The world is made up of causes and effects. Hurricanes cause storm surges. Hitting a cue ball hard into a break causes pool balls to scatter. A bad earnings report causes a company’s stock to go down. And so it goes in business, sports, life and cities. High crime rates cause visitors to stay away from a city. High taxes slow development. High college acceptance rates attract students to schools. This is what economists spend their time thinking about.
Most people think about these causes and effects abstractly. Common sense tells them that one thing ought to affect another. For instance, an after school program keeps kids off the streets and therefore should reduce the crime rate because kids on the streets sometimes commit crimes. Another example might be, making a city’s inspections process less expensive to lower development costs and stimulate investment. Or perhaps, opening a new museum will increase tourism.
Most people are comfortable making statements like the above, but generally don’t know the details. For instance, they can’t answer questions like: If we spend $1,000,000 on foot patrols how many FBI index crimes will be avoided? Or, if we lower inspections fees by 50%, how much incremental investment should the city expect to see over the next 5 years? These are fair and important questions. Most citizens can come nowhere close to answering these types of questions, and that would be OK but sadly, most policy-makers in a city like Trenton can’t answer them either.
So how can normal citizens get better at thinking through the policy issues that face us every day?
Without researching every policy assertion that’s ever made, how can we begin to really understand causes and effects?
We make better choices by knowing whether a policy has a 1st order or 2nd order effect and whether the effect is strong or weak. Of course we need to start with clarity on our goals (investment, crime, education, population, income). But after being clear on goals we must carefully consider causes and effects so we can begin to decide whether policy assertions are important. This kind of thinking is often called “systems” thinking and is used to better understand complex things, like cities.
There’s a difference between 1st and 2nd order effects
In pool, when the cue ball strikes another ball and knocks it directly into a pocket, we call that a 1st order effect. One thing caused another. However that same pool shot may have left the cue ball well positioned to allow the player to sink the next ball. That’s a 2nd order effect. The difference is that in order for the good “leave” to have happened, many more effects of physics had to take place over and beyond the just hitting the first ball in. The cue ball had to be deflected just so, the spin had to be just right and perhaps the cue ball needed to bounce off the bumper with just the right angle. The good “leave”, assuming it was intentional, had a much less likely chance of success than hitting the first ball in.
And so it is with city policies. An afterschool program will most definitely get some kids off the street. Getting kids off the streets is a 1st order effect and can be measured fairly simply. It’s the number of kids in the program minus the percentage of those kids who would have otherwise stayed at home or in the library. For instance: of the 100 kids in the after school program we might say 40 of them would have been home. So the program got 60 kids off the street.
But how does an after-school program affect crime? It’s not likely that a kid staying home would cause a crime. But what about the 60 who would have hung out on street corners. It’s a bit harder to say because crime reduction is a 2nd order effect. For example, not all of those 60 kids would have ever committed a crime. Of the several who might be inclined to commit a crime they might do it when they weren’t in the after-school program. But then again, maybe the program has a long term effect on the child, or maybe it doesn’t. As you can see, the 2nd order effects begin to get murky. This is why sophisticated policy makers don’t depend on them and often point to 2nd order effects as “potential side benefits”.
In Trenton, we shouldn’t base our important policy decisions on 2nd order side effects.
Strong vs. weak effects and the importance of context
Even when causes and effects are 1st order, the linkage between the two can be weak. For instance in buying a used car, high mileage may not dissuade you from buying it. This is a 1st order effect but not a strong one because you’ve already decided you could accept a few miles on the car. However, dented side panels may just completely turn you off. The big dents might be a strong 1st order effect and keep you from buying the car.
It’s the same with public policy. Let’s return to inspection fees for a new home. Let’s say we want to stimulate growth by reducing the fee from $1000 to $300. That’s a big drop. And because it directly affects the price of the house, it’s a 1st order effect. However, that $700 drop in cost is fairly small in comparison to the $300,000 that you’ll eventually spend on the house. Other things like lumber, labor, land and property taxes easily dwarf the inspection cost. So while the reduction in inspection fees may be annoying to the builder, it has a weak effect (though 1st order) on the eventual buyer.
2nd order effects can be weak and strong as well. For instance, we can imagine a school retention program that lowers the high school drop-out rate. This program might have a good 1st order effect on education but also a 2nd order effect on crime reduction. That 2nd order effect might be considered strong because we know there’s such a high correlation between high school graduation and likelihood of committing a crime in the future. Compare that to an after-school basketball program which should have a 2nd order effect on crime reduction (as we discussed above) but that effect may be weak. Certainly the research and evidence linking graduation to crime reduction is stronger than that linking basketball to crime reduction. That’s not to say there’s no effect, it’s just not likely to be as strong.
The cause and effect of crime also varies widely. Economists have shown that each incremental index crime in a city leads to one person moving away. However, the rate of emigration is 5 times higher for high income people and 3 times higher for families with children. Poor, single people are much less likely to move away due to a high crime rate. Therefore we can say that a high crime rate has a strong effect on high income people leaving a city but a weak effect on the poor leaving (likely because they have fewer choices).
Just understanding this differences in the effects of crime, even in the abstract, should have a profound impact on how we think about policy in a city like Trenton. Sadly, you’ve never heard a government official make the above distinction.
It might be good to focus on strong 1st order effects rather than weak 2nd order ones.
In the world of policy making and particularly in a cash-strapped city like Trenton, we need to make hard choices. We don’t have either the money or the man-power to do everything we’d like. So it’s important for citizens to lobby for the most important policies and for government officials and activists to help clarify 1st and 2nd effects and strong vs. weak linkages.
We can use crime reduction as an example of a good objective. Criminologists know that high rates of incarceration have a beneficial effect on the crime rate (most people get this). There is a strong 1st order cause and effect between building good cases against criminals that lead to long sentences. On the other hand, we may spend the same money we would have spent on an extra detective on a mentoring program. The mentoring might have a 2nd order effect on crime reduction and likely a weak one at best.
When we talk about programs and policies in Trenton politics, we need to keep these things straight and always keep our core goals in mind as well as cost-benefit.
Policies that have multiple 1st and 2nd order effects are generally more impactful than others
Finally we should remember that sometimes policies can have multiple effects. You’d likely trade a $1,000,000 program to reduce crime that has single strong and effect on the crime goal, for a $1,000,000 program to stimulate development that might have a strong 1st order effects on the economic growth goal, a strong 2nd order effect on the crime goal and a weak 2nd order effect on the education goal. Some policies give us broader “bang for the buck”.
Policies that positively affect multiple goals in Trenton (investment, crime, education, population growth and income growth) will not only strengthen the city and stretch our dollars, but will find broader political support.
Every minute of every day, Trentonians have policy discussion on Facebook, at barber shops, in civic association meetings, over drinks and at City Hall. We discuss crime, trash pick-up, taxes, parades and any number of topics. It’s important for Trentonians to move past sentimentality and misguided assumptions in our discourse. We need to get on the same page. To do that, not only do we need shared goals, but we need a common vernacular for discussing policy. To the extent we can begin to discipline our thinking by keeping our goals clear and then breaking causes and effects down into 1st and 2nd order and then strong vs. weak, we’ll have a more constructive civic dialogue.
Note: I wrote this article for my blog 2 weeks ago, before the TESC deal for Glen Cairn Arms came up and was having it edited. I had no way of knowing we would be having a important policy debate about this very subject. I held off publishing it in favor of reporting on and providing thoughts about the proposed TESC deal. However now is a good time to start talking about causes and effects in policy discussion.