Why are richer people more successful?

It is a fact that kids from more advantageous backgrounds have better life outcomes: kids whose parents are richer and/or educated are themselves more likely to be rich and educated. Why is this the case?

Of course, there is a lack of social mobility involved: rich kids have better opportunities. But today, in most developed countries, poor kids have the very same opportunities in theory. This is because education all the way through high school (and in many countries even beyond that) is free or heavily subsidized. What is it then that makes kids from a more advantageous background more successful? On what dimensions do these kids differ from disadvantaged ones? And does socioeconomic status matter once we account for these differences?

Schildberg-Hoerisch et al. (2013) measure some important personality/behavioral traits for 732 children in a variety of experiments to shed light on these questions.

The idea is to see whether children from families with high socioeconomic status (SES) systematically differ from children from low SES families on any of these traits. And if yes, then through what channels do these differences arise (e.g. parenting, family structure). Finally, the authors ask whether high SES children are still at an advantage even after these potential personality differences have been taken into account. This way, we can examine whether the advantage of high SES children stems completely from personality/behavioral differences. And if not, then what percentage is unexplained by these differences.

The authors’ sample includes 732 children from Bonn and Cologne, ages 7-9. They seemingly designed and conducted the data collection themselves. The children participated in a set of experiments. They and their parents also filled out a questionnaire. This is some nice original data. The following personality/behavioral traits were of interest.

Time preferences. This refers to the fact whether a child is long-term or short-term oriented. It is measured by an experiment in which children receive 7 pieces of 20-cent coins and can decide how many to take now, and how many to put in a piggy bank. The amount put in the piggy bank will be doubled, but will only be received a week later. The mean was 4.49 coins in the bank with a standard deviation of 2.12.

Risk preferences. In other words, whether the child is risk-averse or risk-seeking. Basically, here the experiment allowed the child to choose between a guaranteed payoff of 3 stars (which were an imaginary currency used in the experiments) and a lottery where one could receive 0 or 7 stars with equal probability. Then the children were also asked to choose between a guaranteed 4 stars and the very same 0-vs-7 lottery. This allowed the authors to divide children into four groups: risk-averse (never likes a lottery), risk-neutral (takes lottery over 3 stars, and 4 stars over lottery), risk-seeking (always takes lottery), and inconsistent (the rest). Inconsistent children (11%) were removed from the subsequent analyses.

Altruism. Children had the option to divide 2 stars between themselves and an unknown child in a dictator game. They could either share (1 star for each) or not (keep both stars). 84.4% kept both.

In addition, the children’s fluid and crystallized IQs were measured. Fluid IQ refers to general logical reasoning in new situations and processing speed. This is largely hereditary. Crystallized IQ is more about acquired traits such as vocabulary. The two are then combined to arrive at an overall IQ score. An interesting side note: both the crystallized and the overall IQs are approximately normally distributed, but fluid IQ seems to follow more of a power law with a long right tail.

Socioeconomic status is measured by parental education (average of mom and dad) and household income (accounting for the number of people in the household).

First, the authors looked at whether personality traits differ by SES. They find that high SES, older and male children are more patient. Males are more likely to be risk-seeking. And high SES and older kids are less likely to be risk-seeking. Finally, older children and females are more altruistic.

In sum, time and risk preferences are affected by SES. Higher SES kids are more patient and risk-averse. Some interesting gender differences show up too. Males are more likely to be patient and risk-seeking. The former is likely a fluke, because other studies find the opposite. The latter, however, is supported by many other studies. If males are more risk-seeking that can potentially explain why we observe higher variance in male life/school outcomes. Males are disproportionately more likely to be on the top of their fields (science, politics), but also to be very unsuccessful (homeless, unemployed).

IQ is also higher for children of high SES families. This is true for both crystallized and fluid IQ. The effects are stronger for the former suggesting that learned (as opposed to inherited) factors are somewhat more important. But biology cannot be completely ignored as we’ll see later. Males have a higher crystallized IQ, but fluid IQs don’t differ by gender. This suggests that gender differences in IQ are learned rather than inherited.

So we can clearly see that high SES kids are more patient and risk-averse, and these two traits are quite important for life outcomes. They also have higher IQs. So it indeed seems to be the case that children of richer kids are just better equipped to succeed in life. But through what channels do they acquire these advantageous personality traits?

To answer this, the authors look at a bunch of environmental factors that can be roughly categorized into five groups: differences at birth; quality of time spent with parents; parenting styles; mother’s IQ, education and preferences; and family structure.

Before we look at whether these can explain the personality differences between high and low SES kids, let us ask how these five factors differ between high and low SES families. The authors find the following.

  1. High SES kids have better initial conditions: higher birth weight, and being born in a later week of gestation.
  2. High SES parents spend more quality time with their children: less time is spent in child care, museum visits instead of TV, etc.
  3. Parenting styles markedly differ. Low SES parents are more likely to be inconsistent (e.g. promise punishment but then not go through with it), use psychological control (e.g. not talking to one’s child because they did something wrong), be strict, and less likely to be warm (e.g. use praise).
  4. Maternal personality differs by SES. High SES mothers are more patient, more risk-averse, and have higher IQs. There is no difference in altruism though.
  5. Finally, high SES kids have fewer siblings (which translates into more attention) and older mothers.

Now let’s move on to our last exercise: what if we account for all these environmental differences. Will SES still be important? The answer is: it depends.

Having parents with more education still leads to high patience, and higher overall and crystallized IQ; even after controlling for all the environmental differences between high and low SES households. And similarly, being born in a richer household is still positively associated with overall and crystallized IQ scores.

The conclusion here is that time preferences and crystallized (learned) IQ are affected by SES. But risk preferences and fluid (inherited) IQ are not. This is true once we control for environmental differences between high and low SES households. So to the extent that eliminating environmental differences is not costly (in terms of money), low SES households could merely change their behavior, and they could thus close the gap in risk preferences and in fluid IQ between their children and high SES children.

Namely, having fewer kids altogether and having more quality interaction with one’s children could reduce low SES kids’ preference for risk-seeking. And healthier behavior while pregnant (so as to increase the baby’s birth weight) as well as eliminating inconsistent parenting practices could increase low SES kids’ fluid IQs.

Thus, there are biological differences between high and low SES households: low SES households’ children have lower birth weight. This leads to lower fluid IQ. But this biological difference doesn’t seem to be genetic or something that cannot be changed. This is good news for low SES parents.

Now, for patience and crystallized IQ the effects of SES are not completely eliminated once we control for environment. But this doesn’t mean that low SES parents cannot change their behavior in order to create a better environment. It just means that change in behavior will probably not guarantee a complete elimination of the gap in time preferences and crystallized IQ.

What can low SES parents do to help their kids be more patient and smarter? The two seem somewhat interlinked: kids with higher IQ are more patient. So low SES parents can increase fluid IQ (as explained above), and crystallized IQ (as explained in the next paragraph) in order to help their kids be more patient. In addition to this, having fewer older siblings increases a child’s patience as well.

As for crystallized (learned) IQ, low SES parents can do a lot to improve this. They could spend more time with their kids in everyday interactions (e.g. talking, joint meals), spend less time with them watching TV, and avoid using psychological control in parenting. Having more older siblings is also detrimental to crystallized IQ. Low SES parents could also make an effort to increase their own IQ, because parental IQ is correlated with a child’s crystallized IQ.

Altogether, after taking environmental differences into account SES still remains an important predictor of certain traits (patience, crystallized IQ), but not of others (risk preferences, fluid IQ). I drew two main conclusions from this paper:

  1. Low SES parents can theoretically do a lot to improve their children’s chances in life. But at the same time, some barriers may exist, e.g. low SES parents may simply not have enough time, money, or information to make the desired behavioral changes. So there may also be room for some government policy to help low SES parents be better parents.
  2. Environmental differences can explain a lot of the low-high SES gap in children’s personalities. But they can’t explain it all for certain traits. Money and a good parental education, which come with being high SES, are still important in and of themselves for certain personality traits. So statistically speaking, the entire SES gap will probably never be eliminated, but dedicated low SES parents can still make a huge difference in their children’s life outcomes.

One thought on “Why are richer people more successful?

  1. Pingback: ZeeConomics | Are parents more risk-averse?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s