Madam Chairwoman and Subcommittee Members:

On behalf of the more than 600,000 federal employees represented by the American Federation of Government Employees, AFL-CIO (AFGE), I thank you for the opportunity to testify today on the question of whether a shift from the General Schedule to some sort of "Pay for Performance" plan would be in the interests of taxpayers, agencies, and federal employees.

The buildup to this hearing has included an ideological campaign to impugn the General Schedule and create a sense that replacing it is an urgent need for our nation. The campaign has included an Office of Personnel Management (OPM) White Paper, General Accounting Office (GAO) condemnation and high-risk designation, a blue-ribbon commission's blessing upon the manufactured conventional wisdom, conferences by interested parties parading as disinterested experts touting "studies" that demonstrate the fatal shortcomings of the General Schedule and glories that await their own design for a new federal pay system.

The manufacturing of an echo chamber consensus on the need to replace the General Schedule with pay for performance has been impressive. Replacing the General Schedule, according to the ideological campaign, is the answer to the government's self-inflicted human capital crisis, the reason the Bush Administration has had to force agencies to privatize 850,000 federal jobs, and perhaps most absurdly, the best way to make sure the government succeeds in preventing further terrorist attacks. If only they could figure out a way to link the General Schedule with obesity, traffic jams, overcrowded schools, and prescription drug prices, perhaps even federal employees could be convinced.

Some of the campaign's signature slogans include the charge that the General Schedule is a system that rewards only "the passage of time" rather than performance, and that it is an anachronism designed for a late 19^th and early 20^th century government populated mostly by clerks and typists rather than the "information based" government of today. Neither charge is really true, but every good marketer knows that repetition breeds plausibility, and today they have at least the ring of truth. After all, the General Schedule does provide financial recognition for experience gained over time, and the federal government has had a pay system since the late 19^th century, so the slogans aren't outright lies. They are, however, highly misleading.

In my testimony today, I will focus on both the broad question of whether individualized pay for performance is a wise choice for the federal pay system, as well as the specific system that has been rumored to have emerged as the favored candidate to replace the General Schedule, the Navy's China Lake Demonstration Project Pay for Performance Plan. The question of whether the China Lake Plan is a worthy successor to the General Schedule is a worthwhile one, especially in response to pay for performance advocates' usual practice of limiting the discussion to whether a fantasized perfect model is preferable to the easily-maligned real system. In addition, I will discuss briefly the General Schedule, since it too deserves an accurate description so that proposed alternatives are not considered or evaluated against an easily dismissed or derided "straw dog."

The version of the General Schedule I will discuss is the one that was established as a result of the enactment of the bipartisan Federal Employees Pay Comparability Act (FEPCA) in 1990. Despite the insistence of some anti-General Schedule ideologues who claim that it is an aged and inflexible historical relic untouched by history, the fact is that the General Schedule has been modified numerous times, in some cases quite fundamentally. FEPCA's distinguishing feature, the locality pay system, has not even had a full decade of experience, since its implementation began only in 1994 after passage in 1992 of technical and conforming amendments to FEPCA that established both locality pay and Employment Cost Index (ECI)-based annual pay adjustments.

FEPCA introduced a panoply of pay flexibilities into the allegedly rigid and never-changing General Schedule:

• special pay rates for certain occupations
• critical pay authority
• recruitment and retention flexibilities that allow hiring above the minimum step of any grade
• paying recruitment or relocation bonuses
• paying retention bonuses of up to 25% of basic pay
• paying travel and transportation expenses for new job candidates and new hires
• allowing new hires up to two weeks advance pay as a recruitment incentive
• allowing time off incentive awards
• paying cash awards for performance
• paying supervisory differentials to GS supervisors whose salaries were less than certain subordinates covered by non-GS pay systems
• waiver of dual compensation restrictions
• changes to Law Enforcement pay
• special occupational pay systems
• pay flexibilities available to Title 5 health care positions, and more.

In addition, FEPCA retained agencies' authority for quality step increases, which allow managers to reward extraordinary performance with increases in base salary that continue to pay dividends throughout a career.

The basic structure of the General Schedule is a 15-grade matrix with ten steps per grade. Movement within a grade or between grades depends upon the satisfactory performance of job duties and assignments over time. That is, an employee becomes eligible for what is known as a "step" increase each year for the first three years, and then every three years thereafter up to the tenth step. Whether or not an employee is granted a step increase depends upon performance (specifically, they must be found to have achieved "an acceptable level of competence"). If performance is found to be especially good, managers have the authority to award "quality step increases" as an additional incentive. If performance is found to be below expectations, the step increase can be withheld.

The federal position classification system, which is separate and apart from the General Schedule and would have to either continue or be altered separately and in addition to any alteration in the General Schedule, determines the starting salary and salary potential of any federal job. As such, a job classification determines not only initial placement of an individual and his or her job within the General Schedule matrix, classification determines the standards against which individual worker's performance will be measured when opportunities for movement between steps or grades arise. And most important, the classification system is based upon the concept of "equal pay for substantially equal work", which goes a long way toward preventing federal pay discrimination on the basis of race, ethnicity, or gender.

The introduction of numerous pay flexibilities into the General Schedule under FEPCA was only one part of the pay reform the legislation was supposed to effect. It was recognized by President George Bush, our 41^st President, the Congress, and federal employee unions that federal salaries in general lagged behind those in the private sector by substantial amounts, although these amounts varied by metropolitan area. FEPCA instructed the BLS to collect data so that the size of the federal-non-federal pay gap could be measured, and closed gradually to within 90% of comparability over 10 years. To close the pay gap, federal salary adjustments would have two components: a nationwide, across-the-board adjustment based upon the Employment Cost Index (ECI) that would prevent the overall gap from growing, and a locality-by-locality component that would address the various gaps that prevailed in specific labor markets.

Unfortunately, neither the Clinton nor the George W. Bush administration has been willing to comply with FEPCA, and although some small progress has been made, on average federal salaries continue to lag private sector salaries by about 22%. The Clinton administration cited, variously, budget difficulties and undisclosed "methodological" objections as its reasons for failing to provide the salary adjustments called for under FEPCA. The current administration ignores the system altogether, and for FY04 has proposed allocation of a fund with 0.5% of salaries to be allocated via managerial discretion. Meanwhile, the coming retirement wave, which was fully anticipated in 1990, has turned into a full-fledged human capital crisis due to highly irresponsible and untargeted downsizing and privatization in the intervening years, as well as a stubborn refusal to implement the locality pay system which was designed to improve recruitment and retention of the next generation of federal employees.

The Navy's China Lake plan started out as a demonstration project under title 6 of the Civil Service Reform Act. It was initiated in 1980, modified in 1987, expanded in 1990, extended indefinitely in 1994 (made into a "permanent" alternative personnel system), and expanded again in 1995. The employees covered by the China Lake plan are approximately 10,000 scientists, engineers, technicians, technical specialists, and administrative and clerical staff—a workforce that is not typical of any government agency, or even a minority of work units in any one agency.

Although the China Lake plan is often referred to as a model for pay for performance, the rationale given to OPM at its inception, and to Congress in its progress reports, was to improve the competitiveness of salaries for scientists and engineers. Nevertheless, the China Lake model is a performance-based pay system that differs from the General Schedule in terms of its classification of jobs into pay bands that are broader than the grades and steps in the GS matrix. Thus it is often called a broadbanding system.

OPM's evaluation of the China Lake plan was positive. They judged it a success in improving overall personnel management at the two demonstration laboratories studied. OPM cited the "simplified delegated job classification based on generic standards" as a key factor in the demo's success, as the time spent on classification actions was reduced, and the official report was that conflict between the affected workers and management declined. In the 10-year period of evaluation, average salaries rose by 3% after taking into account the effects of inflation. The China Lake plan made an explicit attempt to link pay increases within its "broad bands" to individual performance ratings. Starting salaries were also "flexible" within the bands.

It is important to note that the China Lake demo predated the passage of FEPCA by a decade. Indeed, China Lake's experience was invoked throughout the debate over reforming the federal pay system in the years leading up to FEPCA's passage in 1990, and many of FEPCA's flexibilities were based upon positive experiences accumulating in the China Lake demo.

It is worth describing at length the mechanics of the China Lake pay for performance system, apart from its equally elaborate classification system. I do this in part to show how China Lake's design may be appropriate to some scientists and engineers, but not to all federal employees since many are in occupations and workplaces that place extreme or even total limitations on creativity, individual initiative, or individualized performance. I also include this description to show that administrative ease is not one of pay for performance's virtues if the pay for performance system attempts to build in safeguards that limit the role of bias, favoritism and prejudice, as has been attempted at China Lake.

Instead of the General Schedule's 15 grades, China Lake has five career paths grouped according to occupational field. The five occupational fields are Scientists/Engineers/Senior Staff, Technicians, Technical Specialists, Administrative Specialists, and General Personnel. Each career path has classification and pay levels under the broadband concept that are directly comparable to groupings of the General Schedule. Within each career path are included many types of jobs under an occupational heading. Each job has its own career ladder that ends at a specific and different point along the path. Each broad band encompasses at least two GS grades. The China Lake plan describes itself as being "anchored" to the General Schedule as a "reality check." For those keeping count, the China Lake broadband has at least as many salary possibilities as the General Schedule, and at most as many as 107,000, since salaries can really be anywhere between the General Schedule's minimum or maximum.

Movement along an individual career path is the key factor to consider, as the overall plan has been suggested as a pay for performance model. As such, it is important to note that although some individuals may have an opportunity to move up to the top of a career path, not all can. Each job has its predesignated "top out" level. The promotion potential for a particular position is established based on the highest level at which that position could be classified, but individuals' promotions will vary. Promotion potential for a given position doesn't grow just because movement is nominally based upon performance. The only way to change career paths is to win a promotion to another career path altogether, i.e. get a new job. One can move along a pay line, but one may not shift to a higher pay line.

The description of the China Lake system involves pages and pages of individualized personnel actions involving the classification and reclassification of workers, and the setting of salary and salary adjustments. It is certainly neither streamlined nor simple, and asks managers on a continuous basis to evaluate each individual worker on numerous bases. In terms of bureaucratic requirements, and a presumption that managers have the training, competence, available time, commitment, and incentive to be as thorough as this system expects them to be for every single employee under them, the China Lake plan seems unrealistic at best. Further, the plan lacks adequate opportunity for employees to appeal their performance appraisals and the attendant pay consequences.

Unlike some of the radical "at will" pay and classification systems advocated by those who believe that any rules or regulations or standards or systems constitute intolerable restrictions on management flexibility, the China Lake plan retains a requirement to tie salary to job duties and responsibilities, not an individual worker's personal characteristics.

AFGE's invitation to testify today included a request to address the broad issue of pay for performance as it was addressed in the Volcker Commission report. The following represents AFGE's views on the question of whether replacing the General Schedule with a pay for performance system is appropriate or desirable.

The rationales offered by proponents of pay for performance in the federal government have generally fallen under one of four headings: improving productivity, improving recruitment prospects, improving retention, and punishing poor performers. Perhaps the most misleading rationale offered by advocates of pay for performance is that its use has been widespread in the private sector. Those who attempt to provide a more substantive argument say they support pay for performance because it provides both positive and negative incentives that will determine the amount of effort federal workers put forward. Advocates of pay for performance wisely demur on the question of whether pay for performance by itself is a strategy that solves the problem of the relative inferiority of federal salaries compared to large public and private sector employers. That is to say, when pay for performance is referred to as complying with the government's longstanding principle of private sector comparability, what they seem to mean is comparability in system design, and not comparability in salary levels.

Does a pay system that sets out to reward individual employees for contributions to productivity improvement and punishes individual employees for making either relatively small or negative contributions to productivity improvement work? The data suggest that they do not, although the measurement of productivity for service-producing jobs is notoriously difficult. Measuring productivity of government services that are not commodities bought and sold on the market is even more difficult. Nevertheless, there are data that attempt to gauge the success of pay for performance in producing productivity improvement.

Although individualized merit pay gained prominence in the private sector over the course of the 1990's, there is good reason to discount the relevance of this experience for the federal government as an employer. Merit based contingent pay for private sector employees over the decade just past was largely in the form of stock options and profit-sharing, according to BLS data. The corporations that adopted these pay practices may have done so in hope of creating a sense among their employees that their own self interest was identical to the corporation's, at least with regard to movements in the firm's stock price and bottom line. However, we have learned more recently, sometimes painfully, that the contingent, merit-based individual pay that spread through the private sector was also motivated by a desire on the part of the companies to engage in obfuscatory cost accounting practices.

These forms of "pay for performance" that proliferated in the private sector seem now to have been mostly about hiding expenses from the Securities and Exchange Commission (SEC), and exploiting the stock market bubble to lower actual labor costs. When corporations found a way to offer "performance" pay that effectively cost them nothing, it is not surprising that the practice became so popular. However, this popularity should not be used as a reason to impose an individualized "performance" pay system with genuine costs on the federal government.

Jeffrey Pfeffer, a professor in Stanford University's School of Business, has written extensively about the misguided use of individualized pay for performance schemes in the public and private sectors. He cautions against falling prey to "six dangerous myths about pay" that are widely believed by managers and business owners. Professor Pfeffer's research shows that belief in the six myths is what leads managers to impose individualized pay for performance systems that never achieve their desired results, yet "eat up enormous managerial resources and make everyone unhappy."

The six myths identified by Professor Pfeffer are:

1. labor rates are the same as labor costs;
2. you can lower your labor costs by lowering your labor rates;
3. labor costs are a significant factor in total costs;
4. low labor costs are an important factor in gaining a competitive edge;
5. individual incentive pay improves performance; and finally,
6. the belief that people work primarily for money, and other motivating factors are relatively insignificant.

The relevance of these myths in the context of the sudden, urgent desire to impose a pay for performance system on the federal government is telling. Professor Pfeffer's discussion of the first two myths makes one wish that his wisdom would have been considered before the creation of the federal "human capital crisis" through mindless downsizing and mandatory, across-the-board privatization quotas. Pfeffer's distinction argues that cutting salaries or hourly wages is counterproductive since doing so undermines quality, productivity, morale, and often raises the number of workers needed to do the job. Did the federal government save on labor costs when it "downsized" and eliminated 300,000 federal jobs at the same time that the federal workload increased? Does the federal government save on labor costs when it privatizes federal jobs to contractors that pay front-line service providers less and managers and professionals much, much, much more?

Salaries for the 1.8 million federal employees cost the government about $67 billion per year, and no one knows what the taxpayer-financed payroll is for the 5 million or so employees working for federal contractors. But as a portion of the total annual expenditures, it is less than 3%, according to Congressional Budget Office (CBO) projections. Regarding the relevance of low labor costs as a competitive strategy, for the federal government it is largely the ability to compete in labor markets to recruit and retain employees with the requisite skills and commitment to carry out the missions of federal agencies and programs. Time and again, federal employees report that competitive salaries, pensions and health benefits; job security, and a chance to make a difference are what draw them to federal jobs. They are not drawn to the chance to become rich in response to financial incentives that require them to compete constantly against their co-workers for a raise or a bonus.

Professor Pfeffer blames the economic theory that is learned in business schools and transmitted to human resources professionals by executives and the media for the persistence of belief in pay myths. These economic theories are based on conceptions that human nature is uni-dimensional and unchanging. In economics, humans are assumed to be rational maximizers of their self-interest, and that means they are driven primarily, if not exclusively by a desire to maximize their incomes. The inference from this theory, according to Pfeffer, is that "people take jobs and decide how much effort to expend in those jobs based on their expected financial return. If pay is not contingent on performance, the theory goes, individuals will not devote sufficient attention and energy to their jobs."

Further elaboration of these economic theories suggest that rational, self-interested individuals have incentives to misrepresent information to their employers, divert resources to their own use, to shirk and "free ride", and to game any system to their advantage unless they are effectively thwarted in these strategies by a strict set of sanctions and rewards that give them an incentive to pursue their employer's goals. In addition there is the economic theory of adaptive behavior or self-fulfilling prophesy, which argues that if you treat people as if they are untrustworthy, conniving and lazy, they'll act accordingly.

Pfeffer also cites the compensation consulting industry, which, he argues, has a financial incentive to perpetuate the myths he describes. More important, the consultants' own economic viability depends upon their ability to convince clients and prospective clients that pay reform will improve their organization. Consultants also argue that pursuing pay reform is far easier than changing more fundamental aspects of an organization's structure, culture, and operations in order to try to improve; further, they note that pay reform will prove a highly visible sign of willingness to embark on "progressive reform." Finally, Pfeffer notes that the consultants ensure work for themselves through the inevitable "predicaments" that any new pay system will cause, including solving problems and "tweaking" the system they design.

In the context of media hype, accounting rules that encourage particular forms of individual economic incentives, the seeming truth of economic theories' assumptions on human nature, and the coaxing of compensation consultants, it is not surprising that many succumb to the temptation of individualized pay for performance schemes. But do they work? Pfeffer answers with the following:

Compensation consultants like the respected William M. Mercer Group report that just over half of employees working in firms with individual pay for performance schemes consider them "neither fair nor sensible" and believe that they add little value to the company. The Mercer report says that individual pay for performance plans "share two attributes: they absorb vast amounts of management time and resources, and they make everybody unhappy."

One further problem cited by both Pfeffer and other academic and professional observers of pay for performance is that since they are virtually always zero-sum propositions, they inflict exactly as much financial hardship as they do financial benefit. In the federal government as in many private firms, a fixed percentage of the budget is allocated for salaries. Whenever the resources available to fund salaries are fixed, one employee's gain is another's loss. What incentives does this create? One strategy that makes sense in this context is to make others look bad, or at least relatively bad. Competition among workers in a particular work unit or an organization may also, rationally, lead to a refusal on the part of individuals to share best practices or teach a coworker how to do something better. Not only do these likely outcomes of a zero-sum approach obviously work against the stated reasons for imposing pay for performance, they actually lead to outcomes that are worse than before.

What message would the federal government be sending to its employees and prospective employees by imposing a pay for performance system? At a minimum, if performance-based contingent pay is on an individual-by-individual basis, the message is that the work of lone rangers is valued more than cooperation and teamwork. Further, it states at the outset that there will be designated losers – everyone cannot be a winner; someone must suffer. In addition, it creates a sense of secrecy and shame regarding pay. In contrast to the current pay system that is entirely public and consistent (pay levels determined by Congress and allocated by objective job design criteria), individual pay adjustments and pay-setting require a certain amount of secrecy, which strikes us as inappropriate for a public institution. An individual-by-individual pay for performance system whose winners and losers are determined behind closed doors sends a message that there is something to hide, that the decisions may be inequitable, and would not bear the scrutiny of the light of day.

Beyond compensation consultants, agency personnelists, and OPM, who wants to replace the General Schedule with a pay for performance system? The survey of federal employees published by OPM on March 25 may be trotted out by some as evidence that such a switch has employee support. But that would be a terrible misreading of the results of the poll. AFGE was given an opportunity to see a draft of some of the poll questions prior to its being implemented. We objected to numerous questions that seemed to be designed to encourage a response supportive of individualized pay for performance. We do not know whether these questions were included in the final poll. The questions we objected to were along the lines of: Would you prefer a pay system that rewarded you for your excellence, even if it meant smaller pay raises for colleagues who don't pull their weight? Do you feel that the federal pay system adequately rewards you for your excellence and hard work? Who wouldn't say yes to both of those questions? Who ever feels adequately appreciated, and who doesn't secretly harbor a wish to see those who appear to be relatively lazy punished? Such questions are dangerously misleading.

It is crucial to remember that the OPM poll was taken during a specific historical period when federal employees are experiencing rather extreme attacks on their jobs, their performance, and their patriotism. The Administration is aggressively seeking to privatize 850,000 federal jobs and in many agencies, is doing so in far too many cases without giving incumbent federal employees the opportunity to compete in defense of their jobs. After September 11, the Administration began a campaign to strip groups of federal employees of their civil service rights and their right to seek union representation through the process of collective bargaining. The insulting rationale was "national security" and the explicit argument was that union membership and patriotism were incompatible. Some policy and lawmakers used the debate over the terms of the establishment of the Department of Homeland Security as an opportunity to defame and destroy the reputation, the work ethic, loyalty, skill and trustworthiness of federal employees. And out of all of this has come an urgent rush to replace a pay system based upon objective criteria of job duties, prerequisite skills, knowledge, and abilities, and labor market data collected by the BLS with a so-called pay for performance system based on managerial discretion.

In this historical context, federal employees responded to a survey saying that they were satisfied with their pay. In fact, 64% percent expressed satisfaction and 56% believed that their pay was comparable to private sector pay.

But as the representative of 600,000 federal employees, AFGE would suggest that they are satisfied with their pay system, not their actual paychecks. Since the alternatives with which they have been threatened seem horrendous by comparison, expression of satisfaction with the status quo in a survey sponsored by an agency determined to give managers discretion or "flexibility" over pay is no surprise.

Perhaps more important for the subject of pay for performance in the context of the survey is the fact that 80% report that their work unit cooperates to get the job done and 80% report that they are held accountable for achieving results. Only 43% hold "leaders" such as supervisors and higher level management in high regard; only 35% perceive a high level of motivation from their supervisors and managers, and only 45% say that managers let them know what is going on in the organization.

In this context, it seems reasonable to ask if the majority of employees are relatively satisfied with their pay, why the frantic rush to change? If federal supervisors and managers are held in such low regard, how will a system which grants them so much new authority, flexibility, unilateral power, and discretion be in the public interest? How will a pay system that relies on the fairness, competence, unprejudiced judgement, and rectitude of individual managers be viewed as fair when employees clearly do not trust their managers? Given that less than a third of respondents say managers do a good job of motivating them, is pay for performance just a lazy manager's blunt instrument that will mask federal managers' other deficits?

No discussion of federal pay is complete without consideration of funding. To the extent that pay for performance is proposed as a replacement for the General Schedule that would be "budget neutral" and exclude additional funding, AFGE will work in opposition. Federal salaries are too low, and they are too low not just for prospective employees, or employees in "hard to fill" positions or employees who intend to stay in government for short periods – federal salaries are too low for all federal employees. There may be legitimate disputes about the size of the gap between federal pay and non-federal pay, but it is indisputable that federal salaries are too low across-the-board.

AFGE is supportive of Congressional attention toward the inadequacy of federal pay. We are also supportive of those who are looking for ways to reward federal employees financially for excellent and extraordinary performance. But at the same time we must caution that rewards for excellence and extraordinary acts must be supplements to a fully funded regular pay system, not substitutes; and these supplements must be fully and separately funded.

We are also highly concerned about the introduction of managerial discretion over pay in the context of recent aggressive attempts on the part of this administration to disparage and dismantle important elements of the merit system and provisions of title 5 which protect federal employees from discrimination in hiring, firing, pay, classification, performance appraisal, and which provide for collective bargaining. The current system makes sure that winning a federal job is a matter of what you know, not whom you know. The current system makes sure that the salary and career development potential of that job are a function of objective, job description criteria, not a manager's opinion of an individual worker's "competency" or skin color, gender, religion, age, political affiliation, or union status. Deviations from these protections are not warranted. Our nation has prospered and our government programs have benefited from having a professional, apolitical civil service that is strongly protected from corruption and discrimination. Introducing individualized pay systems that grant enormous power to federal managers regarding pay represents a grave danger to this protection.

The advocates of pay for performance in the federal government have the burden of demonstrating exactly how and why the General Schedule prevents federal managers from managing for excellence and productivity improvement. They must demonstrate exactly how and why each of the merit system principles will be upheld in the context of political appointees' supervision of managers who will decide who will and will not receive a salary adjustment, who will receive a higher salary for a particular job and who will receive a lower salary for the same job. They must demonstrate exactly how and why individualized pay for performance is superior to systems that provide financial reward for group and organizational excellence. They must demonstrate exactly how and why paying some people less so that they can pay others more will contribute to resolving the federal government's human capital crisis and attract the next generation of federal workers to public service. They must demonstrate exactly how and why agencies will invest in the training, oversight, and staffing necessary to administer elaborate, federal employee by federal employee pay for performance plans fairly and efficiently. And they must demonstrate that they will be able to secure adequate funding so that pay for performance does not degenerate into a false promise, where discretion is exercised to award higher salaries only to recruit and/or retain particular individuals rather than to reward actual performance.

This concludes my testimony, and I would be happy to answer any questions Members of the Subcommittee may have.