What values underlie the decisions of experts and politicians? How do these differences affect their policy decisions? Robert Behn, a Harvard scholar who works on understanding the challenges of improving performance of public agencies, wrote a famous essay called “Policy Analysis and Policy Politics” which begins to answer these questions. There, Behn argues that policy analysts (i.e. “experts” – often bureaucrats, but also academics and think-tankers) and politicians see the world in very different lights. Lately I’ve been thinking a lot about the differences in the values and attitudes of politicians and policy experts. These differences have major implications downstream – that is, for the types of policies ultimately adopted. Throughout this post I’ll use Behn’s ideas as a jumping-off point so a full-reference is provided at the end of the post.

The first difference has to do with how these two groups measure and understand policy “success.” These measurements and projections will have major effects on policy selection and implementation. Policy analysts tend to consider the outputs of programs (marginal increases in military defense capability, provision of services, flood control, etc.) compared to the costs of implementing those programs (development of new missile systems, construction or expansion of a social welfare agency, building infrastructure such as levees and spillways, etc.). In short, “output” is the amount of a good that is obtained via a given policy – economists would call this the “net social benefit”. Thus analysts are concerned with mathematical, economic analyses of policy alternatives in the traditional cost-benefit mode, and seek to maximize “net social benefit.”

Politicians on the other hand, tend to focus on “inputs”, especially how many jobs/how much money will flow through their constituencies in the operation of a policy. Put differently, politicians are primarily concerned with the distribution of costs and benefits – that is, who receives the benefits and who bears the costs of a given policy – particularly within their constituency. Politicians must consider the political feasibility of a policy, which includes a careful consideration of what interest groups or lobbies might mobilize for or against that policy. Thus politicians consider policy alternatives in light of their political feasibility rather than their economic efficiency.

The differences in these measures are fairly obvious. Analysts will, in theory, advocate for the policy that they estimate will provide the most benefit to the greatest number of people and at the lowest cost. Politicians, in contrast, will advocate for policies that benefit their constituents regardless of the costs to other parties. For decades the different emphases of politicians and analysts have led many to advocate for a depoliticized policy-making process – who wouldn’t support more efficiency and less influence from organized interest groups? In theory, turning policy-making power over to bureaucrats should have the effect of reducing the influence of rent-seeking political interests, increase policy efficiency, and make governance more economical. In theory.

In reality, however, increasing analyst – often bureaucratic – discretion is not all rosey. Recall that analysts’ economic methods emphasize “net social benefit”, the fundamental question being: does the projected benefit outweigh the projected cost – and if so, by how much? This criterion ignores who gets the benefits and who pays the costs, and this can be a very important question in policy-making. Even very efficient policies can be problematic.

I’ll illustrate with a hypothetical example – though it’s important to remember that this hypothetical is a significant oversimplification, still it remains instructive for helping understand the differences in values underlying expert decisions and political decisions.

Imagine an agency is considering a flood protection policy which, they hope, will protect a large portion of the nation from floods, a policy they value at $1,000,000 of “public benefit”. This policy requires land – land adjacent to the river, specifically. The agency intends to “store” water on this land in times of flood (it would accomplish this by allowing water over or through the levees and thus siphoning it off the river). Now, the agency has to acquire this land somehow. It could purchase the land, it could rent or lease the land in years that it anticipated possible flooding, it could use the land without asking, or something else altogether. Each of these options has a cost associated with it: land is expensive, in fact, it would cost the agency $700,000 (given the $1,000,000 anticipated benefit, this option would yield a “net social benefit” of $300,000, in the analysts’ estimation); leasing or renting the land would be far cheaper, only $100,000, yielding a much higher net social benefit, but has its own problems – renting or leasing would incur the cost each time a flood threatened, renting might be too slow a process, etc. Simply using the land without asking would be free, however, netting a social benefit of $1,000,000. Basing its decision solely on economics, the agency ought to simply use the land without asking – spend no money and receive all the benefit.

This policy, however, would abuse the property rights of the people who own the land around the river. In essence, those property owners would bear the full costs of the agency’s policy and would receive none of the benefit. The distribution of the costs (onto the heads of a small number of property owners unfortunate enough to be located adjacent to the river) and benefits (everyone who lives in river’s floodplain) may not enter in to the agency’s calculations – or may even appear from the policy perspective to be effective or desirable.

In Behn’s terms the distribution of cost and benefit is a political question, not a policy question. That is, it is up to the political process to protect the interests of individuals and groups. In our political system, this typically amounts to representatives protecting interests among their constituents. Indeed, we would hope that politicians who represent the landowners near the river would stand up for their constituents’ rights before the agency could adopt such a policy. The central point is this: there are important differences in the fundamental values that underlay the policy decisions of experts and politicians.

Even if the above hypothetical represents an argument ad absurdum, it still conveys the appropriate point: agencies – and the bureaucrats that staff them – often do not consider the distribution of costs of their policies. And sometimes the costs of policies – whether monetary or otherwise – fall disproportionately on some segment of the population. Indeed, even when agencies attempt to consider the interests of parties affected by a proposed policy, they do not always do so correctly. That is, bureaucrats sometimes err in making assumptions about the wants and needs of parties with which they are not familiar. For these reason, it is critical that politics be allowed to operate in the policy-making and policy-implementing processes. To be clear, I am not arguing that politics is a panacea for the policy-making process. We need look no further than today’s headlines to see the dysfunction in contemporary federal governance. The larger point I want to make is that “de-politicizing” – that is, bureaucratizing – policy-making is not the cure-all that many commentators seem to suggest.

Robert D. Behn. 1981. “Policy Analysis and Policy Politics.” Policy Analysis 7(2): 199-226.

The article was reprinted in The Science of Public Policy, (ed. by T. Miyakawa), and can be found here.