Sunday, March 15, 2015

Solving the problem we think we know rather than the problem we actually have

One of the issues associated with arguing about complex, dynamic, self-regulating systems is that every outcome has multi-causal, inter-related roots. Contributors to poverty include such factors as physical ability (health), mental ability (IQ), mental health, education attainment, class, religion, geographic location, temporal location, language, economic cycle, government policies, government structure, cultural norms, personal values, personal behaviors, familial structure, technological innovation, etc. Seizing on any one of these causes can yield a productive discussion and argument both about the size effect of the cause, the causal mechanisms (for example, does education attainment correlate with positive life outcomes because you actually know more than others, or is it that higher education trains you in more alacritous thinking, or is it that higher education requires desirable attributes such as curiosity and perseverance, or is it that education attainment works as a signaling mechanism, or something else?), the quality of evidence, etc.

But the discussion about any one factor, such as education, quickly becomes a standalone conversation separated from all the other factors, (health, class, etc.) which are also known to be contributive to outcomes. When that happens, almost instantly, the conversation no longer remains particularly salient to the primary focus on poverty. You might come up with some robust, defensible conclusions to encourage greater education attainment but without tying that back to the other factors, you haven’t really accomplished anything. Whatever you do in terms of increasing education attainment is going to have some beneficial, neutral or detrimental impact on all those other root causes. But what? Hence the frequency of unintended consequences arising from any major policy decision.

All this was brought to mind by the opening paragraph of a post by DarwinCatholic, Is Capitalism Destroying the Family? A worthwhile question.
Apparently one of the ideas going around on the left is that if conservatives really cared about marriage, children getting to live in an intact family with both parents and other related issues, they would turn around and support progressive economics: unions, higher minimum wage, etc. Elizabeth Stoker Bruenig had a piece in The New Republic earlier this week titled "Poor People Don't Need Better Social Norms. They Need Better Social Policies." Today Jeff Spross takes to the virtual pages of The Week with a similar piece entitled "The conservative obsession with moral values doesn't explain the plight of the working poor".
I think DC is accurate about how the argument would go.
Conservative: In order to reduce poverty we need to shore up families so that we do not have so many children growing up in straightened circumstances and without parental and moral structure. More jobs and economic growth will do wonders for family formation and sustainability which will in turn do wonders for both the economy and the community.

Liberal: Correct, and in order to bolster families we need to increase unionization and the minimum wage.
The Liberal counter is perfectly reasonable, absent context. We can argue ad infinitum about the relative quality of the evidence and the differences between short term consequences and long term consequences of unionization and minimum wage laws but there is at least a plausible case to be made, even if it is wrong, that family formation and family sustainability will increase with greater unionization and higher minimum wages.

However that argument is resolved, it ignores all the rest of the context of a complex, dynamic, self-correcting system. There is a correlation between unionization and lower poverty and lower inequality but there is also a correlation between unionization and slower growth. Do we want that trade-off? Higher unionization is also associated with less technological change, and lower economic competitiveness, and inflation, and lower education attainment, and, and, and. Unionization, minimum wage, marriage rates, education attainment, etc. are all interwoven with one another and while it makes it simpler to argue about only two relationships at a time (say, unionization and poverty), that discussion is in many ways a red herring. All the pertinent and material relationships have to be discussed simultaneously to begin to understand the net impact of any given change. And that is hard.

Imagine being in that discussion and having to have a working knowledge of economics, education policy, health programs, anthropology, sociology, history, geography, religion, psychology, technology, demographics, etc. It is no wonder that people choose to focus on a small subsection that they want to discuss in isolation. And of course, advocates are especially prone to tunnel vision.

DarwinCatholic explores a lot of the specific issue about family and unionization in detail with a good deployment of data but my take-away is somewhat different.

Discussions about complex, dynamic, self-correcting systems are always difficult. They are made more difficult because we fail to acknowledge the complexity of what we are talking about, usually focusing on only one, two or three attributes and only occasionally discussion how they might affect one another. But what makes such discussion almost impossible is the advocacy element. Advocates are almost always aligned with a single, simple solution rather than being open to resolving the argument in whatever fashion the data supports. That means that whatever their particular issue might be serves as an anchor to the argument and effectively as a filter to ensure that all the other attributes are ignored. We force arguments to be simpler than they usefully can be.

Does that mean that we can’t debate complex, dynamic, self-correcting systems? Absolutely not. It means that we have to be conscientious and not foreclose discussions about all the other attributes involved, taxing and difficult as that might be.

No comments:

Post a Comment