The Fox OFCCP Report


October 2012

OFCCP's New Formulas to Identify Federal Contractors For Audit

I cannot currently prove it in a court of law, but I would bet my next paycheck that I know how OFCCP is now going about the task of selecting federal contractors for audit.

OFCCP roiled the waters last year in Fiscal Year 2012 with new and secret procedures to select contractors for audit. While the President promised a transparent federal government, Pat Shiu never promised to make OFCCP policy and practice changes transparent. And she has not. One of the current difficulties between OFCCP and the contractor community, indeed, springs from the contractor community's perception that OFCCP now routinely makes unilateral and non-collegial-body decisions and regularly deploys new operating procedures, especially concerning audits. Contractors learn about these matters "by hook and by crook" which has made the role of OFCCP vendors knowledgeable across a wide spectrum of geographically dispersed OFCCP District Offices even more precious and more valuable. (My daily e-mail "inquiry and benchmarking" traffic from law firms and vendors is at an all-time high relative to any other administration.)

While OFCCP is not telling how it currently selects contractors for audit, I am going to break that news below. While I do not have any "inside baseball" leak on this one, I am fairly certain I have "cracked the code" and understand what OFCCP is now doing to select contractors for audit. OFCCP argues that there are good reasons not to reveal its selection protocols so that contractors will not know who might be audited. That explanation sounds fair and right on the surface of it. But as many contractors have pointed out, that explanation does not in fact "hold water" for at least three different reasons:
  1. No other administration has had the fear that discussion of its audit selection procedures would reveal which particular contractors and which particular contract establishments will be audited. When I built OFCCP's first audit selection tool in 1982 with Robert Gerlerter and his Policy Division Branch, we had no hesitation to publicly explain the mechanics of the system. I called the audit selection system we built "EEDS: the Equal Employment Data System".

    In 1978, the United States Supreme Court had decided the seminal 4th Amendment case involving civilian agency on-site searches of companies: Marshall v. Barlow's, Inc. 436 U.S. 307(1978). Marshall was Ray Marshall, the Secretary of Labor in the Carter Administration. Barlow, Inc. was a small plumbing company Bill Barlow ran in Pocatello, Idaho. OSHA wanted to do an on-site inspection of Mr. Barlow's offices. He declined to permit the on-site inspection when the OSHA investigator could not produce a warrant to allow entry onto Mr. Barlow's land. OSHA then sued Barlow's seeking access. The case wound its way up to the U.S. Supreme Court where the court laid down its now famous holding that the 4th Amendment protected commercial entities (i.e. businesses) and that to come ON-SITE (that is the operative word), a federal civilian agency had to have "probable cause" to enter the business. A civilian agency could make that showing, however, if it could later show (upon challenge in federal court) that it had either (1) "specific evidence of an existing violation" or (2) "reasonable legislative or administrative standards for conducting an inspection".

    When Secretary Donovan appointed me to OFCCP to help Ellen Shong run the place, I was terribly concerned that OFCCP was daily violating the 4th Amendment. At the time of my arrival at OFCCP, the agency was selecting contractors for audit using any variety of local cockamamie systems which lacked both evidence of prior violations or, in the alternative, any neutral administrative selection standards. And, of course, IN THOSE DAYS, OFCCP's regulations required OFCCP to go on-site (NOTE OPERATIVE WORD) 100 times out of 100 in every audit.

    I asked Robert to suggest a "neutral administrative plan" OFCCP could deploy as soon as possible. It turns out he and his team had been thinking about this and suggested, as a stopgap tool, that OFCCP use the EEO-1 data to select contractors for audit. The EEO-1 file had the advantage that it was current (many databases were outdated and/or incomplete—especially the Dunn & Bradstreet database) and/or did not identify which employers were federal contractors. The Federal Data Procurement System identified, accurately, and in real time, all prime federal contractors, but not covered federal subcontractors and not the race/sex/ethnicity of their employees. So, the EEO-1 file was useful in that it provided reports of establishments with at least 50 employees and with "Subpart C" information (noting which contractors confessed to being a covered federal contractor or subcontractor with at least one establishment of 50 or more employees).

    This was a great start, but we had a few gnawing technical problems: how should OFCCP pick which companies to audit in each of OFCCP's then 65 District Offices spread around the country and which establishments, exactly, within those companies should OFCCP audit? We were painfully aware OFCCP could not have a "cop on every corner". We were doing 7,000+ audits per year (Supply and Service and construction) and there were almost 200,000 establishments out there from among 100,000-120,000 federal contractors each year (the list changes each year as the federal government lets new contracts and old ones expire and as companies expand, contract and go out of business).

    I had tried quite a few lawsuits by the time I joined OFCCP and the process of picking contractors for audit struck me as a lot like the process of picking jurors: it was not so much "picking" as it was "deselecting". So, I looked for ways to determine who not to audit. One of Robert's team made the significant observation that the EEO-1 data ALSO came with detailed race/sex/ethnic information for the 9 EEO-1 categories (subsequently expanded to 10 in 2000) of information for each establishment and each company had an industry SIC code (Standard Industrial Classification…since 2000=NAICS or North American Industry Classification System).

    Brilliance struck in my office as Robert and I hammered out the architecture of the audit protocol: we would ask the EEDS computer to sort the EEO-1 database by EEO-1 category within SIC code and find the like establishments reporting the FEWEST minorities and women, relative to that company's SIC code comparators. We could compare the employment and the vertical penetration of minorities and women up the 9 EEO-1 categories and see who had more and who had fewer. We also figured out from OFCCP's budget plan how many audits each District Office was supposed to accomplish in the upcoming year. We were then able to instruct the EEDS computer to divide the EEO-1 file of Supply & Service establishments by OFCCP District Office. We then instructed the computer to find the contractor establishments with fewer minorities and women and select them for audit.

    I frankly now forget how we instructed the EEDS computer to apportion audits across different SIC codes, but I am sure we must have done so to avoid driving all of OFCCP's audits into one or a small number of industries. Maybe we told EEDS to run an audit in each industry (since in those days we figured there were only about 55 primary industries), and most OFCCP offices started and finished about 100 audits per Fiscal Year. Of course, there were a few dozen other little details Robert had to work out--like what to do if there was no other competitor to a contractor in the same OFCCP District Office jurisdiction (yes, Robert created an algorithm to collect together "like" SIC code establishments across different geographical locations and adjust in some rough fashion the availability of minorities and women so we could figure out whether to audit the tool and dye plant in a nearly all-White county in Ohio or the tool and dye plant in a nearly all-Black county in Mississippi).

    We did not mind telling contractors how the audit selection system worked since contractors could not replicate it. One would have to have had not only the entire EEO-1 database, but also would have had to have known the precise boundaries of the OFCCP District Offices and the number of upcoming scheduled audits. The system was not random (random would be selecting with no algorithm), but it was neutral. AND, significantly, creation of the EEDS auditing system and publication of its procedures quieted overnight contractor and civil rights critics of OFCCP's selection procedures. Lawsuits contractors were threatening evaporated. Discussion of this issue came off industry seminar agendas. All were suddenly content that there was an OFCCP audit plan and it was fair.

    I thought of EEDS as a 2-year stopgap system to allow OFCCP time to develop a more sophisticated system to meet its 4th amendment obligations. Instead, the system endured until last year…although the Clinton Administration changed its name to the Federal Contractor Selection System: FCSS.

  2. The second reason this Administration should not worry about describing generally how OFCCP's audit selection process works is that OFCCP now announces, in advance, which contractor establishments will soon be audited. Where's the mystery, then, in who is going to be audited?…although to be sure, the pre-warning is typically only a few days to about 6 months? The CSAL Letters (Contractor Selection for Audit Letters) nonetheless eliminate the opportunity for surprise…which is OFCCP's current articulated reason to cloak the audit protocols in the first place (so a contractor won't know it is susceptible to audit).

    So, the CSALs undo most of the asserted rationale to keep secret OFCCP's audit protocols. Indeed, the CSALs go one-step beyond: their issuance makes irrelevant the OFCCP audit selection algorithm since OFCCP publishes the results of the algorithm in advance of the audit. As long as the CSALs remain, they signal the very thing OFCCP does not want to tell us about its auditing system: who is going to be audited.

  3. Transparency is required to know OFCCP audit protocols, at least in general enough terms to convey at least a sense of fairness to the contractor community. OFCCP audits are a VERY expensive activity for a contractor (not to mention cash-strapped OFCCP). Defense of an OFCCP audit is now a $30,000-$50,000 off-budget expense which just landed on a VP of HR's desk. Contractors want to understand why they were suddenly so favored and have been so financially taxed. When contractors thought the OFCCP audits were neutral, it was a regrettable, but understandable, event when OFCCP's audit Scheduling Letter arrived. And, a contractor's receipt of 10 or 20 OFCCP audit Scheduling Letters in the same year, as now routinely occurs, can spell economic and workload disaster for an already overworked HR Department and its lawyers.

Again, while I cannot prove it in a court of law, here is my best thinking as to how OFCCP picks Supply and Service Contractors for audit. I think OFCCP has discarded entirely and completely the EEDS/FCSS architecture Robert and I laid down (which compared EEO-1 participation rates of minorities and women). Rather, while OFCCP still uses the EEO-1 file, I think OFCCP now uses it (along with Dun & Bradstreet and the "Next Generation: Federal Data Procurement System") simply to find confessed federal contractors and the locations of their establishments. The selection formula from among this universe is, I believe:
  • by industry;

  • with a "recidivist algorithm" (my language) which causes perhaps up to 5 audits in the next audit cycle of those contractors which have signed a "paperwork Conciliation Agreement (CA)" (meaning a CA which resolves mere recordkeeping and documentation deficiencies: up 400% in this Administration); and

  • with a second "recidivist algorithm" which causes perhaps a minimum of 10 audits in the next audit cycle of those contractors which have signed a Conciliation Agreement curing alleged discrimination law problems (meaning a CA in which the contractor has agreed to pay back pay).
The industries I think OFCCP has chosen to audit include, I believe, at least: energy; health care; financial institutions; transportation; food; and manufacturing. The sum total result of the changes to OFCCP's selection-for-audit rational is that fewer companies are being audited, but many more establishments are being audited among those (fewer) companies OFCCP has selected for audit.

Accordingly, while the vast majority of contractors are sighing with relief, since OFCCP has now vanished from their doorsteps entirely, other industries and companies in those industries which OFCCP is targeting are feeling very taxed, ambushed, surprised, and angry that they are suddenly encountering numerous audits (some reportedly with as many as 100 audits in one year). Defense of these audits potentially cost heavy six and even seven figures to defend with a compliance staff built to handle typically 1 or maybe 2 audits per year. And, are the audit results any different? Are OFCCP's back pay collections skyrocketing? No. OFCCP backpay collections are actually shrinking in size and number (to less than 1% of all audits/yr), despite an OFCCP work force about 20% larger in size.