on capital goods (i.e. "investment"), then we will not count it as dissaving. So what you get is an accounting identity: Saving = Investment What the world bank data is telling us is that historically Eve borrows and spends about 25% of GDP on investment goods. Now imagine something starts to change Adam starts to save more and more, call it his "saving glut" He needs Eve to borrow his additional savings He convinces her to do that, but only after making it cheaper for her to borrow by lowering his interest rate However, Eve is not interested in spending her additional borrowing on capital goods, or investment Instead Eve decides to "consume" it - throws a party, goes on vacation, that sort of stuff The accounting convention will start counting Eve's additional spending as dissaving What this all means is that even though Adam *is* saving a lot more than before, total saving measured by our accounting standards will remain the same. Yes, there is a "saving glut" - But aggregate national accounts will not show it! So if you really wanted to see the saving glut, you will have to see it indirectly You will see it as a rise in total credit (remember Eve is borrowing more and more), as a fall in interest rate, and as a rise in Adam's personal savings I think this is the world we actually live in now Adam is the top 1% And I have no idea why Eve does not invest more, but consumes her additional borrowing ps: in recent years government has been borrowing and spending on Eve's behalf... but same thing more or less”..."
Ed Comment: “A few thoughts. A lot of value gets captured by real estate. So eve may not save, but her real estate may nevertheless permanently rise in value as owners save, invest, take risk and innovate. She may borrow against the increased value of her home and consume the proceeds to adjust her desired amount of savings An investor might own stocks in companies that pour money into risky investments, and seem over valued. They may hold cash as an offset to manage their risk levels. I start by assuming the world is fairly logical. People determined to change the world/defeat the status quo start by assuming the world is mistaken/broken. They are confident they can out think markets/the economy, as if they could out-think nature and evolution. Most likely they are wrong because the world is too complex and their analysis is too simple to accurately account for all the tradeoffs.”