How hot labor markets can lead to misleading median comparisons.

September 10th, 2019 at 3:07 pm

The Census income and earnings data sometimes have a confusing characteristic that is not uncommon in these sorts of data, especially in periods of tight labor markets, as was 2018. The issue has to do with changes in medians from one year to the next.

For example, the data that came out this morning showed that for both men and women full-time, full-year workers, real annual earnings rose 3.4 and 3.3 percent, respectively, 2017-18. But for all ft/fy workers, combining both genders, earnings fell 0.6 percent. The decline was statistically insignificant, but jeez–that’s confusing, right? Why would earnings fall, overall, in a year with a clearly solid job market, especially when both genders did pretty well?

A number have folks have asked me about this today–a similar dynamic is in the data for median household income: the real median went up  significantly last year for the two household types–family and non-family (individual) households–but not for all households (i.e., it rose slightly, by 0.9 percent, but the change was insignificant).

One way this often occurs, especially in hotter labor markets, is that the composition of workers or families change in ways that can make it hard to figure out what’s up and what’s down. Consider the first two columns below, arrayed from lowest to highest, assuming no change in the composition of people or HHs in the data. The median goes up in ways that we expect in positive economies.

No comp change                   Comp change

Year 1 Year 2 Year 1 Year 2
1 2 1 1
2 3 2 1
3 4 3 2
4 5 4 3
5 6 5 4
5
6

Numbers in bold italics are medians.

But now look at columns 3 and especially 4. What happens there, relative to column 2, is that two people with very low incomes or earnings enter the distribution. Imagine, for example, that these two “1’s” were sitting it out on the sidelines of the labor market, but got pulled in, hoping to take advantage of the tight conditions. This shifts everybody else up that year such that median slips from 4 to 3. A comparison of the two years would suggest no change at all in incomes and earnings. In fact, the number of ft/fy women grew more than twice that of men last year (0.7 million men vs. 1.6 million women).

At any rate, I suspect something like this is going on in these data.

The 2018 Poverty, Income, and health coverage results: a tale of three forces.

September 10th, 2019 at 12:51 pm

This morning, the Census Bureau released new data on health insurance coverage, poverty, and middle-class incomes. While the data are for last year, they shine an important light on key aspects of families’ living standards that we don’t get from the more up-to-date macro-indicators, like GDP and unemployment.

As the economic recovery that began over a decade ago persisted through 2018, poverty once again fell, by half-a-percentage point, from 12.3 percent to 11.8 percent. Other results from the report show that anti-poverty and income support programs lifted millions of people out of poverty, including 27 million through Social Security alone. Though the real median household income—the income of the household right in the middle of the income scale—increased slightly less than 1 percent last year, the increase was not statistically significant. Median earnings of full-time men and women workers both rose significantly, by over 3 percent for each (for reasons discussed below, sometimes earnings rise significantly but income does not).

Health coverage, however, significantly deteriorated last year, as the share of the uninsured rose for the first time since 2009, from 7.9 percent to 8.5 percent. In total, 27.5 million lacked coverage in 2018, an increase of 1.9 million over 2017. This result is partially driven by actions of the Trump administration to undermine the Affordable Care Act (note that Medicaid coverage was down by 0.7 percentage points), and in this regard, it should be taken as a powerful signal of the impact of conservative policy on U.S. health coverage.

Taken together, the poverty, income, and health coverage results tell a tale of three powerful forces: the strong economy, effective anti-poverty programs, and the Trump administration’s ongoing attack on affordable health coverage. A strong labor market is an essential asset for working-age families, and the data are clear that poor people respond to the opportunities associated with a labor market closing in on full employment. Anti-poverty programs are lifting millions of economically vulnerable persons, including seniors and children, out of poverty. But while a strong labor market and a responsive safety net help to solve a lot of problems, the history of both U.S. and other countries shows that it takes national health care policy to ensure families have access to affordable coverage. The ACA was and is playing that role, but efforts to undermine its effectiveness are evident in the Census data.

Poverty, Income, Inequality

The Census provides two measures of poverty: the official poverty measure (OPM) and the Supplement Poverty Measure (SPM). The latter is a more accurate metric as it uses an updated and more realistic income threshold to determine poverty status, and it counts important benefits that the OPM leaves out. While the two measures often track each other, year-to-year, that wasn’t the case last year, as the SPM rose an insignificant one-tenth of a percent, from 13.0 to 13.1 percent, while the OPM fell a significant half-a-percent, from 12.3 to 11.8 percent. Because the SPM has a higher income threshold than the OPM, 4.4 million more people were poor by that more accurate measure.

Because it counts anti-poverty policies that the official measure leaves out, one particularly useful characteristic of the SPM data is that it breaks out the millions of people lifted out of poverty by specific anti-poverty programs. For example, refundable tax credits, such as the Earned Income Tax Credit and the Child Tax Credit lifted about 8 million people out of poverty in 2018; SNAP (food stamps) lifted 3 million more out each, and Social Security was the most powerful poverty reducer, lifting 27 million out of poverty in 2018, 18 million of whom were elderly (65 and older).

As noted, median household income, inflation-adjusted, rose less than a percent last year, a statistically insignificant change (meaning a change that is statistically indistinguishable from no change at all). Yet, real median earnings of full-time, full-year workers rose more than 3 percent for both men and women. It is hard to square these results, but they are not that unusual and probably have something to do with the changing composition of households and the fact that the median male worker is different from the median female worker and neither are necessarily in the median household. Note, for example, that family households (basically, two or more related people) and non-family households (people living alone) both rose significantly last year. But when the Census smushes them together, we get an insignificant increase.

I conclude from this and other information in the report, like the fact that the number of full-year workers rose 2.3 million, or the evidence showing real wage gains last year for middle and low-wage worker, that the strong labor market helped to boost family incomes in 2018 (though as I show below, these gains are slowing over time). Another key factor pushing up wage growth at the low end of the pay scale were the minimum wage hikes that occurred in 18 states in 2018, affecting 4.5 million workers, according to EPI.

Here’s one way to look at this relationship between labor markets and, in this case, poverty outcomes. It’s a scatterplot of unemployment against the change in poverty rates (using the OPM for which we have a long, consistent time series). It shows how low unemployment correlates with declines in the poverty rate and vice-versa. Why? Because able-bodied, poor people respond to tight labor markets, an important fact that pushes back on the alleged need for work requirements.

Sources: Census, BLS

Unfortunately, over the past few decades, labor markets have not consistently provided the job and earnings opportunities that help to support income growth for families in the bottom half of the income scale and longer-term comparisons show real median income not too far above its pre-recession peaks in 2000 and 2007. Moreover, as inequality has increased, we cannot blithely extrapolate from positive macro-indicators, like unemployment and GDP, to indicators like poverty and median income that will often reflect less improvement in periods when growth disproportionately accrues higher up the income and wealth scale. Though these Census data are less comprehensive than some other sources of inequality data, they do show that in 2018, the highest fifth of households held more income (52 percent of it) than the bottom 80 percent. Though, as noted, the survey has changed over the years such that long-term comparisons should be made with care, in 1967, this share was 44 percent, meaning the bottom 80 percent controlled more income than the top fifth. This increase in inequality is solidly confirmed in much other data.

The table below brings the critical dimension of race into the analysis (note: none of the income changes shown for 2018 are statistically significant). Median household income growth was slower in 2018 relative to earlier years, particularly for Hispanic families. Note also how poverty rates for blacks and Hispanics are multiples of those of whites. The scatterplot shows that lower unemployment correlates with lower poverty, and the table shows this effect to be greater for non-whites, who, over this period, experienced larger declines in unemployment accompanied by bigger drops in poverty. For example, over this period both white unemployment and poverty fell about 1 percentage point. For blacks, the comparable declines are 3 points for both variables. Hispanic poverty was down almost 4 percentage points.

Sources: Census, BLS.

Health Coverage

As noted, as soon as the ACA passed, the expansion of Medicaid coverage and premium subsidies through the exchanges quickly reduced the share of people without coverage. The discussion above—the one noting the increase in the uninsured rate—focused on the main national survey featured by the Census today (the ASEC). But due to its many discontinuities, to compare changes over time it is better to use the other survey results released by Census today, from the American Community Survey (ACS).

This figure clearly shows the historical coverage gains made by the ACA, but it also shows those gains fading in 2017 and this year, in 2018 (the 0.2 point increase in the uninsured rate last year is statistically significant).

Source: ACS

In recent years, gridlock, dysfunction, government shutdowns, and the general unwillingness of Congress to deal with our fundamental challenges has led to a justified skepticism of our federal system. But it’s worth remembering that not too far back, this system passed and implemented the largest and most consequential change in national health policy since the advent of Medicaid and Medicare in the 1960s. And the results, in terms of increased coverage, were equally dramatic.

This insight makes today’s health coverage results extremely concerning, as they reveal the impact of policies to reverse those gains. This attack on affordable coverage, according to my CBPP colleagues, “began on President Trump’s first day in office, with an executive order calling on federal agencies to waive and delay ACA provisions “to the maximum extent permitted by law.”’ They include repealing the individual mandate, anti-immigrant measures that are likely leading immigrants to avoid publicly-provided coverage, cuts in ACA outreach and enrollment assistance, work requirements that hassle people off of the Medicaid rolls, and a wide variety of waivers and eligibility barriers designed to shrink public coverage and shift medical costs onto consumers.

What’s it all mean?

The Census report is a tale of three powerful forces. First, the momentum from the strong economy continues to boost work and wages for low- and middle-income people. Second, anti-poverty programs are reliably helping to lift millions out of poverty. Third, such gains can be reversed by policies hostile to them. It is thus extremely worrisome to consider actions the Trump administration is taking to reduce government support of poor households, especially those with immigrants. Such actions include work requirements that ramp-up administrative demand to hassle low-income people off of Medicaid and SNAP; the “public charge” changes that threaten to block legal immigrants from seeking support they and their children need, changes in poverty measurement designed to make it look like fewer people are poor (and thus reduce their eligibility for assistance), and changes to nutritional support also designed to kick currently eligible persons off the roles.

The economy and complementary work supports are helping many low- and moderate income get ahead. Significant gaps persist, especially with regard to race. But the underlying trends of poverty and income have been favorable. Health coverage tells a different story and we must be vigilant not to let these same political forces do to anti-poverty programs what they’re doing to health programs.

Payrolls slow and the trade war is hurting manufacturing. But underlying job market still solid.

September 6th, 2019 at 9:58 am

Payrolls rose by 130,000 last month and the unemployment rate held at 3.7 percent, close to a 50-year low and the same level as the past 3 months. Still, job growth is cooling (25,000 of this month’s gains were temporary decennial Census workers), as the pace of monthly gains, while still strong enough to support low unemployment, has slowed. Wage growth also stayed parked at about where it has been in recent months, and there’s some evidence that the trade war is taking a toll on factory jobs. However, the job market remains strong, real wages are growing, and consumer spending will continue to be supported by these dynamics.

The slowdown in payrolls

To get a clearer take on the underlying trend in job growth, our monthly smoother shows the average monthly gain over 3, 6, and 12-month periods. This month, however, we add an extra bar to our usual smoother, as we believe it is important to begin to incorporate a recent BLS revision, based on more accurate jobs data, into our assessment of the US job market. This preliminary benchmark revision estimates that employers added 500,000 fewer jobs to US payrolls between April of 2018 and March of 2019 (BLS will officially wedge their final estimate into the payroll data by Feb 2020). The second bar includes the result of this revision, showing that over the past year, payroll growth was likely closer to 150K per month than 175K per month.

To be sure, this is still solid payroll growth at this stage of the expansion and as noted below, in tandem with real wage growth, it’s strong enough job growth to support the recovery and keep unemployment around where it is. However, using the preliminary revised data, the pace of payroll gains has slowed from 1.6% last year to 1.3% this year. Clearly, that’s not a big deceleration, and it’s also not unexpected in a job market closing in on full employment. But it is a slower trend which I expect to persist.

The trade war

The trade war that the Trump administration has been waging is clearly taking a toll on the global economy. While its impact is greater in countries more exposed to trade, like Germany, than the US, our manufacturers have been hit by these new taxes (tariffs) on their imported inputs and by retaliatory tariffs on their exports. To what extent is this showing up in factory employment, hours, and wages?

Manufacturing employment has slowed since the Trump administration began ramping up tariffs at the beginning of last year. Last month, factory jobs rose just 3K and durable manufacturing employment was unchanged. Thus far this year, the factory sector has added 5.5K jobs per month on average, compared to 22K for all of last year.

The product of manufacturing employment and weekly hours yields the aggregate hour index for the sector, a very good proxy for labor demand. The next figure looks at the year-over-year change in this index for blue collar and for all manufacturing workers. Starting about a year ago, a clear deceleration is evident, and for the non-managers—who comprise about 70 percent of the sector’s employment—total hours worked have outright declined in recent months (relative to a year ago).

After slowing in 2018, manufacturing wages for blue-collar workers have picked up pace in recent months and are now growing at about the same rate of other mid-level workers.

In sum, at least in terms of jobs and hours, the trade war is hurting manufacturing workers. I’m sure some will push back that this near-term pain is worth the longer-term gains from a “victory” in the trade war. I find this totally unconvincing, as victory apparently means getting China to be more accommodating to US multinationals. That is, were China to stop insisting on tech transfers, or issue more licenses to our multinationals, we’ll get more, not less, offshoring of US jobs.

Wages still stalled

Wage gains are still stalled, though at a level above inflation, so real paychecks are growing on average (see third figure below). The stalling is clear in the 6-months rolling average, and is not particularly surprising as the job market has not particularly tightened further over this period. That is, low unemployment is providing workers with more bargaining clout than they’d have in less tight job markets, but this force appears to be holding steady for now.

Stronger Household Survey

Participation ticked up and the closely watched employment rate for prime-age workers (25-54) hit a cyclical high of 80%, just 0.3 ppts below its 2007 peak. While we can’t say much about one month’s change, this important measure of core labor market capacity had previously been stalled. If it continues to rise, it will suggest there’s still more room-to-run in the job market, and especially given low inflation, a strong rationale for the Federal Reserve to do what they can to extend the run.

Bottom line, the job market will handily support consumer spending in the near term, staving off any recessionary threats from the trade war and the global slowing to which it has contributed. However, payroll gains have slowed somewhat, especially in manufacturing, and, I suspect, in any other sectors with global connections (i.e., tradeable goods and services). We will continue to monitor this and any other fragilities related to the trade war or whatever other unforced policy errors are forthcoming.

The NYT wrote a woefully imbalanced piece on Opportunity Zones.

September 3rd, 2019 at 9:30 am

A number of people (OK, four…but it’s early) have asked me to respond to the NYT piece from last Sunday on how the Opportunity Zone tax break is nothing but a boon to the rich. As I’ve written in a few opeds, I’ve been a cautious supporter of the program, though I’ve been careful to make the points that a) it’s too early to say much about outcomes, and b) while OZs have the potential to become a wasteful tax shelter mechanism, some early signs are hopeful. And, as the Times points out, some early signs are not.

The problem is, the piece was a list posing as an analysis. It just lists many examples of rich people getting the tax break through the program without a shred of evidence that poor people and places aren’t getting helped. That’s largely because, as noted, it’s simply too early to make this foundational assessment, which is why it’s too early to conclude that OZs are failing to have their intended effect.

In essence, the piece makes two points, neither of which should surprise anyone: rich people have capital gains, and rich investors are taking advantage of the OZ tax break. It then cherry picks a bunch of cases that look bad, where Opportunity Funds are supporting the building of luxury dwellings that would have been financed without the tax break.

Here’s a good example of what’s wrong with the article. In a typical example of why OZ’s don’t work, the Times writes:

Many others [taking advantage of the program] are lesser-known business executives who recently sold small companies or real estate and are looking for ways to avoid large tax bills.

Paul DeMoret, for example, recently sold his auto-industry software company in Oregon. He said he was using some of those capital gains to help finance a Courtyard by Marriott in Winston-Salem, N.C., and an apartment building in Tempe, Ariz., among other projects in opportunity zones. He is making the investments through a private equity firm, Virtua Partners.

This is not anywhere close to evidence of the article’s thesis that OZs are not having their intended impact. To the contrary, it’s showing the part two of the plan is working: investors are tapping the tax break to invest in the zones (part one was picking the zones; as I and others have argued, and the Times agrees, most—not all—of the zones were well chosen). Apparently, the reader is supposed to assume that building a hotel and an apartment building won’t help folks left behind. But without data on employment outcomes, which can’t possibly exist yet, there is no way of knowing this outcome. I could, based on the facts that hotels employ low-income workers who also often live in “apartment buildings” declare success! But that would make no more sense than declaring failure based on this anecdote.

Same with the line of argument that goes: X is a known, greedy jerk. X invests in OZs. Thus, OZs won’t help the poor. Again, not exactly the trenchant analysis we’re looking for from the paper of record. The point of the program is to incent patient capital investment in places that face historical disinvestment. There’s no requirement that the investors are good guys and gals.

And guess what? As Steve Glickman, one of the early designers of the program, points out in this Twitter thread, there are lots of great people—I mean serious non-jerks—who are investing in OZs. But you’ll learn almost nothing about them is this unbalanced piece.

I myself spoke to the authors of the piece for hours about such nuances, and earnestly gave what I believe was a balanced assessment of the program’s promise and risks. I’m perfectly willing to admit that my message—on-the-one-hand-this-on-the-other-hand-that—along with my strong assertion that it’s just too soon to tell, wasn’t definitive, sexy, or even that interesting. But it’s just irresponsible journalism to leave out informed voices (I was co-author, with Kevin Hassett, of the white paper that first introduced the idea) that don’t fit the authors’ slant.

A few other rants:

–The piece quotes someone as saying “Perhaps 95 percent of this is doing no good for people we care about.” There’s simply no possible way to know this, much less put a number on it. I don’t see how a supposed “fact” like that gets by an editor (and I hope the Times doesn’t think “perhaps” makes it okay to run false numbers). Ross Baird, who also has an excellent thread on the NYT piece, had the same reaction, and provides useful background on where the 95 percent comes from.

–A good chunk of what the piece bemoans is actually about building in mixed-income communities. Many of us view this as a potentially positive outcome of OZs. There’s solid social-science research showing the benefits to the poor, especially for young children, of growing up in such places relative to high poverty areas.

–Baird also drills down into just what early days these are re OZs: “The narrative that a “wave of developments” is happening is not yet true. Opportunity Zone funds across the country have raised less than 10% of their goals. It is a new market and most people are very cautious.” In fact, to my knowledge, no OZ project is up and running such that we can evaluate the outcomes on the variables about which I care most: jobs, poverty, incomes.

–My biggest concern about an imbalanced piece like this at this early point in the evolution of the program is its opportunity costs. Yes, we need to identify and stop wasteful projects, like some of those identified on the piece. But a more balanced take would have asked what else needs to be done to make sure the program has its intended effect. Most important, in this regard, is what Kenan Fikri of EIG notes in yet another useful thread reacting to the Times: “20 months have elapsed since passage; it’s time to get the data reporting regime in place.”

–The political framing of the piece is off. As Glickman notes: “Referring to this program as a Trump tax break is in itself disingenuous. The #OpportunityZones legislation was cosponsored by nearly 100 Members of Congress, proportionally divided between GOP & Dems, including several Dem candidates for President.” Sen. Booker (D-NJ) was the leading co-sponsor and is now pushing important legislation—again, bipartisan—to get the data we’ll need to make the evaluation that was lacking in the Times piece.

Let me, for the n_th time, be unequivocal about my own position, not because my view particularly matters, but because this is the view of other progressives who share my take. OZs pose promise and risk. At this point, one could write a piece featuring promising projects, as I recently did (while emphasizing the risks) as easily as the opposite tack taken by the Times. I’ve also been careful to cite the nuanced work by my CBPP colleagues who have legitimate concerns that Opportunity Funds could become wasteful tax shelters.

OZ advocates and close observers, including Glickman, EIG’s John Lettieri (cited in the NYT piece), Fikri, Baird, have been quick to point to developments that are counter to the intention of the law. The reason we do so is simple: we want this thing to work! I’ve been working in anti-poverty policy for over 30 years, and I have absolutely zero interest in a wasteful tax break for rich people. To the contrary, since Reagan, I’ve been one of the most outspoken critics of trickle-down tax cuts. But I and others with similar backgrounds (e.g., Bruce Katz) see potential in OZs.

Whether we realize that potential is the huge, outstanding question, currently unanswerable. The Times made a big mistake by assuming the answer is in and the program’s a flop. That’s wrong, and I and others will continue to do our best to make sure it stays wrong.

Recession Readiness and State UI Trust Funds

August 22nd, 2019 at 5:25 pm

[My colleague Kathleen Bryant took the lead on this piece–JB]

Given the recent dramatic spike in media coverage of our economic headwinds and recession readiness over the past week, we decided to take a closer look at the balance sheets of state unemployment insurance (UI) trust funds. While the Department of Labor (DOL) is responsible for overseeing the UI system and paying administrative costs, the basic program is managed and mostly funded by the states. Using the most recent final data available from the Treasury Department, we analyzed the number of state UI trust funds that meet DOL’s recommended minimum solvency standard. This standard is measured using a ratio called the  “Average High Cost Multiple,” where a value of 1 means that trust fund reserves could pay out at least 1 year of benefits during a recession of average depth– states with an AHCM greater than 1 have met DOL’s recommended minimum solvency level.

There are 18 states that have not met DOL’s minimum solvency standard (as of July 2019), including some of the most densely populated states in the country– California, Texas, and New York. Congress should be closely monitoring the balance sheets of state UI trust funds and should be prepared to ramp up federal spending on UI when the next recession hits, considering the financial status of many state trust funds.

Source: The Department of Labor, the Department of Treasury