On the brink of welfare reform -- with an estimated $11 billion in additional system development and operations costs between 1993 and the turn of the century -- we need to understand the factors that contributed to the problem, and how to fix it.
GAO's findings are a serious charge against public officials in an industry that has expended $6.8 billion in federal funds between 1984 and 1992 alone, and over $1.8 billion prior to 1984. The full federal, state and local costs, if they were measurable, could easily be $10 billion or more, as federal participation is limited to 50 percent or less of overall costs up to a maximum of 90 percent on enhanced funded projects. (February 1994 GAO Report, Automated Welfare Systems: Historical Costs and Projections.)
But the issue is much broader. Unfortunately, our largely intuitive belief that automation would result in better administration of programs has not been borne out. There has been no incentive for state governments to explore and target significant business improvements even though the system planning process does require an alternatives cost/benefit analysis, which has undergone some revision since the GAO report -- Ineffective Federal Oversight Permits Costly Automated System Problems -- was released in 1992.
The 1992 report charged that weak federal oversight permitted less than optimal outcomes. Although it was out of the purview of the study, the GAO report could have provided valuable insight by exploring the critical role of local government managers in planning projects that could deliver needed local business changes -- projects capable of exploiting technology to provide long-term cost savings.
The report could have provided leadership by challenging federal approval processes which reward "no change" state projects that only appeared to be "low risk." Both state and federal staffers lack understanding about how "low risk" technology solutions most often lead to very "high risk" business outcomes.
With or without federal requirements to guide the process, however, one would expect that state government administrators more directly accountable to local government constituents would demand concrete business plans before embarking on such major expenditures. One would also expect that a strong sense of responsibility for local dollars would drive business outcomes.
ALTERNATIVES ANALYSIS
Rather than presenting a true business case, project planners most often used alternatives analysis to support a technology solutions direction. This was particularly true when federal approval agencies demanded system transfers based on the mistaken notion that development costs could be controlled if new system code wasn't rewritten in each state. Pressure to automate and pressure to transfer -- irrespective of the business case -- (and 90 percent funding) drove system development efforts. The age-old economic challenge of supply and demand overtook the welfare systems arena, as the cost of development efforts rose from a few million dollars to $200 million for a single state project in the course of 10 years -- and which could cost $1 billion for California alone.
Functional and technical requirements shadowed business requirements. Short-term benefits were easy to define; long-term costs and benefits were most often deemed unpredictable in spite of remarkable case study information available by 1987 to state project teams and federal oversight staffers. Only rarely during this period did a local government produce a planning document that could later be used to measure positive outcomes demonstrating that a system had achieved expected long-term results.
Funding approval processes focused heavily on counting widgets to create cost-savings pictures. Cost avoidance was disregarded as vaporware, demonstrating a failure of government analysts to grasp that defining a different business model would establish a new cost base, providing a natural control mechanism for long-term cost increases in program administration.
In some states, the failure to match necessary levels of higher risk to bona fide business plans before embarking on system planning has resulted in an ongoing "battle of the boxes" to supply the processing power needed to meet client service demands during the online
day. The technology problems define an agency's business model ... rather than a business model defining both technology solutions and business outcomes.
BLOCK GRANTS
An expenditure of over $20 billion during the last two decades has failed to deliver expected results. And we are starting off into unmarked territory, as states -- which have not had to be accountable for business results based on automation investments -- face the probability of block grants for social services/welfare programs.
Colleen Daly, a division director of the HHS State Systems Approvals organization, in a summer 1995 interview said, "Most are concerned about AFDC, those who built FAMIS systems in the late 80s and early 90s with enhanced federal financial participation. More than half of these systems today have reached the end of their useful life cycle. Systems professionals in these states are wondering where the funding will come from to supplement systems, replace them, and modify them to do the extraordinarily different things that welfare reform will require.
"For states that have just started or are starting now, they're unsure whether or not they'll be able to get funds out of the block grant to make the capital investment that it will take to build the systems to run the programs." (Generally speaking, the existence of a FAMIS system presupposes the existence of an integrated system that may include Medicaid registration, food stamp eligibility, and possibly Medicaid eligibility.) Daly made similar comments about states that will not have completed child welfare system development projects by the time they must manage within the capped funding of block grants.
Nickels and dimes are saved by politicized, programmatic reform, while there are hundreds of millions to be saved by designing changed, effective business models. In the same 1995 interview with HHS managers, Robin Rushton, division director of Child Support Information Systems, put the issue into excellent perspective. "In child support, we're finding too many states automating a bad process. In the past ... they automated bad processes and paper, and we see that in the system they ultimately end up with. It really works best if, before you ever start automating, you examine or reengineer your business process and rethink the whole and not fall into the trap of simply automating what is."
If automating "what is" resulted in optimal efficiency or effectiveness the issue would be very different. Unfortunately, there invariably is a mismatch that adds operational inefficiency and cost.
CONTROLS NEEDED
If states are left to financially support existing and future automated systems within block grant funding caps, they will need to set their own standards for viable automation efforts. There are three levels of control that must be exercised to reverse this trend of throwing good money after bad.
The first is at the state policy level as block grant funded operations are defined. The business model should be drawn up before the programmatic alternatives are explored.
Several questions must be posed if the GAO's report findings are to be changed. How much money is there? What administrative business model will cost the least and retain the greatest level of funding for direct benefits as opposed to direct services? How can technology support the least-cost administrative business model in the most sophisticated, seamless manner?
Legislatures must provide leadership to entrenched bureaucracies by rewarding those career officials who are successful in eliminating programmatic boundaries and turf issues. Commitment to bureaucratic process must be replaced by a commitment and accountability to taxpayers for delivery of measurable outcomes.
The second level of control is at the state and local program administration level where responsible, accountable administrators must be retrained in modern business planning methodologies, and trained in how to apply those methodologies to a government environment. From some point forward, political programmatic initiatives should not drive government management processes, rather programmatic initiatives must be easily integrated within an effective business model and within a least-cost implementation framework. The current excuse that we can't determine how we want to do business because we don't know what the programs will look like must be removed. Automated system development or enhancement solutions -- which cannot exhibit a match against a concrete business plan and explicitly state expected results, along with timelines for achievement -- should not be funded.
Past federal funding has enabled long-term expenditures that states would not logically make if the largest burden of cost had to be born locally. There must be heightened legislative demand for measurable performance results and individual administrative accountability.
At the third level, federal block grant funding must be just that -- any retention of a federal oversight or regulatory bureaucracy for state systems will perpetuate waste. The federal bureaucracy is too far removed from the business and economic demands of the local constituency to provide guidance, and federal oversight has been counterproductive.
Across the board, at every level, government must learn to cut its losses.
The lessons learned in case history after case history have taught us -- and the GAO reports have confirmed -- that system development is the tip of the iceberg in taxpayer cost. We must assume management responsibility for scrapping technology solutions that increase rather than decrease administrative costs. Block grants may offer automation success a strange bedfellow as states face the cold hard facts of capped funding.
Under block grant funding, maintaining the status quo in government operations and systems management will break the bank. Government leaders are challenged to remember that "opportunity knocks once," but someone must open the door. Who's going to be first? From the taxpayer's perspective, this truly is a "time for heroes."
Rita Kidd provides commentary and analysis on human services and other large-scale automation efforts in state and local government. Her experiences as former deputy director of Merced County, Calif.'s Human Services Agency, and as director of MAGIC -- Merced County's welfare automation system -- provide valuable insight into the complexities of welfare automation and approaching welfare reform. She is now a government reengineering consultant residing in Cathey's Valley, Calif. Her opinions are her own, and not necessarily those of GT or its editors. E-mail:
