Monday, August 5, 2019

Right and Wrong about Consumer Debt

There is a story floating around in media saying that consumers are drowning in debt. No, they are not. Unfortunately, even the Wall Street Journal gets it wrong.


The Trump economy has put America on a better path than it has been on for at least 20 years. That does not mean we are out of the woods - we still have a welfare state that weighs down our economy, slowly drifting us into a state of industrial poverty. Our economy is struggling to even keep up three percent growth at the height of a business cycle.

I addressed the structural causes of this stagnation in my book Industrial Poverty back in 2014. Until we deal with those causes, i.e., until we roll back the welfare state, we will not be able to get off this path of stagnation.

It is urgent we do so. We have an entire generation of Americans who have come out on the labor market in this stagnant economy, trying to build lives for themselves. With every new class of high school graduates that enter the workforce or go to college, we backfill the working-age population with more people whose prospects of a life in prosperity get dimmer and dimmer.

These problems are now of a magnitude where they are beginning to show up in the media main stream. On August 1, the Wall Street Journal had a long article on consumer debt, reflecting the problems with economic stagnation. It did not, however, discuss the issue in a very good, productive context. Instead, it conveyed an assortment of conventional wisdoms with - to be honest - very limited bearing on economic facts and good analysis:
The American middle class is falling deeper into debt to maintain a middle-class lifestyle. Cars, college, houses and medical care have become steadily more costly, but incomes have been largely stagnant for two decades, despite a recent uptick. Filling the gap between earning and spending is an explosion of finance into nearly every corner of the consumer economy. Consumer debt, not counting mortgages, has climbed to $4 trillion — higher than it has ever been even after adjusting for inflation. Mortgage debt slid after the financial crisis a decade ago but is rebounding.
I have a lot of respect for the journalism over at the Wall Street Journal. Compared to almost every other news outlet they maintain high standards both in editing and in quality content. However, in this case they have not maintained their usually high standards.

Before we get to the details, let us listen a bit more to them:
Student debt totaled about $1.5 trillion last year, exceeding all other forms of consumer debt except mortgages. Auto debt is up nearly 40% adjusting for inflation in the last decade to $1.3 trillion. And the average loan for new cars is up an inflation-adjusted 11% in a decade, to $32,187, according to an analysis of data from credit-reporting firm Experian.
Lots of numbers being thrown out there. The story, again, is that America's middle class is struggling to maintain its standard of living, forcing them deeper and deeper into debt. However, once we start looking more closely at this narrative the wheels start coming off the wagon.

Take the student-loan number as an example. It tells us little since it is not contrasted against either historic-growth numbers, the personal income out of which it is paid for - or the income boost that college graduates experience as a result of their educational investment.

This is not to say that student loans are insignificant to those who have to pay them. They are a burden, and it has not gotten better since the federal government began guaranteeing some of those loans. In fact, the total of student loans owed by the public has increased by 27 percent since 2014 alone.

However, if we are going to get any sort of understanding of the debt burden that households face, we have to put numbers like these in context. The same holds true for the next number the Journal throws out: car loans. The eleven-percent real growth in the average loan over one decade sounds outrageous, but it is equal to just over one percent per year, adjusted for inflation. At that level, it is perfectly easy to explain: we can safely assume that the average car is getting one percent better each year in terms of quality, fuel efficiency and performance.

If someone offered you a brand new 2009 Camry and wanted the same price for it as you would pay for a brand new 2019 Camry, would you buy it?

Plain and simple: the number that the Wall Street Journal reports as contributing to the consumer-debt problem is just a reflection of consumers buying a better product.

Unfortunately, it goes downhill from here. The Journal claims that there is a "debt surge" and that it is caused in part by the Federal Reserve's Quantitative Easing policies. However, if we actually look at consumer debt as a share of personal income - the correct measurement of household indebtedness - we find that consumer debt (which includes all loans, from mortgages to student loans to credit card debt) has not really increased much in the past decade:

-in 2008, when the Great Recession started, consumer debt stood at 21 percent of personal income;
-in 2010, at the bottom of the Great Recession, the share was 20.1 percent;
-in 2014, when the slow Obama recovery was under way, it had risen to 22.4 percent;
-in 2018, consumer debt amounted to 21.8 percent of personal income.

In other words, over ten years consumer indebtedness has increased by a measly 0.8 percent of personal income. In fact, in the first two years of Trump's presidency it has declined by 0.9 percentage points, from 22.7 percent of personal income in 2016.

Better still, if we look at consumer debt as share of disposable personal income - in other words when income is adjusted for taxes - the decline in consumer debt is even more pronounced:

-in 2016 consumer debt equaled 25.9 percent of disposable personal income;
-in 2018, the first year with Trump's tax cuts in effect, the share had fallen to 24.7 percent.

The difference between the decline in debt burden on pre-tax personal income and disposable personal income is due entirely to the cuts in personal income taxes from the Trump reform. It is the equivalent of a debt relief of $63 billion.

The Journal tries to use an anchor for its debt-burden numbers, and when they do - toward the end of their story - they actually contradict the very narrative they are trying to build:
Counting all kinds of debt, including mortgages, consumers aren’t nearly as debt-burdened as they once were. In the fourth quarter of 2007, the last year before the financial crisis struck, households devoted 13.2% of their disposable income to debt service. In the first quarter of 2019, that number was 9.9%, largely due to low interest rates. Partly because of widespread refinancing, mortgage payments since the start of 2017 have claimed the smallest slice of disposable personal income in decades, in the low 4% range, according to Fed data. Other debt, such as auto and student loans and credit-card borrowing, consumed about 5.7% of disposable personal income in the first quarter. That was up from a low of 4.9% at the end of 2012 and back to 2009 levels. In contrast to a mortgage, most of this borrowing went to fund consumption.
That last point is almost derogatory in its tone, but it also reflects ignorance on behalf of the Wall Street Journal - a rare sight, for sure.

The fact of the matter is that debt has not at all become a heavier burden on consumers in recent years. The interest-rate cuts that the Journal refers to have actually not increased consumer debt. As the numbers mentioned above tell us, consumer debt has not become a bigger problem in recent years.

In fact, if we look at the long-term perspective on consumer-credit growth, the latest years are not very dramatic at all:

Figure 1
Source of raw dataFederal Reserve

Historically we have seen some significant swings in the growth of consumer credit:

1. During the recession in the early 1990s there was a dip in consumer-credit growth, primarily in installment loans (non-revolving credit). This dip was isolated to a couple of short years at the depth of the recession.
2. Right after the '90s recession the economy went into its strongest growth period on record. With GDP growing at more than four percent per year at the top, the Clinton economy opened up new markets for consumer credit. Revolving credit grew at rates we have not seen since then; the growth rates in installment loans have since been challenged but not surpassed.
3.  After almost a decade of relatively stable growth consumer credit plummeted into negative numbers during the Great Recession. Mortgages and other installment loans contracted for less than a year - the first time since, and more leniently than, the recession in the 1990s. Revolving credit, on the other hand, went into the negative in January 2009 and did not start growing again until December 2011.

The credit-card led revolving-credit category contracted by almost 15 percent before it turned up again. Due to the barely-noticeable dip in installment credit, total consumer credit increased by 4.1 percent during that period.

However, these numbers do not matter unless we put them in the context of personal income - something the Wall Street Journal fails to do. To begin with, let us look at consumer debt as share of total personal income (in other words, the actual debt, not the cost of debt as the Journal does). Let us also include two other variables that can help us explain the trends we see in consumer indebtedness, namely transfers (entitlements; E) and wages (W) as share of personal income:

Figure 2
Sources of raw dataFederal Reserve (Consumer debt), Bureau of Economic Analysis (Income, Transfers, Wages)

Bluntly: there is no surge in consumer indebtedness. We saw an uptick soon after the Great Recession, but no surge.

What we are seeing, though, is a long-term increase, from debt levels of around 15 percent of personal income in the 1950s to 20-22 percent today. This increase is what really matters, and it has its explanation in the role that wages and salaries play in personal income.

To understand this role we first need to dispel a myth that the Wall Street Journal bullhorns out in its article:
Taking on a mortgage to buy a house that could appreciate, or borrowing for a college degree that should boost earning power, can be wise decisions. Borrowing for everyday consumption or for assets such as cars that lose value makes it harder to save and invest in stocks and real estate that tend to create wealth. So the rise in consumer borrowing exacerbates the wealth gap.
A car loan is a payment for the services a car provides for you while you own it. It is not an investment for the future. Nobody - I dare say not even Wall Street Journal reporters - would buy a car for everyday use that is also an investment object. This is so common sense I should not have to point it out.

Likewise, the basic idea with a mortgage is to purchase "installments" of housing. Over time, though, the nature of a mortgage has changed: a home has become an investment object for middle class families who are unable to save sufficient amounts for their own retirement. One main reason for this is explained in Figure 2 above, namely the steady decline of wages and salaries as part of personal income.

The cause of this decline is not the suppression of wages that some pundits like to throw around (for the record, the Journal only hints at it) but is instead almost entirely attributable to the slow drift of the American economy into a state of economic stagnation. As GDP growth has slowed down, so has growth in personal income:

Figure 3
SourceBureau of Economic Analysis

The long-term problem with growth in household earnings is even more pronounced when we take taxes into account. The next figure explains the difference in real growth between personal income before and after taxes. Each column represents the average, annual difference for a decade.* For example, in the 1950s disposable income grew 0.07 percent more slowly per year than personal income before the deduction of taxes. While personal income grew at 3.93 percent per year, disposable personal income grew at 3.86 percent per year:

Figure 4
Source of raw dataBureau of Economic Analysis

Here, the interaction between household income and tax policy shines through:

1. In the 1950s and 1960s the differences in growth rate between total personal income and disposable personal income were moderate. There were no drastic changes to fiscal policy, except for the tax cuts during President Kennedy.
2. Things started to go haywire in the 1970s, when inflation conspired with a federal personal income tax that shoved income earners upward in the brackets without regard to their ability to maintain their standard of living. We saw a similar but less pronounced effect in the 1990s, when the Bush Sr. and Clinton administrations worked with Congress to roll back some of the Reagan tax cuts. This increased the wedge between total and disposable income, as visible in a bigger difference between growth rates.
3. The Reagan tax cuts definitely reduced the growth-rate difference from the 1970s. In terms of tax brackets alone, Reagan took us from fifteen to two; the top tax rate dropped from 50 percent to 28 percent. A similarly positive effect is visible in the Bush Jr. years of the 2000s. The bottom tax rate - applicable from the first dollar earned - fell from 15 percent to 10, while the top rate declined from 39.6 percent to 35.
4. Absurdly, the Great Recession and the slow Obama recovery allowed disposable income to move closer to total personal income. The reason was not in tax reforms, but in the fact that incomes generally grew more slowly, with more people losing income and falling into lower tax brackets.

Figure 4 teaches us two things about the role of ]debt in the finances of American consumers:

a) Taxes began eroding consumer finances already in the 1970s; and
b) If it had not been for the Reagan and Bush tax cuts, in the past 40 years we would have seen a notably larger growth wedge between total and disposable personal income.

Because of the invasion of taxes into household earnings, America's middle class has been forced to make their homes an investment. With that, mortgages have turned into a capital expense aimed at producing returns over the long term.

Then the Wall Street Journal turns on the income-distribution argument, again missing vital pieces of information that, when omitted, distort the content of the information they provide:
The median net worth of households in the middle 20% of income rose 4% in inflation-adjusted terms to $81,900 between 1989 and 2016, the latest available data. For households in the top 20%, median net worth more than doubled to $811,860. And for the top 1%, the increase was 178% to $11,206,000. Put differently, the value of assets for all U.S. households increased from 1989 through 2016 by an inflation-adjusted $58 trillion. A third of the gain—$19 trillion—went to the wealthiest 1%, according to a Journal analysis of Fed data.
The reason for this is fairly simple. As explained earlier, wages and salaries have declined as share of personal income. This decline is concentrated to higher-income layers, whose earnings from dividends, interest and capital gains have surpassed their work-based earnings. As a result, households in those higher income brackets have seen earnings grow faster than the average household.

The relative rise of equity-based income over work-based income has its explanation in - yes - tax policy. With taxes being lower on equity-based income, every $100 earned leaves a big difference in the taxpayer's wallet depending on how he earned the money. Consequently, higher-earning households have not only maintained a higher standard of living, but also had opportunities to invest themselves into higher wealth.

Nowhere in their article does the Wall Street Journal attempt to put differences in income, wealth and debt into the most obvious context: tax policy. This is a shame, because it leaves the reader of their article with the impression that consumers are drowning in debt and that the rich are running away with all the wealth. Neither of which is true, of course.

*) A decade in this particular analysis runs from the eighth year to the seventh year, e.g., the 1950s is counted as 1948-1957. The reason is availability of data that is not distorted by World War II and its immediate economic aftermath. It also aligns the beginning of the analytical period with its end in present time.

No comments:

Post a Comment