Cause-and-effect is underrated

Sometimes I’ll write 3,000 words criticizing some proposed policy, but I think many times my argument just boils down to “You should be more explicit about what you think this will accomplish and walk through exactly how you think it will work.”

The city’s Equity and Inclusion commission met on November 3, 2022, and they spent a fair amount of time discussing racially restrictive covenants and how to get rid of them. These covenants flourished in the early- to mid-20th century and, at one time, prevented black families from owning, leasing, or occupying many homes in Wauwatosa. I’ll come back to them.

But first, a few examples.

A.

One way to improve the world is to find something that could actively make people’s lives better and do it. Alternatively, one could notice something that is actively making their lives worse and try to stop it.

Lots of people in Africa get malaria and many die from it. One thing you could do is give them anti-malaria bed nets that they can sleep under to prevent mosquitos with malaria from feeding on them. The result is they don’t get sick and they don’t die. Alternatively you could develop a malaria vaccine, send it to Africa, and make sure everyone gets a shot. Same outcome: they don’t get malaria, don’t get sick, and don’t die. There is a clear mechanism of action here—it is easy to see the relationship between mosquitos, people, and sickness, and how they interact. You could even make a diagram to explain it.

It also implies potential interventions. You could imagine just getting rid of all the mosquitos so that the gray oval above never even appears. Or perhaps genetically engineering mosquitos so they can’t carry malaria. You could imagine eliminating that arrow between the gray oval and the green/yellow oval. A vaccine wouldn’t eliminate mosquitos and wouldn’t keep them from carrying malaria, but they wouldn’t really be able to infect an individual. Or you could imagine therapies or treatments that block the arrow between the green/yellow oval and the blue oval—someone might get malaria but there are treatments that make them less likely to get sick or die from it.

Sometimes causal mechanisms aren’t always so straightforward and easy to witness. There might be many causes each with individually very small effects and they might act over years or decades instead of days or weeks.

Lead is a toxic heavy metal whose ingestion, especially at young ages, impedes physical and mental development, lowers IQ, increases impulsiveness and criminality, and has a host of other long-lasting effects. This wasn’t always obvious. The Romans used a crystallized form of lead (lead acetate) as a sweetener for their food. In the United States, we used it for water pipes and for house paint and as an additive to gasoline. But there were big gains from getting rid of it. Many studies have linked the gradual elimination of leaded gasoline in the 70s to declining crime rates 20 years later as children with lower exposure become less impulsive adults.

So you could imagine removing this lead from the soil, the walls, and the pipes and measuring people’s blood lead levels or the number of cases of lead poisoning to observe the effect in the short-term, and years or decades in the future looking at things like school performance, rates of criminality, or even wealth. Though the steps between the cause—lead—and (one of) the things you care about—crime—are more numerous and complicated, there is at least a theory of how it all works.

However, as the number of interacting effects increases, the relationships between all them tends to become noisier, more uncertain, and harder to measure. Does the reduction in lead contamination account for 7 percent of the subsequent reduction in crime in the United States, or 28 percent? This suggests that root causes, even if identifiable in the abstract, can be hard to pin down and targeting them might be exceedingly difficult or expensive. In between the ingestion of lead in childhood and the commission of crime decades later are a literally infinite range of mediating factors that might also be increasing or decreasing the likelihood of committing a crime. And while it’s hard to think of too many downsides to eliminating lead additives (maybe we’ll find out that the ethanol we replaced lead tetraethyl with as a gasoline additive is actually worse in some way) those root causes might have lots of other effects besides the one you care about.

Nevertheless, there is a certain clarity of thought here.

B.

Some other examples, hopefully confusing in light of the above.

When Ald. Lowe proposes to ban ‘no knock warrants’ in the City of Wauwatosa, he cites high profile incidents in other parts of the country where the police raided the wrong house or injured or killed someone they didn’t intend. But then during discussions between the Chief of the WPD and the Common Council, you realize that the police department almost never executes no knock warrants and that the constraints Ald. Lowe is suggesting be placed on their use are actually already the stated policy of the police department. It’s not clear what additional effect he imagines his proposed ban would have, why he continues to bring it forward for consideration, and why he continues to be upset when it is rejected. He does mention something like, “it will make people feel safe and more willing to move to Wauwatosa.” But how does this actually work?

Or this past spring, former-Ald. Nancy Welch proposed an ordinance to ban pet stores from selling cats and dogs in the City of Wauwatosa in order to prevent the proliferation of puppy mills. The policy’s proposed mechanism for accomplishing this—that prohibiting pet stores from selling puppies reduces the pet store’s demand for them from puppy mills and, therefore, the supply of puppy mill puppies would shrink in response—is a pretty straightforward economic argument. And yet, there were no current pet stores in the city of Wauwatosa that sold puppies nor were there any plans to open one that the ordinance would have prevented.

More recently there was a proposal from Ald. Meindl and Lowe to “decriminalize” marijuana possession. In this case, the term “decriminalize” implied a mechanism of action where there was none since the City of Wauwatosa, as a local municipality, has no authority to change those laws which are set by the state and the federal government. Ald. Lowe suggested that it would be a great source of tax revenue, but this too implied a cause-and-effect that had no clear mechanism of action, since he wasn’t actually proposing to tax the sale of marijuana and also because local municipalities in Wisconsin have no authority to do something like that anyway.

Also in September, Ald. Lowe proposed a gun buy back program where gun owners could turn in their old or unwanted firearms at a designated time and location in exchange for some type of gift card. When I wrote about it before, I said:

What his presentation did not include, perhaps considering it to be self-evident […] was much explanation of what the goal of this effort was and why a gun buy-back program would best accomplish it. He did assert that it would be “a good idea” and that “we know crime is going up,” but never specifies what types of crimes he’s referring to or why the city buying back old guns from residents would reduce them.

Further, in the past gun buy back programs have been popular but more recent research suggests they have very minimal effects.

Last week I mentioned Public Health 3.0, a new framework for action proposed by the Public Health Department during their recent budget meeting. Not content with restaurant cleanliness inspections or vaccination drives, Public Health 3.0 will look beyond serving individuals in a one-on-one capacity and begin to address the larger systemic issues and “root causes” that affect people’s health. According to the Director of Public Health, the Public Health Strategist they intend to hire “would really be looking at equity from a systems level. And so that position would look at not necessarily just STIs but sort of these things that affect people’s health decisions.”

Ald. Fuerst then asks what the term ‘health equity’ means.

Like I talked about before, we’ve done a lot of one-on-one individual health work. That’s great. It helps one person make healthier choices. But there are all of these—we call them social determinants of health—so someone’s education can affect their health outcome, the job they have, the neighborhood they live in, the house they live in—all those other things affect their health and their ability to have the same opportunities to have positive health outcomes. […] So looking at those disparities and addressing the systems that affect them.

These effects likely exist. Having a crappy job with poor benefits that you’re constantly worried you’ll be fired from probably does lead you to eat worse food and feel stressed and perhaps—decades hence—die a few weeks, months, or even years earlier than you might otherwise have. It’s just unclear to me what a Public Health Strategist can do about this and how one would evaluate whether they’ve succeeded.

The city does have a Development Director, and you could imagine they might, if the opportunity existed, try to engineer a deal to bring a large employer to the area. But I don’t think the motivation for this would be long-term public health implications because there are other more short-term effects—people can get good paying jobs near communities they can afford to live in—that are much easier to identify, evaluate, and which the people affected will consider more relevant anyway.

It’s the systems, people! (Source)

The Equity and Inclusion commission’s meeting on November 3rd was dedicated to a discussion on eliminating racially restrictive covenants from home deeds in Wauwatosa. Pretty much everyone considers these a sordid and embarrassing holdover from a much more prejudiced past, and the city would like to do something about it. As Mayor McBride said at an earlier meeting, removing these is symbolically important and “akin to pulling down a statue of a Confederate general.”

I am not a complete materialist. Things like increasing income, reducing lead poisoning, increasing life expectancy, and discouraging crime aren’t the only things that matter, and it seems obvious to me that ideas and symbols are important and can influence the world in significant but hard-to-measure ways. Sometimes small, more poorly resourced armies defeat larger, better trained ones for reasons that are inexplicable in terms of the number of soldiers and the sophistication of the weapons they can field. Sometimes principles and a rousing speech can inspire people to act when they otherwise might not have.

However, I feel like symbols and ideas are most useful, from a policy perspective, for the ways they can be used to coordinate some kind of tangible action and observable effect in reality. And that they have the biggest effects when they’re highly visible and also costly for the people who hold them. But the discussion on racially restrictive covenants revealed that:

• Racially restrictive covenants for many decades have been legally unenforceable. The text might be in the deed but it can’t do anything anymore.

• Also as a matter of law, the text of these covenants can’t legally be published. When someone requests a title report during the sale of the home, the covenant is listed, but a disclaimer also gets printed that says something like, “This commitment does not republish any covenant, condition, restriction, or limitation contained in any document to the extent that it violates state or federal law based on race, color, religion, sex, sexual orientation…” etc.

• The City of Wauwatosa can’t legally do anything to remove these covenants on their own. All they can do is encourage individual homeowners to take some private action to have them changed.

If this effort is a symbolic act akin to pulling down a Confederate statue, it’s a Confederate statue that only you can see and then only if you look really hard. And when you take it down, no one will know because they won’t see you do it and it’s presence or absence has no observable impact on the world beside your personal knowledge of the action you have taken.

C.

All the examples above illustrate slightly different failure modes as well as some implied lessons.

→ Some policies or actions just logically cannot work in the way they’re described, and we should think about them a little more.

This seems to describe (3) where as soon as you articulate a theory of how this particular policy is supposed to function, you realize, “Wait, no, that can’t possibly work. Why are you even…” There’s no way an ordinance proposing to reduce the civil fine for marijuana possession will create tax revenues for the city.

→ Some policies or actions might work in one place but won’t work here because the local conditions are different.

This seems to describe (1) and (2) where you have a clear mechanism of action and an effect that is both plausible and real but for various reasons will not work in the local context it’s applied in. Banning no knock warrants won’t do much if the police don’t perform them in the first place. The anti-puppy mill ordinance won’t do much if we have no pet stores that sell puppies.

→ Some policies or actions that plausibly might work in the way they’ve been described might turn out to be ineffective after further research is conducted, and we should take this into account before implementing them.

This seems to describe (4).

→ Some policies or actions require a level of resources and coordination that aren’t really available and have impacts that are really difficult to evaluate, and we should take this into account before implementing them.

This seems to describe (5) which has a thoughtful, sober consideration of a real and complex problem but where it’s unclear local government has the capacity or leverage to really do anything meaningful. Also expending lots of resources to attack “root causes” based on tenuous, uncertain effects decades down the line that are mediated by a dozen other things along the way is, I think, just a harder sell to the public.

→ Even symbolic policies or actions should have some reasonable mechanism for creating an observable impact that we care about and that everyone can agree has occurred if it does occur.

This seems to describe (6), a symbolic act that doesn’t seem costly or public enough to symbolize much. My best guess for the way it’s supposed to work is that the city does a lot of widely advertised public education and encouragement to private homeowners to excise these restrictive covenants from their deeds, some number of them do so, feel personally good about it, and perhaps tell others? One member of the Equity and Inclusion mentioned giving people a sticker or a badge that they could display when they go to sell your house. Maybe there’s a follow-up newspaper article that talks positively about the whole thing? Maybe this encourages the state to take some kind of legislative action? Maybe.

D.

Of course this assumes that people are trying to do things that are effective and useful. People can have other goals. They might just be self-aggrandizing or seeking publicity and acclaim. They might be trying to consolidate power. They might simply want to show others that they care. And there may just be things going on behind the scenes that are difficult for the public to see.

But if we do assume that people want to do things for the community that are effective and useful, a final kind of meta-point is that doing things like the above instead takes resources from actions and initiatives that might be more transparent, more straightforward, and more likely to fix the things we care about.