Steve Sailer`s “Test Case”–Why Civil Service Testing Is Important

Here`s an American Conservative article by me that`s never been online before:

Test Case
By Steve Sailer
Bureaucracy fails when civil servants aren’t put to the test.

The American Conservative
September 10, 2007

You might think that liberals who want to expand the federal government and conservatives who want to shrink it could at least agree to improve how well it works. Yet, good government projects, such as boosting the quality of the federal workforce, have largely dropped out of media discussion despite ample evidence that the federal government no longer functions as well as it once did, relative to what’s now technologically feasible. In its mid-20th Century prime, the federal government matched up reasonably well in efficiency and effectiveness against, say, Sears-Roebuck. Today, however, it’s blown away by Wal-Mart’s relentless improvements.

For example, in June, while the Senate was blithely considering mandating a convoluted new immigration system for the federal bureaucracy to administer, the State Department’s nearly century-old responsibility for issuing passports was melting down under the strain of merely a moderate increase in demand predictably caused by a law passed three years before. In an era of cheap networked computing, many Americans still had their summer travel plans ruined by federal incompetence.

Everything about the federal government is extraordinarily complicated, and thus there are many plausible explanations, both specific and general, for its current malaise.

Democrats, for instance, have denounced Bush Administration appointments. Indeed, the latest political picks seem prone to “marketing major post-modernism,” the assumption picked up in college that some egghead over in France proved there’s no such thing as truth, so there’s no need to feel guilty about shamelessly spinning everything for maximum political benefit. Still, there are roughly 600 civil servants for each Presidential appointee, so the nefarious impact of the thin top layer can be overstated.

Much less debated is what Steve Nelson, director of the Office of Policy and Evaluation at the Merit Systems Protection Board, a federal watchdog agency, calls the “human capital crisis” facing the federal civilian workforce of nearly two million (not counting the Post Office).

Neither party has much incentive to tackle the problem. Democrats don’t want to mention government worker ineptitude because that raises doubt about their hopes to expand the government’s power over health care and other areas. Conversely, Republicans don’t want public confidence in government to increase, so they can expand the outsourcing of federal duties to for-profit firms.

Clearly, growing economic inequality leaves the civil service hard pressed to compete for the finest workers versus Goldman Sachs’s bonuses and Google’s stock options.

Ameliorating the pay gap would be expensive. Much cheaper, yet seemingly unthinkable in the current climate, would be for the federal government to do a better job of choosing among its job applicants by employing a tool used by both colleges and the military in picking whom to take: standardized testing.

In fact, the feds themselves once had an excellent test for entry-level job applicants. One of the last malignant relics of the Carter Administration is the enduring hash it made of civil servant hiring by abolishing the Professional and Administrative Career Examination (PACE) in January 1981.

That this disastrous step has disappeared down the memory hole exemplifies the reigning prejudice in modern America against publicly discussing about how best to select people. In private, selection is increasingly an obsession, with the competition to win admission to elite colleges (and even, among the New York media class, elite preschools) ever-growing. Ironically, one of the most popular hobbies to emerge in recent decades has been “fantasy football,” which is nothing but selection: fans draft players and then see whose “team” has the best statistics each Sunday.

Yet, nobody wonders about how to select better civil servants.

It will never be an easy job. While the 19th Century federal workforce was “an army of clerks,” the need for technically skilled bureaucrats has accelerated since the Progressive Era, with new agencies like the FDA, SEC, NASA, and EPA needing high-end talents. The Volcker Commission on Public Service reported in 2003, “In 1950, 62 percent of the basic federal workforce was in GS grades 1-5, with only 11 percent in the top five grades; by 2000 those relationships were reversed: 15 percent of the federal workforce was in the bottom five grades, compared to 56 percent in the top.”[PDF]

This professionalization of federal employees has changed for the better the greater Washington D.C. area, which was once an economic wasteland with little private enterprise. In 1985, a woman told me her father had owned a factory in Washington. Never having heard before of a normal business in the capital, I asked her what it made. She replied, “Rubber stamps.”

Today, the D.C. region ranks with the San Francisco Bay Area for most residents with graduate degrees, providing a beneficial spillover on private industry. The spouses and children of highly educated civil servants are often employed by the region’s thriving telecom businesses.

Federal workers are, on average, intelligent, but are we hiring the best available under the mishmash of supposedly temporary selection techniques concocted after the Carter Administration abolished the highly evolved civil service examination?

Job testing began in imperial China two millennium ago. Early Jesuit missionaries to the Middle Kingdom were so impressed by the efficiency of mandarin administrators that they brought back to Europe the notion of competitive examinations. Prussia was the first to try it.

Initially, the American Republic relied upon a spoils system with the President appointing his campaign volunteers to sinecures, but after a disappointed office-seeker assassinated President James Garfield, Congress created the civil service merit system in 1883. Early tests were crude, but beginning in the 1920s, the civil service began to use more scientific cognitive exams. It’s not a coincidence that the Volcker Commission said: “The middle decades of the 20th century were in some ways a golden age for public service recruiting and retention.”

The U.S. military, which subjected all draftees to the newly invented IQ test during WWI, preceded the civil service in mental testing. It’s not widely publicized, but the armed forces remains devoted to their IQ-like Armed Forces Qualification Test, having accumulated abundant data showing that IQ is one of the best predictors of both trainability and performance. With the downsizing of the military after the Cold War, the Pentagon immediately took the opportunity to raise the virtual minimum IQ for enlistment to 92 (the 30th percentile). Even today, despite the difficulties of recruiting during the Iraq War, the typical new boot private has a higher IQ than the national average.

Testing has been shown to work well for selecting federal white-collar employees as well. A 1986 study by Frank L. Schmidt of the federal Office of Personnel Management found that hiring “on valid measures of cognitive ability, rather than on non-test procedures (mostly evaluations of education and experience), produces … a 9.7% increase in output among new hires.” Indeed, problem-solving skills may be more useful in government than in private industry because having a salesman’s personality is less important.

Compared to soldiers, testing for entry level hiring is perhaps even more crucial for civilians because civil servants are notoriously hard to fire. Moreover, the feds mostly promote from within, seldom headhunting for middle level managers from the private sector.

Hence, government workers are rather like students at the top universities, who are almost never flunked out. At Harvard, 98 percent of freshmen are allowed to graduate, which puts intense stress on Harvard’s admission process to not let in clunkers. So, despite the SAT’s infamous political incorrectness, Harvard demands high SAT scores, with incoming students averaging about 1500 out of a possible 1600. Whatever their other failings, their SAT scores ensure they have the smarts to make it through Harvard.

Similarly, the federal civil service once invested in increasingly sophisticated brainpower tests to identify young people who could prove competent senior managers in future decades. The Junior Management Assistant test debuted in 1948, followed by the Federal Service Entrance Examination (FSEE) in 1955, a test roughly comparable to the Graduate Record Examination (GRE) now required by grad schools.

In 1972, a lawsuit claimed that that the FSEE was biased because blacks and Hispanics didn’t score as well as whites on average. So, the Nixon Administration deep-sixed it and introduced the sophisticated PACE, which was elaborately validated as predicting performance in 118 federal jobs. The PACE consisted of multiple subtests, which could be weighted differently for each post.

Frustratingly, despite PACE’s impressive predictive power, blacks and Latinos continued to tally lower on it. In another federal discrimination case, the outgoing Carter Administration signed a consent decree in January 1981 agreeing to abolish PACE. Workarounds were “temporarily” implemented until a non-discriminatory general test could be devised.

Twenty-six years later, the Luevano decree‘s makeshifts still control federal hiring procedures. (No such new test has proven feasible.) Federal hiring has devolved into a decentralized hodge-podge. There is some job-related testing, but most agencies emphasize credentials, and assess them in a mindlessly mechanical fashion to boot. A 2005 article in Government Executive by Shawn Zeller observed: “It doesn’t matter whether a candidate earned his economics degree at Harvard or the University of the District of Columbia. Both are considered equal.”

Nelson of the Merit Protection Systems Board notes:

“Our reports show that demonstrating training and experience is not the best indicator of a candidate’s future job performance. Neither is grade point average, yet we use these two methods alone too frequently.”

Emasculating testing and focusing more on qualifications for beginners’ jobs promoted ethnic diversity, but at the expense of hiring youngsters with future upper management aptitude. The old-timers chosen via testing have been retiring, leaving the Luevano generation in charge.

In effect, our post-Carter hiring methods serve as crypto-quotas. Unfortunately, crypto-quotas produce an inferior workforce compared to honest ethnic quotas, which at least would hire the best from each group.

Shifting demands from applicant’s intelligence to work experience has also contributed to what Nelson calls the “upward entry-age spiral:”

“The average age of employees has crept up every year because … we face barriers to bringing in younger people right out of college. Our most common method of evaluating candidates is a system that assesses candidates based on their training and experience, making it difficult for a 22-year-old college graduate to compete with more experienced, but not necessarily higher potential, candidates. Thus the average age of new hires is 34 and 29 for professional and administrative fields respectively… A potential retirement tsunami looms over the Federal workplace.”

By age 34, most people have gotten a life, taking on time-consuming responsibilities for a spouse, children, and aging parents. Thus, they can’t work cheap and they can no longer put in the relentless hours asked of younger professionals in the private sector.

But, hey, nobody seems to mind. Evidently, it’s good enough for government work.