Past as Prologue:
Revised Histories and Extraordinary
In the New York City Economy
PDF version available
Forecasting the city’s economy has a number of inherent challenges. These challenges have grown in recent years as the annual revisions to employment and labor force data have been large and hard to anticipate. Getting a handle on the past has been further complicated by some unusual recent trends in local employment.
IBO has taken a close look at the latest revisions and examined some underlying trends in New York City jobs data. Among our observations:
From 2010 through 2017, the city averaged close to 100,000 new jobs a year. Monthly data released during 2018 indicated a slowdown to about 71,500 jobs added over the year. But the March 2019 revision effectively eliminated the slowdown, retroactively raising job growth to nearly 93,000 in 2018.
In the revised numbers, nearly two out of three private jobs added in New York City in 2018 were due to extraordinary employment surges in a few sectors—employment services, individual and family services, and, especially, home health care services.
Last year, more home health care service sector jobs were added in New York City (27,100) than in the entire rest of the nation (24,300). Still more home care-related employment swelled the city’s individual and family services payrolls.
One of the main drivers of the home health employment surge has been the Consumer Directed Personal Assistance Program, which enables eligible seniors and the disabled to choose their own home-care providers, including family members, who are paid through Medicaid. In 2018 there were almost 100,000 providers employed through the personal assistance program in New York City, two-thirds in home health care services and the remainder mostly in individual and family services. Impending changes in state oversight of this program make it difficult to predict how long it will sustain its rapid expansion.
Labor force data (New York City resident adult population, employment, and unemployment) are also revised every March. Over the past few years these revisions have been large. The latest March revision sharply lower city population and resident employment growth since 2013.
Data from previous years is an essential component of forecasting the local economy for the upcoming months and years. When statistical portraits of the past are revised substantively from one year to the next—and reveal outsized impacts from extraordinary circumstances—it adds more uncertainty to forecasters projections of the city’s economic future.
| CLARIFICATION: This summary was revised on May 16, 2019 to clarify the explanation of key findings in employment trends.
Prediciting the Past
There is a saying, variously attributed, that “it’s difficult to make predictions, especially about the future.” There are times when it is even difficult to predict the past—especially in the world of economic data that are subject to periodic revision. This is one of those times, thanks to recent major reworkings of both the payroll employment and labor force histories in New York City.
These changes have yielded somewhat contradictory portraits of the city’s recent economic past, but they are not the only factors complicating predictions about the future. Highly unusual recent trends within city employment are compounding our challenges.
Through the first eight years of the current expansion (2010 through 2017), New York City averaged growth of close to 100,000 jobs per year. But as the city’s monthly employment numbers were released by the federal Bureau of Labor Statistics (BLS) over the course of 2018, the new stats depicted what appeared to be a substantial slowdown in the local economy. Those numbers were based on employer survey data, however, and each March BLS revises prior years’ figures to bring them more in line with actual headcounts from the employer census based on unemployment insurance statistics, a procedure called benchmarking.
In recent years benchmarks have yielded large revisions to the city’s employment growth history, and the benchmark released this March was no exception. The initial survey data showed 71,500 jobs added in 2018 (column 6 in the table on below), with more than half of the growth packed into the fourth quarter. The revised numbers show 92,900 jobs added (column 7), with the gains distributed fairly evenly across all four quarters of the year.1 The 2018 slowdown appears to have been revised out of existence.
But on closer inspection, a few exceptional factors had an outsized impact on recent city employment growth. First, there was a jump in local government headcount in New York City resulting from a change in the source for employment data. Starting in 2018, the Quarterly Census of Employment and Wages is basing its count of city government employment on data from the city’s own automated payroll system. Prior to that, the employer census based its tabulations on data from the state unemployment insurance system—and evidently the latter did not fully capture city government headcount.
The new benchmark adopted that change with a retroactive phase-in of the revised headcount. This boosted seasonally adjusted government payrolls in New York City by 33,400 as of the last quarter of 2017 and by 41,600 as of the last quarter of 2018, thus adding about 8,300 to last year’s employment growth. The change transformed what had been a 2018 loss of 400 jobs in government under the old benchmark into an apparent 2018 gain of 7,900 jobs in government under the new benchmark. But the gain was a statistical artifact: those "new" employees were already there on the city’s payroll—just not counted by the Bureau of Labor Statistics. (The phased-in impacts on government employment growth in prior years were much smaller, generally running about 1,700 per year.)
The second big factor fueling employment growth in New York City is an extraordinary surge in home health-related payrolls. This has turbocharged employment growth not only in the home health care services sector (which is part of ambulatory care), but also within the individual and family services sector (which is part of social assistance). Recent benchmark revisions have revealed just how extraordinary this surge is. First, the March 2018 revision raised combined 2017 growth in home health care and individual and family services from 16,800 to 29,200. Then, the March 2019 revision raised combined 2018 growth in the two sectors to 39,700.
Looking at just the home health care services sector, which has been expanding rapidly in the city since the turn of the millennium, employment in this sector grew from less than 30,000 in 1999 to over 140,000 in 2016. But growth has only accelerated since then, with more than 50,000 home health care services jobs added just in the past two years alone.
Indeed, home health care services generated nearly a quarter of the city’s private sector employment gains in 2017, and almost a third of all private sector gains in 2018. Indeed, home health care services generated nearly a quarter of the city’s private sector employment gains in 2017, and almost a third of all private sector gains in 2018.
This explosion in home care services employment far outpaces the growth in New York City’s elderly population—and appears to be unique to the city. Indeed, in 2018 more home health care jobs were added in New York City (27,100) than in the entire rest of the United States (24,300).
Not all of these new home care jobs were in traditional home health care agencies. A contributing factor to the upsurge in city home care employment has been the rapid expansion of the Consumer Directed Personal Assistance Program (CDPAP) in New York. This program allows eligible elderly and disabled people to choose their own in-home care providers—including family members—with payroll and administrative services supplied by “fiscal intermediaries” (FIs) who function as co-employers of the providers and are reimbursed (through the state) by Medicaid.
CDPAP uptake appears to have been particularly strong in New York City, where employment at fiscal intermediaries grew from 29,300 in the first quarter of 2010 to 71,500 in the fourth quarter of 2016, and then swelled further to 97,500 by the second quarter of 2018.2
Nearly two-thirds of FI payroll is classified under home health care services (63,900 as of the second quarter of 2018). Fiscal intermediaries account for 37 percent of total employment in this sector, and the share has been growing, for FI employment in home health services has more than tripled since the first quarter of 2010, while all other home health services employment “merely” doubled.
Most of the remainder of FI payroll is classified under individual and family services (28,500). Fiscal intermediaries account for 20 percent of total individual and family services employment—and have been responsible for virtually all of the employment growth in this sector since the first quarter of 2010.3
Individual and family services itself generated an unusually large share of overall city employment growth in 2017 and 2018. Individual and family services employment grew by 5,900 in 2017, the biggest jump in 25 years, and then by another 12,600 in 2018, an all-time high. CDPAP fiscal intermediaries have driven nearly all of the recent surge in this sector’s payrolls.(For a detailed table presenting the recent revisions to Current Employment Statistics data and the growing importance of fiscal intermediaries for these sectors over the years 2010-2018, click here.)
Another exceptionally large share of recent New York City job growth came from employment services. This sector gained 9,600 jobs in 2017, the largest increase in almost 20 years, and then topped that with a record 15,400 jobs added in 2018. Virtually all of the growth appears to have been generated by professional employer organizations (PEOs). PEOs in fact perform much the same function as fiscal intermediaries, providing human resource management services for client businesses and households and serving as the co-employer of clients’ employees for payroll administration and benefits purposes.
Here again, we are witnessing a decidedly city-centric phenomenon: indeed, while employment services PEO payrolls in New York City nearly doubled from the last quarter of 2016 (20,300) to the second quarter of 2018 (39,600), the payrolls of such establishments in the rest of the United States actually declined over this period. The recent city PEO payroll growth has been mostly a function of expanding services to professional and technical companies4
Altogether, home health care services, individual and family services, and employment services accounted for less than a tenth of all private employment in New York City as of the end of 2016, but were responsible for 2 out of every 5 private jobs added in 2017 and for nearly 2 out of 3 private jobs added in 2018. When we look past the extraordinary trends in these sectors (and the change in the counting of local government workers), it is evident that the bulk of the city economy did slow markedly last year, with all other private sector job growth sliding from 57,300 (1.7 percent) in 2017 to just 29,900 (0.9 percent) in 2018.
Implications for Forecasting. The outsized impact of an unusual benchmark revision and exceptional industry trends on city employment in 2018 presents considerable challenges for forecasting employment growth in 2019 and beyond.
All else being equal, we would expect some slowing this year because the switch to the new source of government payroll data is complete and will not artificially inflate growth in 2019 as it did in 2018. This by itself translates into about 8,000 fewer jobs added to city employment in 2019 than were added in 2018.
But what of the nearly 40,000 jobs added in home care and individual and family services in 2018? Can we project a repeat this year (or next)? So far there appears to be little sign of letup, with employment up by 5,900 in home care services and by 2,800 in individual and family services in the first quarter of 2019. But because the recent local trends have been so unusual, the range of possibilities over the longer term is remarkably wide for these sectors.
The real wild card here is the Consumer Directed Personal Assistance Program. The new state budget includes a major revision in the way that fiscal intermediaries will be regulated and reimbursed, with the aim of realizing economies of scale in the administration of the program. According to the state, the consolidation of fiscal intermediary services will not affect CDPAP eligibility or hours of care provided.5 But it remains to be seen whether home health and family and individual services payrolls will sustain their extraordinary growth trajectories under the new regulatory regime.
It will also be hard to predict how far PEOs continue driving rapid growth in employment services in New York City—and almost uniquely in New York City. Overall employment services growth did slow in the first quarter of 2019. But this is another sector where national trends serve as no guide to what has been happening at the city level.
In sum, the fate of a large portion of New York City’s ostensibly strong economic momentum going into 2019 appears to rest with an assortment of exceptional circumstances.
The Local Area Unemployment Statistics (LAUS) program of BLS produces monthly data on the employment status (employed, unemployed, or not in the labor force) of the civilian noninstitional population. These labor force data are derived from household surveys and measure employment status by place of residence—in contrast to the payroll employment data, discussed above, that are based on employer surveys and censuses and measure employment by place of work.
As it does for the payroll employment data, BLS also retroactively revises its labor force numbers every March to include the latest information on population growth and incorporate any changes in methodology. Often there are oscillations in the prior year’s employment and unemployment data that are smoothed out in the new revision. Over the past few years, however, the data prior to revision have been extremely volatile and the scale of the revisions exceptionally large.
Thus, after the labor force history was revised for 2015 and earlier at the beginning of 2016 (dark purple line on the chart on below), reported resident employment first rose, then plunged—falling by 129,000 in five months—and then rose again, resulting in a sharp net decline over the year (light purple line). This was all eliminated when the labor force history was revised again at the beginning of 2017, and previous years’ growth was brought down and smoothed out (solid green line).
But following that revision, reported resident employment shot up through the first half of 2017—if the numbers were to be believed, city resident employment increased by 150,000 in six months—and ended the year with a sharp net increase (light green line). This too was all eliminated when the history was revised at the beginning of 2018 (dark red line), and previous years’ growth was smoothed upwards.
After this, the labor force data were relatively “well-behaved”—no large gyrations—over the course of 2018 (lighter red line). Nonetheless, there are once again substantial changes to the history in the latest revision (dak blue line). The revised history shows a sharp downward adjustment in resident employment in January 2014 followed by much slower growth in this series than estimated the year before.
The new revision reflects a general downward recalibration of city population growth since the 2010 Census. This shows up in lower civilian adult population estimates starting in April 2010; indeed, according to the revised series, the city’s adult population peaked in July 2016 and has been declining ever since. As of December 2018, New York City’s civilian adult population was 182,000 lower than previously estimated, household employment was 103,200 lower, the number of unemployed was 3,900 lower, and the number of adults not in the labor force was 74,900 lower.
These adjustments had little effect on the city’s year-end unemployment rate, which was unchanged at 4.0 percent in December. The year-end employment/population ratio (58.4 percent) and labor force participation rate (60.9 percent) were also virtually unchanged.
The resident employment count, however, dropped by 16,000 over the first three months of 2019 (light blue line), while the number of unemployed jumped by 10,600. This has pushed the city unemployment rate up to 4.3 percent as of March. Pending other indicators that the city economy is actually contracting, these sharp changes may instead signal that further substantial labor force revisions will be in the offing.
Two Trends in Employment
Payroll employment (the job count from surveys of firms) generally surpasses resident employment (employment status from surveys of households) in New York City, as expected given the significant share of commuters in the city workforce. Still, both employment series broadly reflect the same economic activity, so questions are raised when they do not move in step. One consequence of the revisions of employer and labor force data is that the trends in these series now diverge even more sharply than the earlier data showed.
The newly benchmarked payroll data (dark blue line in chart below) show New York City employers adding 921,000 jobs since the September 2009 trough. Over the same period, however, the revised household data (dark red line) show a rise of just 373,500 in the number of New York City residents employed. (These numbers are through March 2019.) The contrast over the past three-plus years is even more stark: 285,000 jobs added by city employers since the end of 2015, but only 40,100 more residents employed.6
Divergences between employer trends and labor force trends are not new or unusual. For example, from January 2001 to that September 2009 trough, city payroll employment fell by 71,000 while resident employment grew by 75,000. And from January 1990 to January 2001, payroll employment was up 144,000 but resident employment rose 377,000.
Attempts have been made to reconcile these series, focusing on three factors. First, payroll employment would grow faster than resident employment if commuters picked up a disproportionate share of the new jobs created in an expansion. Second, payroll employment would grow faster if residents took on more secondary employment in an expansion (if a working resident takes a second job, that still counts as one person employed in the household survey, but the payroll count increases by one in the employer survey). Third, payroll employment would grow faster if residents shifted from self-employment to working for an employer in an expansion (self-employment is included in the household survey, but not in the employer survey).
One challenge for such explanations is that payroll employment does not always outpace resident employment in an expansion; sometimes (as we have just seen), resident employment grows faster. In any case, an IBO study—based on data from 2011, a year in which the gap between payroll employment and household employment was smaller than it is today—found that “none of the three explanations explain more than a small share of the discrepancy between payroll and resident employment gains.”7
Over the very long run the divergences between payroll and resident employment trends tend to even out. But the current divergence between the two series is particularly large. It is possible that subsequent data revisions (most likely on the resident employment side) may narrow the gap. Meanwhile, it adds yet another layer of uncertainty to our efforts to predict the past of the city economy, and from there predict the future.
Prepared by David Belkin
with Cole Rakow and Melinda Elias
1Employment growth is measured on a seasonally adjusted quarter-to-quarter basis. Annual growth is the sum of four quarters of growth, which is the equivalent of current year fourth quarter employment less the previous year’s fourth quarter employment (Q4 over Q4). Except where noted, seasonal adjustments are done by IBO. IBO obtains seasonally adjusted total employment for New York City from the sums of seasonally adjusted industry employment. This total almost exactly matches the seasonally adjusted total nonfarm employment series generated by BLS. (BLS does not seasonally adjust at the local industry level.)
2All FI numbers are from the Quarterly Census of Employment and Wages and are seasonally adjusted by IBO. They are based mostly on a list of fiscal intermediaries registered with the state Department of Health (DOH), along with CDPAP venders listed by the city’s Human Resources Administration (and not on the state DOH list), as well as a few FI’s whose names came up as members of the trade association, the Consumer Directed Personal Assistance Association of New York State. This is very likely not a complete count of fiscal intermediaries, not least because the published state DOH list has not been updated since February 2018.
3There were also 2,600 employed or co-employed by fiscal intermediaries classified under business services, and another 2,000 employed or co-employed by FIs classified under other categories of social assistance.
4There is additional payroll employment (upwards of 22,000 as of the second quarter of 2018) associated with New York City PEOs that are themselves classified as professional, scientific, and technical service establishments. All PEO numbers are seasonally adjusted.
5See https://www.health.ny.gov/health_care/medicaid/redesign/2019/docs/2019-01-23_exbudget_briefing.pdf, p. 9.
6Even excluding home health services, individual and family services, and employment services, city employers added 155,200 jobs since December 2015; city payroll employment grew almost four times faster than city resident employment.
7Julie Anna M. Golebiewski, A Matter of Data? Unraveling the Discrepancy Between City Job Growth & A High Unemployment Rate (IBO Fiscal Brief, February 2013, https://ibo.nyc.ny.us/iboreports/febacsemployment2013.pdf).
PDF version available here.
notification of free reports by e-mail