Article and illustration by Mark Beauchamp
Yesterday on his blog, Michael Mandel asked why jobs in the financial industry have fared well — especially considering the upheavals the industry has gone through. Between 2007 and 2010, the percentage employment losses in construction and manufacturing have been in the double digits, yet sectors like finance & insurance and commercial banking stayed (easily) under 10%.
We thought it was was good question — after all, we just went through the biggest financial upheaval in a generation. The trouble (borne out by the story below) is that financial jobs stats are like the financial markets — complex, volatile, and opaque. Nuance and detail is very important, but the data applied to the question can change employment estimates for the financial sector by millions of jobs.
Here is Mandel’s take on why financial jobs looked like they were doing well:
My theory is that as long as the U.S. is running a big trade deficit, financial sector jobs are going to do very well. The rest of the world has to lend large amounts of money to the U.S. to keep the global economy going, and all of that money has to be funneled through Wall Street, which [creates well-paying] jobs.
In some sense, Wall Street’s gain is proportional to Main Street’s pain. The big trade deficits means less production at home, but the deficits needs to be financed by debt–and any debt issuance, including U.S. Treasuries, ends up going through Wall Street.
Reuters financial blogger Felix Salmon weighed in here:
…while Wall Street does do quite a lot of debt finance, I don’t think that activity explains big headcount trends nearly as well as Mandel thinks it does.
So what’s my theory? If you look at the chart, it turns out that the job losses in finance are put into two buckets. There’s “commercial banking”, on the one hand, which has had very small job losses: people have just as many checking accounts and bank loans as they always did. And then there’s “finance and insurance”, which is what we generally think of as Wall Street, but which also includes the enormous number of employees in the insurance industry. And just like commercial banking, the insurance industry is pretty steady, and is going to have seen very few job losses indeed. What’s more, it’s probably bigger, in terms of total headcount, than the investment-banking industry.”
Before reading too much into these numbers, then, I’d like to see a bit more disaggregation. It might be true that Wall Street hasn’t seen condign punishment in terms of job losses. But on the other hand, it might not.
That said, we pulled up a full breakout of all sub-sectors under the financial sector in the nation and New York State using Analyst, and passed it along. Salmon included the Excel file in his second post.
In it you’ll see there’s clearly a lot going on under the hood — Salmon points out that debt, and debt servicing, were not behind the low losses in financial employment:
But “credit intermediation and related activities” saw job losses of 8% between 2007 and 2009. The gains were in things like commodity contracts dealing, which saw 7% more jobs; “trusts, estates, and agency accounts”, which saw 11% more jobs; and of course “miscellaneous financial investment activities”.
The insurance industry in general was flat from 2007 to 2009, while investment banking generally (”securities, commodity contracts, investments”) saw its total employment rise by 1% over that time.
In fact, the biggest job losses of all are exactly where Mandel expected to see gains. ” Nondepository credit intermediation” saw total jobs fall by 18%, while “Real estate credit” fell by 31%. “International trade financing” was exactly flat.
The financial sector did get hit hard — in certain parts. Other parts (as we well know) have done spectacularly well.
But there’s another part to this story. Mandel had pointed out that there was something odd with the unemployment rate for the financials:
Equally odd, the unemployment rate in finance and insurance, at 7.3%, is substantially lower than the 10.1% for the economy as a whole. And yes, part of that reflects education, but that doesn’t explain why finance unemployment is substantially lower than the unemployment rate in the information sector, which is also a high-education sector.
One theory we’d like to offer is this: A lot of financial-sector workers didn’t exist in the usual employment data sets, and so weren’t there to lose in the first place.
The reason for this theory lies in the data sets we build, and the single data set used most often for this kind of analysis, the Quarterly Census of Employment and Wages. Produced by the Bureau of Labor Statistics, “the QCEW program produces a comprehensive tabulation of employment and wage information for workers covered by State unemployment insurance (UI) laws and Federal workers covered by the Unemployment Compensation for Federal Employees (UCFE) program.” (See here for more). Mandel’s numbers that extend into the first quarter of 2010 are from the Current Employment Statistics, which is essentially a “preview” of full QCEW.
(Click image for full-sized chart.)
But not all workers are covered by unemployment insurance — sole proprietors are one example. In certain financial sub-sectors, envisioning a one-man shop in commodities or investments is easy. Another kind of worker not covered by unemployment insurance would be an independent contractor; someone who files a 1099-MISC. Stockbrokers, sales reps, analysts, and specialists of all sorts can classify themselves as such. Employers can keep personnel-related costs down to a minimum by staffing a company with independent contractors, because worker’s comp premiums and unemployment insurance withholding are not required for 1099’s.
In this paper (PDF), the Department of Labor mentioned that the financial and insurance industries would sometimes use 1099 statuses to:
…recruit employees, train them for a year, then make them switch status to
independent contractors, but continue to use them under the same terms and
conditions as before. Minneapolis-based financial advisors of American Express
filed a lawsuit alleging this practice. In another federal lawsuit in California
(AllState Insurance) agents alleged that the employer retained the authority of an
employer without shouldering the accompanying financial responsibilities. The
agents who sell products only for AllState got slightly higher commissions by
switching employment status, but lost most of their benefits and business expense
reimbursements, while the employer maintained all prior elements of
direction and control.
This is not to say that all independent contractors or their employers are up to something shady. It should be pointed out that this paper was written in 2000 — and while it is very likely that enforcement against abuses has stepped up since this report, we can also expect that companies have concurrently developed more complex methods to achieve essentially the same results. This is, after all, the financial sector we are talking about.
There is a financial incentive for employers to hire independent contractors — and in a boom market for financial and insurance products, there would be big incentives for employees to accept this sort of position and do well in it, and employers would probably kick some of their savings in premiums and withholdings back into commission and contract amounts. Everyone wins — for a while.
In the meantime, this activity would not even register in the QCEW, because an independent contractor is technically not an employee. There would be no gains in employment in QCEW, because they weren’t hired, they were contracted. There would be real employment going on, but because of it’s classification, it eludes the unemployment insurance rolls, and therefore wouldn’t show up in QCEW.
In order to get data on those workers who are working in the economy (but not covered by unemployment insurance) we use State and Local Area Personal Income and employment data from the Bureau of Economic Analysis as well as Non-Employer Statistics and County Business Patterns (both from the Census Bureau).
By comparing the numbers derived from this process to the numbers in the QCEW, we are able to look at a ratio of covered (by unemployment insurance) to non-covered workers.
In the U.S. finance and insurance sector, we estimate that in 2008, 34% of the workers are not covered by unemployment insurance. And that’s the average, spread out over the sector. For comparison, in 2008, 4.43% of Manufacturing workers and 36% of Construction workers were not covered under unemployment insurance. (FYI: The bulk of non-covered workers in Construction are in residential construction and specialty trades).
Sub-sectors like Central Banks, Commercial Banks, Savings Institutions all had low, single-digit percentages of non-covered workers. However, when we get into sub-sectors like Securities, Commodity contracts and Investments, the majority of workers in the field (66%) are not covered by unemployment insurance.
The further we push into more esoteric forms of finance, the higher the percentage of non-covered workers: Miscellaneous Intermediation (84%); Portfolio Managment (80%); Trust, Fiduciary, and Custody activities (81%), and so on. We’ve included the full breakout in .xls here.
So our theory runs like this — in the finance and insurance industry, there was likely a widespread use of the 1099 status, evidenced by the high rate of non-covered employees in the sector nationally. When the financial crisis hit, independent contractors were “laid off” from companies, but because they weren’t employees of the firms, they would not show up as a decline in the Quarterly Census of Employment and Wages.
As an independent contractor, they would have no unemployment insurance withdrawn from their paycheck. They would be responsible for putting money towards that themselves. In that position, they would have a huge incentive to enter into any kind of employment quickly, even if it meant a serious decrease from their last paycheck.
Further, their status (as a “fired” Independent Contractor) would not even show as a decline in even the non-covered data sources unless they (a) made so little money that they didn’t need to file a 1099, or (b) they got hired as a regular employee, and didn’t file at the end of the year as a 1099. And if the latter, they would show up as a gain in employment in the QCEW data, and then there would be a lag time before they disappeared from the non-covered data.