'The success sequence is just correlation,' answered point by point

There is a sentence that ends many a debate about poverty, and it goes like this: "That's just correlation." It is delivered with the calm of a man laying down a trump card. The speaker has not disproved anything. He has pointed out that a statistical pattern is not, by itself, a proof of cause, which is true and which every freshman statistics course teaches in week one. What he usually omits is the next step, which is to say what else explains the pattern and whether the explanation survives.

The success sequence draws this response more than most claims. So let us treat it the way it deserves: define it, state the evidence, give the objection its strongest form, and answer it point by point.

What the claim actually says

The success sequence is a short list of choices. Finish at least high school. Work full time. Marry before having children. The claim is that Americans who take those three steps, in that order, are very unlikely to be poor and are likely to reach the middle class.

Notice what the claim does not say. It does not say that everyone who skips a step is doomed. It does not say that every person who follows the steps will prosper. It does not assert that the steps are the only route out of poverty, or that they are equally available to everyone. It is a statement about probabilities, offered as a guide to choices, and most of the heat in the argument comes from people who attack a more sweeping claim than the one that was made.

Two terms need definition before we go on. Poverty here means income below the federal poverty line, which the Census Bureau and the Department of Health and Human Services define and update each year. Middle class in these studies means income at or above a multiple of that line, which the authors of each study state; I describe their definitions in general terms and leave the exact cut-offs to the reports themselves.

The evidence

Two studies carry most of the weight, and one fact about them is worth saying first. They did not come from a single political camp.

The Brookings finding. In 2009 Ron Haskins and Isabel Sawhill published Creating an Opportunity Society. Haskins and Sawhill are senior fellows at the Brookings Institution, which is generally counted on the centre-left. They examined how Americans fare when they complete high school, work full time and wait until marriage to have children. They reported that among those who did all three, only about two percent were in poverty, and roughly three quarters reached the middle class, as the authors defined it. Among those who did none of them, the numbers ran the other way: by the authors' account about three in four were poor, and only a small fraction reached the middle class. The precise percentages and the middle-class cut-off are in the book and the Brookings summaries of it, and I encourage readers to check them there.

The millennial replication. In 2017 Wendy Wang and W. Bradford Wilcox, writing for the American Enterprise Institute and the Institute for Family Studies, published "The Millennial Success Sequence." They looked at young adults in their late twenties and early thirties, the cohort that grew up after the culture's own norms had weakened. They found that the vast majority of those who followed the sequence, 97 percent by their count, were not in poverty, and that 86 percent of those who followed all three steps were in the middle or top income groups. Those who had followed none of the steps were far more likely to be poor. Again, please read the report for the exact figures; I am describing its direction and approximate size.

So the first fact is this. A scholar from the left-leaning Brookings Institution reached the same conclusion as scholars from two conservative-leaning research institutes, eight years apart. That does not prove the claim. It does mean the claim is not the product of a single team's assumptions.

The objection, at its strongest

The objection has force. Here is how its best defenders would put it.

First, the sequence describes people who were already advantaged. Young adults who finish school and hold steady jobs tend to come from stable homes, to have had better schools and neighbourhoods, and to have developed the habits that employers reward. They are also the people who find spouses who are themselves steady. When we see that these people are rarely poor, we are seeing the effect of what they began with, not what they did.

Second, causation might run the other way. A person with a stable job is a better marriage prospect. A person with money and prospects can afford to wait for a spouse before having children. If so, prosperity causes the sequence rather than the reverse.

Third, the steps are not equally available. A young woman in a town where the men have no jobs is not in a position to marry a man with a steady paycheck, whatever she might wish. To tell her to follow the sequence is to tell her to choose from a menu that has been taken away.

Fourth, and morally the most serious, the sequence can become a way of blaming the poor. If poverty is the result of three bad choices, the argument runs, society can stop worrying about the structures that produce it.

That is a serious objection, and some of it is true. Let us answer it point by point.

The answers

One: selection explains some of the pattern, and nobody serious denies it

The first answer is a concession. Of course selection plays a part. People who finish school and hold jobs differ, on average, from those who do not. If the claim were that a random person assigned to follow the steps would have a ninety-eight percent chance of avoiding poverty, the objection would defeat it.

But that is not the claim. The claim is a description of the pattern and a recommendation that people aim at it. And two features of the evidence suggest selection is not the whole story.

The pattern holds across family backgrounds. The Wang and Wilcox report looked at young adults who grew up in different circumstances, and they report that the advantage of marrying before having a child held across racial and ethnic groups and across family-income backgrounds, even after controlling for education and the family's income growing up. A selection story that relied entirely on advantaged upbringing would have trouble explaining that. Selection is part of the account. It is not the entire account.

And the three steps are not three traits. They are three decisions. Finishing school, taking a full-time job and delaying children until after marriage are acts of will, and a person can choose them or not at the margin. A person cannot choose to have been born to wealthy parents. When the objection says "these are advantages," it blurs the difference between what a person inherits and what a person does.

Two: reverse causation is possible, but the sequence is not the same as income

The second objection says that prosperity causes the sequence. In some cases it plainly does. A man with a good job is more likely to marry, and the data show that.

But consider the first step and the third. Finishing high school happens in the late teens, before a person has earned much of anything. Having a child before marrying is a choice that is made at a point when income is typically low, and it is a choice that can be avoided at every income level. The sequence is largely settled by the early twenties. If the pattern were driven by income alone, we would expect those from the same family-income background to do equally well regardless of whether they had children outside marriage. The Wang and Wilcox report found that they did not: marriage-first young adults were about twice as likely to reach the middle or upper income groups, even after controlling for family income growing up.

There is also the mechanism, which is easy to state and hard to dispute. A child born to two married parents has, on average, two incomes and two sets of hands. A child born to one parent has one. A person with a full-time job has an income; a person without one does not. A person who finishes school can compete for more jobs than one who does not. These are not mysteries. They are arithmetic. One person earning $30,000 supports a household of three on a smaller share per head than two persons earning $30,000 each. The sequence works in part because it adds earners and reduces the number of dependants per earner at the moment the household is most stretched.

Three: unequal opportunity is real, and it is an argument for the steps, not against them

The third objection, that the steps are not equally available, is the one I take most seriously. Some men do lack steady work. Some neighbourhoods do have weak schools. Some young women do face a shortage of marriageable partners.

But follow the logic. If the problem is that the steps are hard to reach, the response is to make them easier to reach: better schools through choice and competition, work that pays, a tax and benefit system that does not penalise marriage. I have written about the last of these in the marriage penalty buried in welfare programs, with the math shown, where a single mother's credit drops by thousands of dollars when she marries the father of her children. That is a rule a legislature wrote and could rewrite. And it is precisely the sort of obstacle that the success sequence, taken seriously, tells us to remove.

The objection, in other words, concedes that the sequence is worth reaching. It simply says some people cannot reach it. Fine. Then the conservative agenda is the agenda for opening the path: school choice, work requirements and work supports, pro-family tax rules, safe streets. A person who says "the sequence is a good outcome that is hard to get to" has agreed with the thesis and quarrelled only about the policy.

Four: teaching the sequence blames no one

The fourth objection is moral, and it deserves the plainest answer.

Does teaching a sequence of choices amount to blaming those who did not follow it? No, for the same reason that telling a teenager to wear a seat belt does not blame the victims of crashes. Advice is addressed to the future. It is given to people who have a choice still to make, which in the case of the success sequence is mostly young people, the ones who have not yet had a child or left school.

Consider what the alternative amounts to. Suppose educators, pastors and parents stopped describing the sequence because it might make someone feel judged. Teenagers would then hear nothing from the adults who care about them, and everything from a culture that treats these choices as private lifestyle options with no consequences. The silence is not neutral. It is a decision to withhold information that the best evidence says matters.

And there is no reason the advice must be delivered with scorn. A single mother who finished school, works two jobs and raised her children with courage has done something admirable, and no honest description of the sequence contradicts that. The point is not that those who followed a different path are failures. The point is that the next generation deserves to know which paths have the best odds, and that it is a service to young people to tell them.

What the objection gets right

An honest answer concedes what the other side has earned.

Correlation is not causation, and the sequence studies establish a strong association, not a controlled experiment. No one is going to randomly assign teenagers to marry or not to marry. The best evidence we can get is observational. Where I part from the critics is in what follows from that. In most of social life, and in most of the decisions a family makes, we act on strong observational evidence combined with a plausible mechanism, because the alternative is to act on nothing. We tell people to exercise and to save for retirement on exactly this kind of evidence.

The critics are also right that the sequence is an incomplete picture of poverty. It says little about disability, illness, a layoff, a divorce no one chose, or the death of a spouse. A sequence that has no place for those events is not a theory of poverty. It is a list of ways to reduce a family's exposure to it. That is modest, and it is useful.

The longer view

The success sequence is not a new idea. The conviction that a household is made stable by work, education and marriage is as old as the tradition that grounds the family in covenant, a point made in the family came before the state, so why does the state act like a parent?. What is new is the data, and the data have tended to support what grandmothers said.

The sequence also connects to the question of fathers. A man who finishes school and works full time is the sort of man a woman can reasonably marry; a man who is absent from the child's life is a cost the child bears for decades. That is the theme of fathers are not optional, and the culture that says they are is wrong. And the sequence can be reinforced by policy, as the options compared in baby bonuses, tax credits or cheaper houses, comparing pro-family policies show.

There is a practical side too. A family that follows the sequence still has to manage a budget. The household finances of a young couple, the temptation of debt and the use of a family plan for childhood are all part of the picture, and A Family Playbook for a Childhood With Fewer Screens and More Freedom is one place to see how parents try to keep a household's priorities in order.

A fair verdict

Here is the verdict, in four sentences. The success sequence is a strong correlation, supported by research from the centre-left and the right, with a plausible mechanism and a pattern that survives across family backgrounds. It is not a proof of causation, and nobody who understands it claims that it is. The objection from selection is a good reason to be careful about how much any one person's outcome depends on the steps, but it is not a reason to stop recommending them. And the moral objection, that the sequence blames people, has the matter backwards: the failure would be to know what helps and decline to say.

The case in brief

1. The success sequence says that finishing high school, working full time and marrying before having children makes poverty very unlikely and the middle class likely, and it is a statement about odds, not destiny.
2. The finding appears in a centre-left Brookings study from 2009 and in a 2017 report on millennials, which is evidence that it does not belong to one political camp.
3. The selection objection is strongest when it says that advantaged people are more likely to take the steps, but the pattern holds across family backgrounds and the steps are choices rather than inheritances.
4. Where the steps are hard to reach, the answer is to open the path through schools, work, safer streets and fair tax rules, not to stop describing the destination.
5. Teaching the sequence blames no one, because advice is addressed to those who still have a choice to make.

What you can do: find the Haskins and Sawhill chapter and the Wang and Wilcox report, read the tables for yourself, and share them with a teenager who has not yet made these choices.