Defending Conservatism, Part 3: In Favor of Decentralization and Subsidiarity

For the sake of argument (see Part 2), we’re accepting that we don’t fully understand the world around us. One very straightforward implication is that we should prefer the local to the national. A single solution at the national level may well be wrong because of our knowledge problems and our issues understanding policy reality. Getting policy wrong at a national level affects us all, and often locks us into bad outcomes. Preferring local policies is better: some of those will likely work better than others, and states and localities can gravitate, over time, to models that prove more effective. Moreover, the calamities are more contained, and there are ready-made solutions to consider in response, with actual data and a track record worth evaluating.

Let’s take health care policy as an example. Three fictitious states are trying to figure out how best to provide health insurance in their states:

• Columbia: Columbia adopts a “three-legged stool” model and provides subsidies for the purchase of private insurance: all individuals must have a policy, either through their employer or via purchase on an individual market. Insurance companies cannot deny coverage for any reason, and cannot charge sicker people larger premiums or deductibles.
• New Eagle: New Eagle opts for a market-oriented “Health Savings Account” model, providing tax credits to all of its citizens for spending on routine services, and insurance coverage for more unpredictable expenses. The state sets up a “reinsurance” system for catastrophic expenses, where insurance companies can have the state government partially cover patient expenses above a certain amount.
• Cascadia: Cascadia goes for a “single-payer” model. Insurance companies exist only for their administrative expertise, but ultimately, all reimbursements come from the state government.

I have my suspicions about which state model would work best. But the reality is that we’re just guessing until we try the models out. And even if we do get data, no two states are exactly alike. If New Eagle’s program succeeds, that does not necessarily suggest that it is the best approach; there may have been confounding factors that were more relevant. It is merely a data point in favor of one position.

The alternative is that the federal government steps in, makes its decision about which model is most effective--or decides to “solve” the issue by hammering out perverse compromises through a challenging political environment--and locks us into whatever solution results. If the federal government chooses poorly, or if the political dealings result in less-than-ideal outcomes, we are stuck with a bad model, complete with the backing of well-funded, entrenched stakeholders.

Decentralization, in contrast, gives us the opportunity for policy experimentation. This is vital in complex domains, because otherwise, we are essentially hoping that the one answer to the policy question we come up with happens to be the best of all possible worlds. The odds of this are not great in a complex domain. At the government level, decentralization points towards federalism; let individual states have discretion to try new things. But more broadly, decentralization suggests that we should look towards markets for new models and new approaches to complex problems. Thousands of decentralized actors will likely create better models for the 21st century than the wisest government bureaucrats; they have many more opportunities to fail, and will recognize their failures quickly. Bureaucracy is much slower.

The only situations where we should prefer a national solution are ones where there are dramatic risks of externalities. In those situations the risk of broad errors and contagion are surpassed by the inability of a local solution to remain a local solution. This is typically a question of prudential balance: one can make an attenuated externalities argument about most issues. For example, education: if students in New Eagle are worse than students in Cascadia, then we all suffer due to the loss of potential human capital. There is certainly a claim worth considering here, but we should not compare our “bad New Eagle” example against a beau ideal; we need to compare it against the reality of what we can and cannot do.

Environmental regulations are one such area where externalities are relevant and immediate. (If New Eagle wants strict air quality standards, and Columbia doesn’t, then New Eagle will not actually be able to implement their policies because of the prevailing winds.) Likewise with climate change: New Eagle dramatically reducing its emissions is practically meaningless, in the overall scheme of things, if other states and countries do not. In other areas, though, like transportation, education, and health care policy, externalities are much less relevant, and as such we should be open to broad and bold experimentation at the local level.

Moving towards a locality-driven approach means that we should become more comfortable with regional differences in our country: they’re not necessarily failures; they’re learning experiences, and opportunities to grow.

Decentralization need not be entirely limited to the states, though. In general, subsidiarity is the way to go: drive the decision-making to as low a level of government or organization as it can go. Encourage people to get involved in their own communities, rather than focusing exclusively on the activities of the federal government. The more people are involved, the more they can take ownership of their lives and society, rather than feeling voiceless when pitted against a cold bureaucracy. Most critically, mistakes are less contagious mistakes this way.

One final point here: not all of this needs to be “the government.” Republican Senator Ben Sasse often cites “Rotary Clubs” as the foundation for American civil society. I’ve never lived in a place with an active Rotary Club that had a real visible presence in a community, but I envy that.

People tend to be more charitable when dealing with the local than with the national or international. Having nongovernmental organizations that can step in and provide stability for struggling kids, and public meals for low-income families, and toy drives, and all of the other things that benefit the less fortunate is a must. These are intrinsic goods, and they help us achieve meaning in ways that simply sending a few extra bucks a year in taxes to Washington DC will ever do. On top of that, we shouldn't reject a state-sponsored safety net, but we should steer it towards the local when and where we can.