Steven M. Fazzari and Barry Z. Cynamon, researchers long interested in consumer behavior, wrote an op-ed for the St. Louis Post Dispatch in October, 2007 in which they predicted that an end to the relentless trend of rising household debt and a subsequent crash in household spending could lead to a surprisingly deep U.S. recession. Soon after, their prediction came true in the aftermath of the global financial crisis.
Fazzari and Cynamon’s current work, which examines how high and rising inequality is holding back the American economy, is part of the Institute’s project on the ” Political Economy of Distribution.” In two papers, “Inequality, the Great Recession, and Slow Recovery” and “Household Income, Demand, and Saving: Deriving Macro Data with Micro Data Concepts,” they explore how the massive debt which led to the Great Recession, the spending collapse that followed, and the stagnation that persists are all linked to income inequality. In the following interview, they discuss how their findings challenge conventional economic views.
Lynn Parramore: Why have you decided to focus on inequality in your research?
Steven M. Fazzari: First, as Barry and I developed our research on household spending, debt, and big movements of the U.S. economy, it became clear that the overall trends before 2008 were unsustainable. Households were taking on debt at a rate that couldn’t continue indefinitely. But for each borrower, there is a lender. We got interested not just in the overall trends of debt and spending, but also who was doing the borrowing and lending; which part of the household sector was spending unsustainably? This question led us to think about inequality in a very general sense: the fact that some households borrow and others lend implies that households are different, that is, “unequal.”
Second, there has been a lot of attention on data that demonstrate rising inequality in the U.S. Thomas Piketty’s big book, Capital in the 21st Century, came out in English last spring, but some of the articles his book is based on appeared earlier. I’ve always had an interest in income inequality, perhaps mostly from the perspective of American values and social justice. So I was curious to look at these new data sources. It turns out that the rise of the income share of high-income households begins at almost the same time as the rise in debt that was the focus of our research in the mid 2000s. It seemed like this correspondence was not a coincidence. I began to see a convergence between the dynamics of inequality and the macroeconomics of U.S. expenditure.
LP: An ordinary person on the street would probably say that if the rich have most of the money, that’s bad for the economy. She’d intuit that if the rest of us don’t have enough money in our pockets to spend on goods and services, the overall economy will suffer. Yet this has been minority view in the field of economics. In fact, many economists have long argued that economic inequality was good for economic growth, that there was some kind of tradeoff between “equity and efficiency.” What do you think explains the persistence of the conventional view?
Barry Z. Cynamon: When you study a discipline for several years, you learn to be skeptical of common sense. The intuition of the ordinary person on the street is not necessarily any better for economic questions than it is for astrophysics. Being a scientist means digging deeper, looking for alternative explanations, and accepting conclusions only when evidence is sufficient to justify them. That said, sometimes a profession can stagnate due to the faith that its members have in the rejection of a particular common sense notion.
The equity-efficiency tradeoff is conventional wisdom among economists, and it stems from two different articles of faith that economists have adopted in order to discipline their thinking and avoid falling for an intuitive but incorrect conclusion. First, economists remind themselves to always look at incentives. The idea of the equity-efficiency tradeoff is that the greater the incentives to be a brilliant success, the harder people will work to achieve success. If the CEO makes 300 times the wage of the average worker instead of 30 times the wage, then there is a much steeper incentive for talented people to work as hard as possible to become CEO. Extracting more hard work from more people is viewed as a path to more output and therefore a larger overall pie. Dividing the pie more equally (equity) would lead to a smaller pie overall (efficiency), because the incentives will be flatter.
There is evidence that getting the best candidate into the CEO role has a huge impact on the profitability of a company. Not only does the CEO of a large company have a large lever to magnify the impact of their decisions, but the incentive effect of the high top salary means that many other people in subordinate positions are trying very hard to maximize their opportunity of becoming a richly rewarded CEO—with the assumed implication that they are doing better work. Thus, the steeper incentive elicits strong effort from many managers and hopefully gets the best person in the corner office making the big decisions.
The second assumption embedded in this reasoning is that not just effort but performance is proportional to monetary compensation and not any other incentive. Said differently, if the top salaries were lower, marginal income tax rates on the highest incomes higher, or capital gains taxes higher, then Steve Jobs would not have created the miracles of Apple and Pixar, and we’d be worse off as a society (smaller pie).
Nobody is debating that if your behavior had no impact whatsoever on your compensation that you would necessarily show up to work every day and do your best. It would be nice if that were true, but it probably isn’t true for a lot of people. That said, the question really boils down to when one effect swamps out the other. If some amount of inequality is good, then that doesn’t mean that more inequality is necessarily better without qualification.
As far as the ordinary person on the street: their view is right up to a point. If the rich don’t want to spend their money, then putting more and more income into their hands will reduce economic growth. That is the crux of our argument. On the other hand, if the rich have most of the income and actually do spend it, then that’s no different really from the fact that we don’t worry about billions of people in emerging markets having tiny incomes and not spending in a way that is of much relevance to the global economy. If we were willing to tolerate a permanent underclass of low-income and low-spending Americans, then we could reach an equilibrium in which the rich would be doing just fine.
If there is no effective price adjustment or monetary policy mechanism to cure the demand problem created by income inequality, we have the potential for persistent, or “secular,” demand stagnation. We argue that this is the current situation for the U.S. economy. Thus, the intuition you describe of the person on the street is, to a large extent, validated by the kind of work we are doing.
LP: What’s distinctively new in your work on the relation of inequality to growth?
SMF: By now, it is widely (although not unanimously) accepted that the financing of demand growth prior to the Great Recession was unsustainable: it required an upward trend in household debt relative to household incomes that would eventually have to come to an end. When the trend did finally end — what we called the “end of the consumer age” in earlier work — we got the most severe recession in seven decades and a historically weak recovery.
Our paper links this basic history not just to unsustainable consumer debt, but also to rising income inequality. We show that U.S. household debt relative to income began to rise at roughly the same time that income inequality started to increase, after being quite stable in the decades after World War II. We use some simple balance sheet relations to argue that when a group’s income growth declines, as occurred for the bottom 95 percent of the U.S. income distribution around 1980, that group needs to cut back the share of its income that it spends to avoid an unsustainable explosion of debt. We present original data, however, that show that the bottom 95 percent did not cut back spending relative to income, and, as predicted by our model, the debt ratio for this group exploded. The balance sheet of the top 5 percent looked just fine. When the Great Recession hit, our data show that it was the bottom 95 percent that cut consumption relative to income, most likely because their access to new credit was cut off. This was the first major decline in bottom 95 percent spending to income in at least 20 years. The top 5 percent actually increased the share of income that they spent during the Great Recession, just like they did in earlier recessions covered by our data. We are the first to show this strongly contrasting spending behavior across income groups on opposite ends of rising inequality. We propose that these facts show that the explosive debt that led up to the Great Recession, the spending collapse that caused the Great Recession, and the stagnation that persists in the aftermath of the Great Recession all have fundamental characteristics tied to rising income inequality.
The bottom line for us is that, ultimately, the “person-on-the-street” view is correct that concentrating too much income in the hands of the rich will hurt business sales and compromise job creation. But the group losing out as the result of rising inequality was, for about two decades, able to postpone these consequences by borrowing. That process has come to an end and it explains much the recent stagnation of the US economy.
LP: Tell us about some of data challenges you faced and how you overcame them.
SMF: It is critical for us to understand how the relation of spending to income evolved across different groups that have been affected differently from rising inequality. It is unfortunate, but there is no good source of consumption data that can be split up this way using straightforward methods. We therefore had to use a rather complex set of techniques, and rely on some detailed work by other analysts, to construct the data we needed for this analysis.
It was also somewhat of a challenge to think through how incomes, spending, and balance sheets are all related. These kinds of linkages seem fairly simple after you work them out. But this kind of analysis has not been very common in macroeconomics, and working out the framework for our analysis involved a number of false starts and revisions to get things right.
LP: Why is the divergence in spending and income between the top 5 percent and the bottom 95 percent dangerous?
SMF: I would identify two big problems. First, as we emphasize in our work, the economy needed the higher consumption of the bottom 95 percent to keep job growth acceptable. The economy wasn’t exactly booming in 2005 and 2006 when there was a lot of debt-financed consumption from the bottom 95 percent. But a lot of this demand was lost when the borrowing of this group collapsed. The top group may be moving along just fine, probably roughly along the trend it was following prior to the Great Recession. But it has not picked up the slack from the debt-financed spending of the bottom 95 percent that we lost, which we identify as a primary reason that the recovery has been so disappointing.
Second, as incomes grow faster at the top and the earnings of most household stagnate, more and more of the production in the society is being generated to serve high-end consumers. We are becoming less of a “middle-class” or “mass-market” society. Some of our recent work suggests that by 2012 the top 5 percent was consuming as much, in total, as the bottom 80 percent! As these trends continue, we are moving more and more to a society in which the work of the middle class (if they are lucky enough to have jobs at all) goes to serve the consumption of the affluent. This trend is inconsistent with norms of shared prosperity that are part of the American Dream. It may ultimately lead to social tension and a growing sense of injustice in our economy.
LP: Many explanations have been given for why growth in the aftermath of the Great Recession has been so sluggish. Why is inequality such a crucial piece of the puzzle?
BZC: “Growth” is measured as the rate of change of GDP. The biggest component of GDP is personal consumption expenditures (PCE). Prior to the Great Recession, PCE was growing faster than the remaining components of GDP. That means that PCE growth was a major driver of GDP growth, on average, between the mid-1980s and 2007. PCE fell dramatically in 2008 and 2009, and the alternative measures we derive and present in our [Institute for New Economic Thinking] research on measuring demand suggest an even bigger plunge than in the standard PCE data. In order for GDP to bounce back, PCE would have to bounce back, too. Alternatively, government spending or business investment would have to grow very rapidly (because they are much smaller portions of GDP). The incentive for business investment is large markets to enter or growing markets to satisfy, so business investment probably wouldn’t lead a recovery. Government spending was constrained at the state and local levels for various reasons and at the federal level primarily due to politics and concerns – misguided in my view, under the circumstances – about rising deficits.
Ideally, government spending would get the ball rolling by putting money in the pockets of workers and contracts in the hands of businesses. In the absence of adequate government stimulus from spending, tax cuts, or transfers, the main channel for inflating the economy from a policy standpoint has been quantitative easing. The problem with monetary policy is that it works through financial channels, and increasing the value of financial assets by reducing the discount rate does not mean the same thing as asset prices rising because future prospects look better (e.g. higher sales and profit expectations). Boosting credit seems like a pretty bad idea if spending is sluggish not just because people have little income growth now, but they also have no good reason to expect an acceleration of income growth it in the future.
And that is the key. If income inequality is steadily rising, it means that the incomes of folks at the top are rising faster than the incomes of folks at the bottom (or shrinking less slowly, but that’s not relevant in the context of growth). When aggregate income growth and rising inequality coincide – as they did, on average – between the early 1980s and the present, it means that the bulk of the benefit of economic growth is flowing to the high income folks. You cannot borrow and spend your way to success unless the spending validates the borrowing. For example, if you start a business and borrow to buy an ice cream machine, then you can use the machine to manufacture ice cream, and the sale of the ice cream in the future will finance the debt you took on to acquire the machine. If all works well for you, the sales will more than make up for the debt payments, and you will be profitable. The logic for consumption is different, because for any individual, their consumption expenditure does improve the economy, but it doesn’t mostly flow back to improving their own personal situation. You don’t get a $1 raise for each extra $1 you spend. But if everyone like you spends another $1 then you might actually see an increase in total output and an increase in your income. However, if the people like you don’t see any increase in income as a result of your increased spending and the increased output, then you would not be able to service consumer loans taken to finance your purchases.
In my view, that is the problem here, and the key contribution of our framework. We are documenting with historical data the way that incomes did not rise for the bottom 95 percent, but their spending did not stagnate to the same extent as their income because they saved less, and the erosion of their balance sheets provided a boost to the overall economy — but not to them as a group – because most of the boost went to folks with the highest incomes.
LP: What’s next in the pipeline for your work? What additional research on inequality and growth needs to happen?
SMF: We have several different projects underway that were spawned by the work that we have discussed in this interview. First, it seems clear to us that the collapse in household spending that took place during the Great Recession is understated because of the way that the government measures personal consumption. We have developed some new measurement methods that magnify the severity of the spending decline. We hope to link these aggregate measures more clearly to the work on inequality and demand. Second, we have developed a framework to measure the financial sustainability of American households more directly. This work should give us more insight about exactly what kinds of households were most squeezed by the income trends of recent years. Third, we hope to build a more dynamic model of how rising income inequality affects the generation of consumption demand, output, and employment over longer periods of time.
We expect that all of these projects will demonstrate how closely linked income inequality is with the way in which our economy generates its spending and creates jobs. In particular, it will expand our insights into why the U.S. economy has under-performed in the post-recession era as the result of rising income inequality. Broadly this work adds a strong “demand-side” perspective to the more “supply-side” perspective of mainstream thinking about inequality and growth.