If you outsourced your angels, would more of them dance on the head of a pin?

It's a nonsense question, of course. Here's another: What fraction of an organization's work should be outsourced, and which fraction should be done in-house?

Again, an impossible question. Without specifics of what the organization does and what services are available in the market, there is no way to know what the right answer might be.

In the same way, any arbitrary conclusions about how many government jobs should be outsourced or insourced should be greeted with skepticism.

This issue is relevant because the federal government, through the Office of Management and Budget, has started pushing "insourcing," setting a goal of reducing contract spending by 7 percent over the next two years.

It is always tempting to set arbitrary goals to drive organizational change. Like most temptations, this one should be resisted.

I fell into this trap in 2000, when I served as the chief domestic policy advisor to then-Governor George W. Bush during his presidential campaign. Looking ahead to how many federal officials would likely be retiring during the eight years of the next presidency, I suggested that candidate Bush pledge to replace only half of them, eliminating the remainder of these jobs through efficiencies and outsourcing.

Once President Bush took office, I served on the Commercial Activities Panel, convened by Comptroller General David Walker. The panel was charged with examining the standards for government outsourcing, what the definition should be of "inherently governmental" jobs, and the quality of the public-sector workforce in certain areas. Time and again, panel members excoriated the artificial target of government managers that President Bush had promised to retire, as well as his goals to increase privatization.

We had gotten it backwards. Any goals in reducing the federal workforce should have come only after an examination of the specific activities of the federal government and an analysis of the commercial availability (and suitability) of competitive services. Instead, goals were set in a vacuum, much like the debates about angels dancing on the heads of pins.

The pendulum has swung, and now OMB is pushing in the opposite direction. I made a mistake in 2000 when I proposed a percentage reduction in employees to President Bush, and I likewise believe that any arbitrary decision to pursue insourcing is equally ill-advised.

This country and its citizens deserve the best possible quality of government services. Government must ensure that those services are produced in an effective, fair and affordable manner. But whether any given service should actually be accomplished by a government employee, a government contractor or a combination of both depends on the service itself and the relative capabilities of the private and public sectors in that realm.

Soon after I became mayor of Indianapolis in the early 1990s, I abandoned an arbitrary target for privatizing services because I realized that it unfairly assumed that one set of employees was inherently more efficient than the other. So instead of committing to privatizing a certain percentage of activities, we became fanatical about competition.

Competition brings out the best in organizations and individuals. Think of the last time you played tennis or golf against a rival. Not only does such competition spur effort, it enables you to "discover" who really is best at a particular activity. As the sports announcers say, that's why you play the game.

So rather than committing to insourcing 7 percent of existing contracts, let's ask some fundamental questions: What are the costs (activity-based costing) per unit of work accomplished? How are outputs and outcomes measured? How are high performing workers acknowledged and rewarded? How is citizen satisfaction measured and translated into accountability? What are the private and public benchmarks by which productivity is compared? Can a part of the service be tested in the marketplace for true comparisons?

In Indianapolis, we found that theoretical examinations of our public agencies were often way off the actual results produced by competition. For example, a consultant told us that we could save maybe 5 percent by privatizing our wastewater-treatment facility. We held a competition, and wound up saving 44 percent. Even in cases where we didn't contract out, public agencies that were exposed to competition found ways to reduce their costs.

Deciding what to contract out and what to do in-house requires hard facts. Are there private contractors or nonprofits that provide similar services? How recently was the market tested? Most important, many complicated services are neither wholly insourced nor outsourced, but rather are accomplished in partnership, with government responsible for quality management and a network of contractors providing the back-office and front-office support.

For the federal government, there will be certain activities that you may decide are inherently governmental. For reasons of national security, you may want to keep certain skill sets in-house, regardless of cost inefficiencies. That's fine. But as a general rule, a competitive process is the best way to discover whether insourcing or outsourcing makes sense. In every case, the goal should be "rightsourcing."

Politicians of both parties have committed us to far more government than we can afford. At the very least, we should be striving to deliver the highest-quality, best-performing government possible. Now is not time to add more costs, or less quality, by artificially dictating insourcing targets in deciding how to go about the people's business.