Michael Grätz: Siblings and their incomes – the same or different over the life course?

In Episode 10 of the Dial Podcast, Michael Grätz from the University of Stockholm talks about sibling similarity in income and what that tells us about their life chances later on.  The research, which uses Administrative Data in Sweden and is published as a Working Paper, was also presented at the DIAL mid term conference in June 2019.


Michael Grätz: Siblings and their incomes – the same or different over the life course?

In Episode 10 of the Dial Podcast, Michael Grätz from the University of Stockholm talks about sibling similarity in income and what that tells us about their life chances later on.  The research, which uses Administrative Data in Sweden and is published as a Working Paper, was also presented at the DIAL mid term conference in June 2019. 

Christine Garrington  0:00 

Welcome to DIAL, a podcast where we tune into evidence on inequality over the lifecourse. In today’s episode, we’re talking about sibling similarity and income over the lifecourse and across generations. Our guest is Michael Grätz from the University of Stockholm he started by explaining by looking at siblings can help us better understand how much their background and circumstances influence how they get on later in life. 

Michael Grätz 0:25

So what we are interested in estimating how much the family and community background of an individual influences his and her life chances. And the siblings approach is one way to try to do that. And so siblings share many characteristics; they have the same parents, they live in the same neighbourhood, they often go to the same schools and all these factors are taking into account if we estimate how similar brothers are and how similar sisters are.

Christine Garrington  0:55

So in this piece of research what specifically did you want to look at and why?

Michael Grätz 0:59

So the siblings similarities approach many others have used before us. The new thing is what we are doing is we are taking a lifecourse perspective. So what we are mainly interested is whether siblings similarities varies over the lifecourse. So on the one hand there is this idea that genetic effects become stronger over the lifecourse, there is also what Christopher Chang has called luck, so there are random life events happening and they might happen more the older people become. So from that perspective you would expect that siblings become less similar to each other over the lifecourse. But there is also this idea in Sociology that there is a process of cumulative advantage. And from that perspective you would think that siblings are becoming more similar the longer they are in the labour market.

Christine Garrington  1:47

And where did you get your information from?

Michael Grätz 1:49

So what we are using here are papers and administrative data from Sweden. So this is tax data, actually tax records and statistics, Sweden has collected them and made them available for researchers dating back to 1968. And for that reason we can for the oldest cohort in our study – for those born in 1950s we can actually cover their income between age 18 and 60 meaning we have income information on them, on their earnings and on their income, age of 18 to age 60 every single year.

Christine Garrington  2:22

And what else about this data is good? Obviously it has got the sort of information you want but presumably its huge as well right? It covers everybody? Is that right?

Michael Grätz 2:29

Well it covers the Swedish population so it is covering everyone who reported their earnings and their income who reported their earnings and income to the tax authorities in that specific year. 

Christine Garrington  2:39

So in that sense it is a really good source of data for you? So what was the overall picture when you compared the incomes of siblings over their lives first and foremost?

Michael Grätz 2:48

So what stands out is that sibling correlations are definitely highest when they are youngest. So age 20, and 25 when you test the effect that siblings often take the same educational pathways. So meaning if they both go to university. They have low income at the beginning. If they take up some vocational education or they start work and they are lucky they have a higher income at the beginning but then of course they don’t have the income development that people have to go to university. If you look at the further lifecourse then it really depends on how we measure the information. So if you look at income at a specific age we find a decrease over the lifecourse. Meaning that siblings become more different. They become more different to each other. But if we are accumulating income meaning we are adding up the income made from one year to another, we actually find that siblings are becoming more similar in their correlated income to each other. This is not very strong but this is what we slightly find so there is a kind of U shape pattern meaning that their overall income at the end of life shows actually siblings are more similar than what was previously also thought in the literature.

Christine Garrington  3:57

And did it matter whether those siblings were brothers or sisters?

Michael Grätz 4:00

So it does matter for the level in the sense that the effect of family background on income was lower for women than for men. This is what other studies have also found. That is what we find here. Meaning we find lower statistical correlations and higher product correlations but the development over the lifecourse is the same for women and for men. 

Christine Garrington  4:22

And what about if you came from a poorer background or a richer background? Did that matter in some way?

Michael Grätz 4:26

So this is an interesting question because that there is good reason to suspect it should matter. So we are referring to the work by Bolton Connolly from Princeton University who had this idea that families who have a lot or resources are investing that in a way to make their siblings more similar and that we should also find that families from advantaged backgrounds should have more similar siblings. So there should be these differences and he actually also found these differences in one study in the United States. So this is what motivated us trying to replicate his results but we did not find evidence for this. So this is not the case. Even though there are good reasons to expect, it doesn’t affect whether siblings are, the level of sibling similarity depending on from which background they are from. 

Christine Garrington  5:18

Can I ask what you make of that? Out of interest? Why the difference? Any thoughts?

Michael Grätz 5:24

Yeah so maybe parental investments are not that important after all. So either, or parents don’t, if they are compensating in the way we presume they are not compensating enough. It at least not make the income of their siblings more similar. Or they maybe compensating with wealth transfers, so direct transfers. Which is not something which we are not observing in this study.  

Christine Garrington  5:51

Now you looked across three different cohorts born in the 50s, 60s and 70s how did the picture differ between those cohorts?

Michael Grätz 5:58

Okay to be honest that is not the very long historical perspective just having three cohorts so from them we find stability which is also what other studies have found before us. But from when we do find a change, and the change goes in the way that sisters become more similar because the three cohorts so the family background impact increases. And what this means at least that the kind of increasing female integration into the labour market and increasing gender equality is associated with a stronger effect of family background, meaning that it has mainly advantaged women who profit at least first from this increasing gender equality. 

Christine Garrington  6:41

And what have you learned about sibling incomes that we didn’t know before then? And what are the important takeaways would you say from this?

Michael Grätz 6:49

What we are claiming that our study shows is that it is really important to take this lifecourse perspective. And the main takeaway from this whole field of stratification research is the message that it is important to take a long, as long information lifecourse as possible. And it is important to accumulate income over, over many years. So the overall level of sibling similarity that we find for both brothers and sisters is much higher if we are looking at the whole lifecourse aged 18 to 60. And add up all this information than if we just look at five years of the lifecourse, or 10 years of the lifecourse, which is what most previous research has done. And we think that this is, this is of course speculation now but this insight is due not only for our study but the whole inequality that it is very important to take this lifecourse perspective and probably also in other countries than Sweden.

Christine Garrington  7:50

Sibling Similarity and Income: a Lifecourse Perspective is research presented at the DIAL midterm conference in June 2019 by Michael Grätz. You can find out more about the NORFACE funded DIAL projects at www.dynamicsofinequality.org. Thanks for listening to this episode of our podcast, which is presented and produced by Chris Garrington and edited by Elina Kilpi-Jakonen.