The aging of the Baby Boom generation, combined with the success of the New Public Management in downsizing the federal government, has led to a rapidly aging federal service, a reduced flow of new blood with creativity and updated skills (Lane, Wolf, and Woodard, 2003), and a looming "tsunami" of retirements that is forcing the federal service to reconsider many of its human resource practices. The difficulty of obtaining good data has severely restricted our ability to measure the extent of the problem in the state government workforce. We examine the changing age distribution of that workforce using the 1980, 1990, and 2000 Census 5% Public Use Microdata Samples and the 2001-07 American Community Surveys, and find that state workforces are even older and have aged more than the federal civil service, suggesting that they may experience the retirement tsunami sooner. We also examine the effects of this aging on institutional memory, race/ethnicity/gender diversity, and educational qualifications.

The challenges posed by an aging workforce are no longer a surprise. Many scholars and practitioners have warned that the federal government needs to prepare for a retirement tsunami, especially in top leadership positions and critical occupations. The federal government has begun to respond with initiatives to slow the pace of retirement, attract qualified young people, and simplify the hiring process. State workforces have attracted less attention, even though state governments employ more people than the federal government. Data problems have hidden the extent of the problems, and no central agency has responsibility for worrying about them. We examine whether the state workforce is experiencing the same aging problem, using the 1980, 1990, and 2000 Census 5% Public Use Microdata Samples (PUMS) and the 2001-07 American Community Surveys (ACS), which provide random samples of state, local, and federal government employees, as well as of private for-profit and nonprofit workers. We examine how the age distributions of the workforce vary by sector and how they have changed over time. The coming wave of retirements offers opportunities as well as challenges. Decreases in institutional memory may be offset by fresh blood, increased diversity, and higher educational levels. To assess the expected impact of cohort replacement on the race/ethnicity/gender diversity and educational qualifications, we compare the characteristics of state government employees in their 20s, 30s, 40s, 50s, and 60s. 3 We begin by examining how the federal government assesses and is addressing its aging workforce problem. We then turn our attention to what we know about state government workforces and consider reasons they might have a less severe aging problem. Our data analysis demonstrates, however, that state workers are even older than federal workers, with a higher percentage over 50 and a smaller percentage under 30 than any other sector. As the state sector is most unbalanced in terms of age, it should feel the retirement tsunami sooner and harder than the federal government. We expect cohort replacement to increase the diversity in state workforces but not to have much impact on education levels.

Data and Method Research on state government employees is hampered by difficulty in obtaining useful data. The U.S. Office of Personnel Management (OPM) and the U.S. Merit Systems Protection Board (MSPB) regularly conduct surveys of federal employees, issue reports on the status of the federal service, and make their data available to researchers. No agency has authority or reason to conduct similar surveys of national samples of state government employees. Some state human resource agencies also conduct surveys and issue reports regularly, but they have little reason to coordinate their efforts in ways that would make data comparable across states. Some researchers have obtained access to samples of federal personnel records; a researcher seeking to do similar research about state government employees would need to seek the cooperation of 50 state personnel departments. Research on local government workers faces even higher hurdles; most use governments rather than individuals as the units of analysis. Llorens (2008) points public administration scholars toward federally gathered data that leaps some of those hurdles. He uses U.S. Bureau of Labor Statistics Current.Population Surveys, which provide data on a random sample of workers, including information on their sector of employment. We find U.S. Census data even more useful; though it comes out less frequently, it provides far larger samples. We use the Public Use Microdata Samples (PUMS) from the 1980, 1990, and 2000 Census. Each provides information on individuals in a random 5% sample of U.S. households. Information on employers is extremely limited, but we know whether the employer is a local, state, or federal government, a private firm, or a nonprofit organization. Only 51 employers are clearly identifiable: the federal government and the 50 states.2 Even restricting the sample to full-time, full-year employees - those who worked for pay at least 35 hours per week for at least 48 weeks during the previous year and were not self-employed -- we have random samples of about 2.9 million in 1980, 3.4 million in 1990, and 4.1 million in 2000. Those include 140,000 state government workers in 1980, 170,000 in 1990, and 210,000 in 2000. Since 2001, the U.S. Census has also conducted an annual American Community Survey (ACS), using similar questions to those used in the Census. Sample sizes have varied but have grown markedly in the three most recent years available. For some analyses we examine data separately for each year, but for the primary analyses we combine 2005, 2006, and 2007 ACS data to compute a "running average" for 2006. This gives us 2.5 million full-time, full-year employees for "2006," including 130,000 state government employees. 2 As the Census provides the locations of workers' residence rather than employment, we make some errors in identifying state employers for residents of one state who work for a different state's government. 10 We begin by examining how the mean, median, and modal ages of employees in the federal, state, local, and private sectors changed between 1980 and 2006.3 We also graph the age distributions in different sectors in different years, showing the percentage of workers at each individual age. To determine whether aging is affecting all states and sub-groups similarly, we compute mean ages for each of the 50 states in 1980, 1990, 2000, and 2006. Using just the 2006 data, we then compute mean ages by race, sex, educational level, and occupation. Because of the changing demographic patterns of American society and the changing hiring patterns of state governments, we expect white male state employees to be older than women and minorities. We expect managers and certain professionals to be older than workers in more entry-level positions. Though the huge sample sizes mean nearly all differences are statistically significant, we use regression analysis to test for this and also to test whether state age differences can be explained by demographic or occupational differences between their workforces. To assess the potential impact on institutional memory, diversity, and educational levels, we use more indirect methods. We assume that institutional memory is strongly correlated with work experience with a state government, which is highly correlated with the age of the employee. We compare the percentage of state employees above various age thresholds in 2006 and 1980 and estimate how long it will be before the percentage above that threshold in 2006 approaches the 1980 percentage. To assess the impact of the retirement tsunami on the diversity of the workforce, we 3 Unfortunately, the 1980 Census does not distinguish between private sector employees working for nonprofit organizations and for-profit firms. 11 compare the composition of workers in different birth cohorts in 2006, assuming that in the coming decade the state workforce will come to look more like current employees born since 1970 and less like employees born before 1940. We also examine educational levels by birth cohort and year. Analysis of State Workforce Aging State Workforce Table 1 presents a picture of aging trends in the U.S. workforce. The mean age of full-time, full-year workers rose by 4.4 years over the past quarter-century, from 38.3 in 1980 to 42.7 in 2006. As expected, it rose more rapidly in the federal service (by 5.6 years) but, contrary to our expectations, nearly as rapidly in state government (by 5.5 years). In 1980, state government employees were, on average, only 1.7 years older than the labor force as a whole and were 0.4 years younger than local government employees. By 2006, the state sector had the oldest workforce: 2.8 years older than the general workforce and 0.5 year older than local government employees. State governments had the lowest percentage of employees under 30 (10.3%) and the highest percentage 50 and over (40.4%).

Conclusions and Implications State workers are even older than federal workers, so the retirement tsunami may hit states sooner and harder than the federal service. State employees have the highest share of old workers and the lowest share of young workers, making them likely to experience the biggest increases in turnover. State governments may want to consider adopting reform initiatives similar to those the federal government is initiating. Interstate variation indicates that some states should move faster than others to confront the aging workforce problem. However, the retirement tsunami offers opportunities as well as challenges. Increasing retirements mean more hiring, giving state governments more options to make their workforces representative of their populations and to hire the mix of skills they need to match the current tasks of government. Lewis (2003) predicted that the retirement tsunami would increase both racial and gender diversity in the federal service. Our analysis paints a similar picture for the state workforce. Further, as the tasks government takes on become more complex, the professionalization of the public sector may have been limited by the stability of its workforce. More hiring can provide more opportunities to recruit better-educated employees with the particular skills governments need. The youth of employees in some professional occupations in state government sends a positive signal that states are taking advantage of opportunities to professionalize their workforces. 18 Reforming human resource management will enable state governments to manage the upcoming retirement tsunami. As the Winter Commission recommended, streamlining the hiring process will help in attracting competent employees. Paying more attention to diversity will assist the effective management of young employees who are more diverse than ever before. Additionally, state government may want to use some ideas and tips adopted by federal agencies. They can rehire old workers (Department of State), make some consortium with non-profit and for-profit companies (Department of Treasury), or utilize mentoring programs (USAID and HUD) (GAO, 2009). Acting on these reform initiatives will give an opportunity for state governments to successfully address the retirement wave while keeping its institutional knowledge.

Critically examine the implications of training/professional development for an aging public sector workforce and give an example of organization where this was implemented.