Robert Frank Falls For the Two-Income Trap:

In today's Washington Post, Robert Frank writes:

But while Congress clearly should not rescue borrowers who lied about their incomes or tried to get rich by flipping condos, such borrowers were at most a minor factor in this crisis. Primary responsibility rests squarely on regulators who permitted the liberal credit terms that created the housing bubble.

Hints of how things began to go awry appeared in "The Two-Income Trap," a 2003 book in which Elizabeth Warren and Amelia Warren Tyagi posed this intriguing question: Why could families easily meet their financial obligations in the 1950s and 1960s, when only one parent worked outside the home, yet have great difficulty today, when two-income families are the norm? The answer, they suggest, is that the second incomes fueled a bidding war for housing in better neighborhoods.

It's easy to see why. Even in the 1950s, one of the highest priorities of most parents was to send their children to the best possible schools. Because the labor market has grown more competitive, this goal now looms even larger. It is no surprise that two-income families would choose to spend much of their extra income on better education. And because the best schools are in the most expensive neighborhoods, the imperative was clear: To gain access to the best possible public school, you had to purchase the most expensive house you could afford.

But what works for any individual family does not work for society as a whole. The problem is that a "good" school is a relative concept: It is one that is better than other schools in the same area. When we all bid for houses in better school districts, we merely bid up the prices of those houses.

In the 1950s, as now, families tried to buy houses in the best school districts they could afford. But strict credit limits held the bidding in check. Lenders typically required down payments of 20 percent or more and would not issue loans for more than three times a borrower's annual income.

In a well-intentioned but ultimately misguided move to help more families enter the housing market, borrowing restrictions were relaxed during the intervening decades. Down payment requirements fell steadily, and in recent years, many houses were bought with no money down. Adjustable-rate mortgages and balloon payments further boosted families' ability to bid for housing.

The result was a painful dilemma for any family determined not to borrow beyond its means. No one would fault a middle-income family for aspiring to send its children to schools of at least average quality. (How could a family aspire to less?) But if a family stood by while others exploited more liberal credit terms, it would consign its children to below-average schools. Even financially conservative families might have reluctantly concluded that their best option was to borrow up.

Professor Frank has missed the ball on this one. As I observed in the Wall Street Journal last August in my column "The Two Income-Tax Trap" the math behind this simply doesn't add up--using Warren and Tyagi's own numbers:

The argument is developed in the book, "The Two Income Trap: Why Middle Class Mothers and Fathers are Going Broke," by Harvard Law School Professor Elizabeth Warren and her daughter Amelia Tyagi. In fact, using their own numbers, it is evident that they have overlooked the most important contributor to the purported household budget crunch — taxes.

Ms. Warren and Ms. Tyagi compare two middle-class families: an average family in the 1970s versus the 2000s (all dollar values are inflation-adjusted). The typical 1970s family is headed by a working father and a stay-at-home mother with two children. The father's income is $38,700, out of which came $5,310 in mortgage payments, $5,140 a year on car expenses, $1,030 on health insurance, and income taxes "which claim 24% of [the father's] income," leaving $17,834, or about $1,500 per month in "discretionary income" for all other expenses, such as food, clothing, utilities and savings.

The typical 2000s family has two working parents and a higher income of $67,800, an increase of 75% over the 1970s family. But their expenses have also risen: The mortgage payment increases to $9,000, the additional car raises the family obligation to $8,000, and more expensive health insurance premiums cost $1,650. A new expense of full-time daycare so the mother can work is estimated at $9,670. Mother's income bumps the family into a higher tax bracket, so that "the government takes 33% of the family's money." In the end, despite the dramatic increase in family income, the family is left with $17,045 in "discretionary income," less than the earlier generation.

The authors present no explanation for why they present only the tax data in their two examples as percentages instead of dollars. Nor do they ever present the actual dollar value for taxes anywhere in the book. So to conduct an "apples to apples" comparison of all expenses, I converted the tax obligations in the example from percentages to actual dollars.

In fact, for the typical 1970s family, paying 24% of its income in taxes works out to be $9,288. And for the 2000s family, paying 33% of its income is $22,374.

Although income only rose 75%, and expenditures for the mortgage, car and health insurance rose by even less than that, the tax bill increased by $13,086 — a whopping 140% increase. The percentage of family income dedicated to health insurance, mortgage and automobiles actually declined between the two periods.

During this period, the figures used by Ms. Warren and Ms. Tyagi indicate that annual mortgage obligations increased by $3,690, automobile obligations by $2,860 and health insurance payments by $620 (a total increase of $7,170). Those increases are not trivial — but they are swamped by the increase in tax obligations. To put this in perspective, the increase in tax obligations is over three times as large as the increase in the mortgage payments and almost double the increase in the mortgage and automobile payments combined. Even the new expenditure on child care is about a quarter less than the increase in taxes.

Overall, the typical family in the 2000s pays substantially more in taxes than the combined expenses of their mortgage, automobile and health insurance. And the change in the tax obligation between the two periods is substantially greater than the change in mortgage, automobile expenses and health-insurance costs combined.

Isolating just the mortgage burden, according to Warren & Tyagi's figures the percentage of family income dedicated to mortgage payments actually declined from the 1970s to 2000s, from 14% of household income ($5310 of $38,700) to 13% of household income ($9000 of $67,800). Again, this is using Warren & Tyagi's own figures.

I learned after writing the column that there is some dispute about the source of the rising tax burden. I followed Warren and Tyagi's lead from the book and attributed the growth in the tax burden primarily to the second-earner bias. Subsequently I found that there is some dispute about the extent to which the growth in the tax burden is attributable to income taxes rather than state and local taxes (including state income taxes). What has not been questioned is that Warren and Tyagi's own numbers show that the growth in the overall household tax burden overwhelms the growth in home mortgage expenses with respect to its impact on the household bottom line.

I like some of Bob Frank's work, especially back before he started writing his really political stuff. But I'm afraid on this one he fell into the two-income trap.


There is some confusion over whether Warren and Tyagi's analysis refers to marginal or average tax rates. It is average. In a long post on which the WSJ column was based I quoted the relevant excerpt from the book The Two-Income Trap (for those who don't have ready access to the book itself). As I noted in that post, the increase in taxes (in dollar terms) is more than three times greater than the increase in mortgage payments (the supposed driver of the "two-income trap"). I present the budget analysis graphically there as well.