Manager's Choice

Learn. Improve. Learn Some More.

To: The Director of Operations, West Dakota Department of Social Services
From: Bob Behn
Re: Making the Performance-Contracting Experiment a Success
Date: April, 2000

Making performance contracting work won't be easy. The theory is nice. But the details keep getting in the way. And there are variety of pitfalls that can hinder improvements in performance — and perhaps even ruin any opportunity for your experiment to succeed.

Bob Behn's Manager's ChoiceYou have already stumbled into some luck, however. Senator Hagen wants you to concentrate on job training. This creates important advantages.

Advantage 1: You can concentrate your attention.

You don’t have to prove that performance contracting can work for every social policy. You just have to demonstrate that performance contracting can work for job training. You don’t have let one contract for job training and another child-support enforcement, and yet another for drug-abuse counseling. Job training will, itself, be enough of a challenge; you are going to have to learn a lot, and so are your contractors. But because your contractors will all be dealing with the same kind of social-service problems, you will be better able to help them, and they will be better able to help each other.

Advantage 2: An obvious (if imperfect) performance measure already exists.

Your performance contracts will need to contain one or more performance measures. Such measures will help the contractors and their staff (and you and your staff) recognize when they have improved performance and when their performance is inadequate. They should also motivate the improved performance that you desire.

But performance contracting is tricky because the performance indicators that are available to incorporate into the contract won’t precisely match the mission you seek to achieve. Most of West Dakota’s job-training programs are designed to move welfare recipients to economic and psychological independence. But you won’t have a nice indicator of any trainee’s increased “independence.” So as a performance measure, you will have to use some kind of surrogate. The obvious one is job placements. The more people placed in jobs, the more people who are likely to become independent. Moreover, you can’t move many welfare recipients to economic and psychological independence without placing them in jobs. Still, the connection isn’t perfect; a contractor can place a lot of welfare recipients in jobs and still not really move many of them to true independence.

Nevertheless, performance measures need to be objective, measurable, and easy to understand. Job placements possess these three characteristics. Moreover, contractors can actually produce these outputs.

As a performance measure, job placements will also help you compare different contractors. You want to know who is doing well and who isn’t. You want to be able to use this information to help everyone improve. If all of your dozen contractors are concentrating on producing job placements, you will all be able to use this performance measure to learn and improve.

Still, to demonstrate some success — to give performance contracting a fair chance at demonstrating its value within West Dakota’s political environment — you need to take four specific actions:

Action 1: Pick contractors with a track record.

Don’t make this a controlled, scientific experiment. You aren’t even sure what performance contracting ought to look like in West Dakota. Only after you have as much experience with performance contracting as you do with traditional regulatory contracting will you be ready to undertake a scientifically valid comparison.

Right now, you just want to make it work — somehow. And to make it work, you need a dozen good contractors. So select the contractors you know have already established a track record (under the existing system of regulatory contracts) of training and placing welfare recipients in jobs. Sure, they’ll have a higher base from which to improve. But, at least these people will have established a base from which you can benchmark their progress. And because these contractors are already doing well, they will recognize how the existing system hinders performance. Indeed, many of them will enthusiastically offer suggestions for how best to design the new performance contracts.

Action 2: Pay vendors for reaching milestones, not just for the final product.

Social services are complicated. Job training is complicated. A contractor can take in a job-training client, but it can’t always convert that individual into the perfect final outcome: an economically productive individual. And even if the contractor does produce that perfect, final outcome, it may take a long time.

Meanwhile, the contractor needs a source of income to pay the rent and the staff. If it has to wait to the end to get paid, it can’t survive. So establish some specific job-training milestones and pay the contractor a portion of the total performance fee every time one client makes it to the next milestone. Such milestones might include two or three steps in the training process, the initial job placement, plus one-month and (say) six-month retention.

Action 3: Monitor lots of performance indicators.

What can go wrong? Lots of things. For example, your contractors will have a clear incentive to focus on the performance milestones that get them paid; this is the idea of performance contracting. In the process, however, they may ignore the true purpose of the contract: to convert welfare recipients into economically and psychologically independent citizens. Or, because you are giving your contractors flexibility in how they use their funds, you may encounter some fraud, waste or abuse.

So, while you use a few indicators to drive the performance of your contractors, you have to check on a lot of other things. You have to monitor a wide variety of indicators of performance. You want to track the recidivism rate; maybe the contractors are placing their clients in dead-end or temporary jobs. You want to track how many clients drop out after the first or second milestone; maybe you are paying too much for the early milestones, and the contractors are taking advantage of this to churn a number of clients through these easy stages. And you have to monitor some traditional indicators of financial integrity and of fairness in dealing with people. If something goes wrong, you want to catch it before the press or the inspector general does.

Action 4: Start simple, and be prepared to learn and change.

You won’t get all this right the first time. You can’t. But this isn’t your objective. Rather, you simply want to demonstrate that performance contracts can produce better job-training results in West Dakota.

So don’t make all this too complicated. Start simple. Learn. Improve. And learn some more. Design your experiment so that, at the end of the three years, your 12 performance contractors have indeed, improved their job-training performance.

Agree or disagree? If you think you have a better way to deal with this month's Manager's Choice dilemma, or would like to expand on Bob Behn's approach, share your thoughts with other readers here. Send your solution to mailbox@governing.com. Please include your name, location, government or business title or job description, and a daytime phone number (for verification purposes).

To see responses posted by other readers, click here.

Copyright © 2000, Congressional Quarterly, Inc. Reproduction in any form without the written permission of the publisher is prohibited. Governing, City & State and Governing.com are trademarks of Congressional Quarterly, Inc.