Jay Greene has responded to my review of the new Brookings Institution school choice report which he co‐​authored, raising a crucial issue for the education policy and research communities. Jay points out that the report is a work of realpolitik rather than scholarship, and as such contends that it must find a compromise between the policies best supported by the evidence and those that have a real chance of being implemented. He makes the related argument that incrementalism is the only realistic path to success.


I agree with Jay that it’s good for analysts to find ways of improving current policy even when the ideal policies are not politically feasible. But these realpolitik recommendations must be clearly distinguished from the ideal policies themselves. Analysts should report both viable compromise reforms AND ideal policies, explaining to policymakers the likely costs and risks associated with the compromises–the reasons why they are inferior. Failing to do this leads to two serious problems:


First, presenting only the compromises robs the public and its elected representatives of crucial information, making it more difficult to build support for the ideal policies and leading to guilt by association when the compromise policies prove disappointing for reasons that should have been — but were not — clearly laid out in advance.


Second, when analysts don’t present their ideal policies and the evidence (if any) on which they are based, there is no way for the public or policymakers to judge the wisdom of their realpolitik compromise recommendations. This is particularly problematic when the analysts’ recommendations conflict with what the available evidence shows to be ideal policy.


As to the need for incrementalism in U.S. policy reform, the evidence is not entirely one‐​sided. The Emancipation Proclamation did not give slaves a 50 percent share in themselves, rising gradually to 100 percent over time. When women won the franchise, it was not at a discounted rate — one female vote equal to 1/3 or 1/2 of a male vote. They won the right to vote outright. Prohibition was not undone gradually, with beverage categories being re‐​legalized in order of alcohol content. I’m sure we could think of other major policy shifts in U.S. history that were not incremental.


In all of the above cases, major social movements were necessary to win the day, and if scholars and advocates who knew better had championed only half‐​measures instead of the policies they knew to be right, it surely would have delayed the eventual victories. Scholars who know what kind of school choice is necessary to best serve children should clearly advocate such policies, especially in any context in which they also offer any interim recommendations they deem more politically feasible. 


And even if we grant, for the sake of argument, that all school choice policies must be incremental, there are incremental policies already in existence that are highly consistent with ideal policy. Existing scholarship donation tax credits such as those in PA, FL, RI, etc., and personal use education tax credits such as those of Illinois and Iowa, are expanding organically over time. Eventually, as that expansion continues, they could be combined and thus ensure universal access to the education marketplace without needing to impose regulations on private schools that the research shows to be intrusive and counterproductive. By contrast, it is hard to see how introducing federal regulation of virtual schools (a Brookings Report recommendation) moves us close in the direction of the minimally regulated parent‐​driven markets supported by the evidence.


So, yes, let’s be realistic in our policy recommendations, but let’s also be clear about the ideal policies indicated by the empirical evidence, so that policymakers and the public hear a consistent message about where we need to go.