During the recent Supreme Court proceedings on the constitutionality of the Affordable Care Act, the notion of "coercive" federalism kept popping up. It is the idea that if states don't adopt this policy or that, then the feds are going to withhold funding in areas ranging from education to transportation. It's the classic top-down, command-and-control approach to intergovernmental relations that has engendered so much conflict between the federal government, states and localities.

But I have a friend who works for the federal government who has long argued for a different model of federalism: One that is cooperative, not coercive. With cooperative federalism, the feds become full partners in an adult relationship that involves the federal government acting as consultant to states, not overlords.

In no area of public policy could such an approach be more beneficial in the long run than in human services. If the feds could position themselves as a collector of and repository for good data on outcomes in relation to human services programs, my friend's argument goes, then we could start to build out a very robust database on what works and doesn't, a system that could be accessed by states and localities interested in learning about best practices and the circumstances under which those practices seem to get the best results.

The tough nut to crack here has always been longevity studies. In human services, what impact this intervention or that actually has on a family or individual often takes years, if not decades, to play out. That leads to a perverse psychology, especially in a political environment defined by two-, four- and six-year terms. The leadership asks itself: Why start a longevity study now if I'm not going to be around when it starts to tell me something important?

A recent report by the firm Mathematica Policy Research on the feasibility of longevity studies in the whole area of disability services has an answer to that question. The report was commissioned by the U.S. Department of Health and Human Service's Office for Planning and Evaluation (OPE). The fundamental rationale for taking up the subject was that OPE is "the principal advisor to the Secretary of the Department of Health and Human Services on policy and development issues, and is responsible for major activities in the areas of legislative and budget development, strategic planning, policy research and evaluation and economic analysis" (the emphasis is mine).

In other words, OPE is supposed to help states and localities figure out what works and whether it's cost effective. The report is an important start as it contains exhaustive and thorough information on what surveys currently exist (and what promise they hold for yielding good information on program effectiveness), what studies are in the works and what the limitations of current data are (e.g., small sample sizes, poor quality data, untimely data, inadequate longitudinal data and so forth).

Data quality, timeliness and depth have long vexed states and localities as has a fundamental understanding of the true meaning of the word "outcome." But as public officials and vendors develop a clearer understanding of all facets of data that help human services systems operate more efficiently, information technology will no doubt continue to evolve and improve to meet that knowledge deficit.

The standing paradox, of course, is that the time you need good longitudinal data is always right now. That's why it is so important to start now, so that practitioners five, 10, 15 and 20 years out will have good, actionable information on program efficiency and effectiveness. Again, it's that old story, one that doesn't resonate well in a political environment that says if we don't invest in this now, we're just setting ourselves up for failure down the road. It's well worth exploring and building partnerships with agencies that get it, like OPE, so that 2020 doesn't become yet another milestone year where failure of foresight again resigns us to what we might have done better in human services in hindsight.