Select text to annotate, Click play in YouTube to begin
00:00:00
[Music] [Music] it's a pleasure to welcome Carlos car
00:00:26
lingren who is a professor and at Upsala University C and Deputy head of the Department of government uh Carlos Scar's research interest include political participation and
00:00:38
representation democratic theory EU politics and education policies is the co-author of the book entitled par parp participatory governance in the EU
00:00:51
enhancing and engering democracy and efficiency you have coauthored and coor more than three dozen article chapter scient ific reports on various topics your academic work has appeared in
00:01:04
leading general and political science journals such as nature the American Political Science review the American Journal of political science and the Journal of politics in 2009 you received
00:01:18
Upsala University's distinguished teaching award for your outstanding accomplishments in uh uh undergraduate teaching and you have Alo been awarded the SJ best paper award and the Lawrence
00:01:31
longle best article award by the American Political Science Association and you will give us two presentations today you will tell us uh you will talk about perpetuity of the past how
00:01:44
political inequality is transmitted across Generations uh and uh I I won't be uh take any longer and I give you the floor and you have one hour uh and of and now
00:01:57
start taking notes thank you so much great pleasure to have you with us yeah thank you very much so nice to be here I have to apologize I don't speak any french I took a class for three weeks it
00:02:08
would be a very short speak if I try to speak in French but I'm going to speak in swinglish instead hopefully you're going to understand me so but the topic of this lecture as I said is the perpetuity of the past and more
00:02:22
precisely we are concerned about what could be said the essence of democracy so if you have this Ida what is it really the democracy is about a common answer to that question is that the
00:02:35
essence of democracy is that all citizens living in a society they should have equal and adequate opportunities to participate and influence public decisions it doesn't
00:02:49
mean that all individuals has to be equally politically active it only means that they should have the same means to be politically active should they so prefer so in that sense democracy rest
00:03:03
on the principle of equality of opportunity rather than equality of outcome however like all important Concepts the exact meaning of equality
00:03:14
of opportunity is obviously contested what is it actually required for something to fulfill the principle of equality opportunity there are many answers to that question but one
00:03:28
influential answered advocated for instance by the economist John rmer is that equality opportunity it requires that the distribution of
00:03:40
different important outcomes should not depend on factors that we as an individuals do not have direct control over but as all good answers obviously that answer raises many new questions
00:03:56
what are these exact circumstances that we as individuals lack controller I guess we can spend the next hour trying to agree on an answer to that question try to make this long list
00:04:09
of these circumstances that we cannot control but I will not be the purpose of the election hopefully unfortunately I would argue we don't need to agree on
00:04:22
that we only need to agree on one thing in order to have a topic for this lecture and that is the that very few of us have any direct control over the
00:04:33
identity of our parents that is we do not get to pick or choose who our parents are which I know at least my daughter sometimes think is a
00:04:45
Pity but the thing is that who our parents are is outside the control of us our individuals and turning back to the
00:04:58
question democracy that also requires that family background should not depend should not influence who get political influence or political power in
00:05:10
democracy and obviously family background has no such importance because in the Democracy we are all equal but I guess we can just look at this picture and we see that apparently
00:05:24
some individuals are a bit more equal than others so I think it's enough to give a names like is we have the gandhis the bush the trul and the leans in order to understand it's very
00:05:37
difficult to think that this can happen by random chance if family background is not important for the distribution of political power in democracy is very baffling I would argue
00:05:51
that we have these parents and children on these important positions in these very large countries and this is just a few stupid examples
00:06:04
but over the years there is increasing evidence that children go up to resemble their parents not only socially and politically but also politically that is
00:06:19
it tend to be the case that individuals who are politically active get children who also politically active themselves and if you go back to the idea of
00:06:32
democracy and think is that the problem I guess one problem with democracy is that we see here and now that some people are more influential than others that might be a problem but an
00:06:45
even bigger problem especially for sort of the starting point of equality or political opportunity that is obviously if these kind of disadvantages or advantages that Ur to the individuals in
00:06:58
one particular generation also passed on to The Heirs of future Generations because then these kind of disadvantages or advantages in terms of political influence are going to carry
00:07:12
through time in in a rather discomforting manner so if you're looking at this this kind of idea that who our parents are matters for our political influence it
00:07:26
have some disturbing parallels with a kind of predom Democratic societies against which Democratic revolutions were once fought so obviously in order to sort of
00:07:41
make democracy more sort of vivid and Alive it would be good if you somehow could try to mitigate or take different kind
00:07:52
of actions or invoke certain policies that could combat this kind of reproduction of politically inequality transmitting across Generations but obviously in order to
00:08:06
take these kind of measures and invoke these kind of policies we need to have a good understanding on the mechanisms underlying this kind of sort of
00:08:20
Persistence of political inequality in democracies because obviously the mechanism for this is not the same as it used to be in this old preem ratic regimes but these kind of persistences
00:08:33
across Generations now take new forms and the question that we're going to discuss today is what is it that drives this transmission of political inequality
00:08:44
across Generations so this is the question that I have sort of paid some attention in my recent research together with colleagues where we had tried to fill this Gap because I would argue that we still have
00:08:57
a rather poor understanding of of the roots of this kind of intergenerational Persistence of political equality so that is what I'm going to discuss
00:09:10
today today we're going to be a bit boring and mainly sort of look at the problem try to see how grave is the problem what is that causes this problem
00:09:23
tomorrow hopefully if some of you come back then we try to be more constructive and ask the question question but hey is there anything we can do about this is there a way to reduce the importance of
00:09:36
family background in modern de Dem democracies when it comes to the distribution of political influence and in particular we're going to look at one factor one kind of policy and that is
00:09:49
improved educational opportunities because being in a hall like that we all like to think that education should matter for this kind of questions don't we and if you come back tomorrow I Canna
00:10:01
tell you whether or not we are correct in hoping that but today I will instead focused on sort of the research I have done together my colleagues looking at sort
00:10:13
of the mechanisms underlying this kind of persistence in political participation and influence across generations and all of my research on
00:10:26
this issue even though I often talk about participation and influence you're going to see that it focuses on participation not on influence and I guess there multiple
00:10:38
reason for that but the most important is that influence is incredibly difficult to measure and observe but I would argue that there are good sort of reasons for studying
00:10:51
participation as well because if you think about democracy what is the most most natural way to have political influence that should be through participation it could be by throwing a
00:11:03
vote in election or it could be sort of standing for office or even holding political office so participation is incremental to the way that you get political influence in a democracy and
00:11:17
we know from Plenty of research that individuals who participate less they around the risk of getting their preferences and interest neglected in the political process so if you're not
00:11:29
there to table when the decisions are made there is a big risk that you're going to lose out on the resulting policies so that's a reason why we're going to focus on participation and also
00:11:41
of course because it's much easier to measure participation than influence so we're going to study different types of participation most likely I most mostly focused on is voting in
00:11:56
elections but also whether or not your nomination ated or elected to political office on different levels but before getting into this empirical
00:12:08
analysis we should start in asking but what are these different theoretical mechanisms that can cause this kind of persistence in political participation
00:12:22
across generation what can drive this observation that we have that politically active parents are more likely to get children that are politically active themselves what drive
00:12:34
this parent child correlation and political participation looking back at the research on this issue there are three main theoretical suggestions here the first one is the
00:12:47
upper curved Arrow saying is about status transmission we know from Plenty of research in sociology and economics that we as parents tend to pass all on our social status to our
00:13:02
children it could be in terms of Education it could be in terms of income wealth or occupational status or different things but we know there is a strong intergenerational correlation in
00:13:13
this kind of status positions and we also know as political scientists that the strongest determinant of political participation in all countries all time periods is the
00:13:28
social status of of individuals that's pretty much as close as we come to social universal law in political science people are better educated have higher incomes and better jobs they tend
00:13:41
to participate more politically than the less fortunate peers so it could be the case then that this correlation in voting or political cacy between parents and children could be
00:13:55
driven by this kind of status transmission that is that we as a par s pass on our social status to our children but it sounds a bit boring if that would be the entire story so
00:14:09
obviously there are other possibilities one possibility is that it could also be about social learning that actually the social learning focuses on the fact that
00:14:21
parents teach their children how to behave through instruction and observation that is politically active political parents they can teach the importance of
00:14:34
political participation by telling the children how important it is that you engage in society or you can just teach them that by taking them with you to the political
00:14:46
meetings or when you go to vote on the election day so the idea here is that there also this possibility of social learning so that parents teach their children how to behave politically Al by
00:15:00
being teachers and Role Models so that's second both of these focus on the social environment that is the idea is that we as a parent we pass on a social environment to our children and that
00:15:14
social environment can be more or less conducive to political participation and that was long the dominant story in this field but then come along this so new field of
00:15:27
behavioral genetics and saying hey it's not only the case that we as a parents pass on a social environment to our children we also pass on our genes and the question is why should
00:15:41
that matter for political participation it's not the case obviously they're going to be some specific genes sort of programming you for voting
00:15:54
or standing for political office but there's still plenty of evidence that many complex social traits like these like voting like becoming
00:16:07
politicians they can have genetic effects they have a number of small genetic effects that all work together to create these kind of outcomes but they do it through all these
00:16:19
intermediary factors and in this case many believe that one particularly important intermediate factor is different kinds of personality traits we know that how we are as individuals
00:16:33
if we're outgoing or introvert whether or not we think it's okay to break Norms or not that is partly not inheritable
00:16:44
from our parents it's gen related to genetics and if you also believe that whether not to participate politically in a democracy can be affected by these
00:16:58
kind of person ality traits so the likelihood that we're going to be Vote or run for political office could depend how we are as persons then that obviously open up the
00:17:11
possibility that our genes could also be important for our political behavior and that also open UPS the possibility that this kind of resemblance that we see in the political
00:17:24
behavior of children and parents at least partly can be due to this kind of genetic transmission and what I would argue is that since these kind of mechanisms and
00:17:38
explanation are pretty different it could be important to understand how important they are in order to know how to elevate or mitigate this kind of
00:17:50
intergenerational political inequality because obviously it could call for different policies depending on whether this correlation is due to status transmission social learning or genetic
00:18:03
transmission perhaps different kind of measure could then be unequally effective in mitigating this depending on which of these stories is the most empirically
00:18:16
accurate so that's what Mar of research tried to do but initially as I said If you're looking back and this is 50 years of research in one slide but the dominant perspective
00:18:33
during the 70s and 80s was it's all about social learning this entire literature was focused on sort of the political psychology and saying that what is matter here
00:18:46
is how effective parents are in transmitting their norms and behavior to the children so there were a lot of focus studying trying to see what kind of factors are hindering or fostering
00:18:59
this kind of transmission from parents to children and the possibility of status transmission was rarely raised and I would argue that the issue of
00:19:11
genetic transmission was even less rarely raised but then come along this new research someone argue for status transmission and saying that that is very important but even more so people
00:19:25
started to think about this genetic transmission and we have this first first generation of genetic studies in political particip in political science and mainly use
00:19:37
twins because twins are nice in the way that they are identical so identical twins are genetically identical and then you have these non-identical twins that are more like ordinary siblings and then
00:19:49
one could come up with these clever ways trying to decompose and disentangle the differences between these genetic factors and these environmental factors
00:20:02
and what these early twin studies found in the area of political science was that these kind of shared environments didn't matter at all for important political trait such as
00:20:14
political participation and political behavior and that sounds a bit surprising giving that these entire literature were saying that it's all about social learning a bit
00:20:27
about state of trans that is exactly that kind of shared environment that the twin studies and these behavioral genetics said that's do not matter at all so we have this
00:20:41
completely different literature why one saying that it's all about social learning the other one saying it's not at all about social learning and I guess the starting point when me and my colleagues started to work on on this was saying that both of these
00:20:54
conclusions in these two phases seems a bit non-intuitive it seems a bit strange that we have these really complex political outcomes like political participation and it should be entirely
00:21:07
driven by the social environment or entirely driven by biology complex things especially when you get older get a bit more boring then you often realize that they
00:21:21
always the glass is all always half full or half empty always depend on many different things so that's what we starting off with this project and we had this idea that perhaps these very
00:21:35
strong conclusions reach in this early research was at least partly due to methological artifacts these twin studies for instance the rest an extremely stringent assumptions that is once these
00:21:47
assumptions do not hold or are relaxed then this nice decomposition results cannot be trusted anymore so they often a rather difficult way to do this so what we argue was that the better data
00:22:01
methods are needed to make progress on these issues that's usually what you argue when you think you have these kind of data and methods to use and that's what
00:22:13
we have and that's kind of the reason that we're going to focus a lot on Sweden today not only because I'm from Sweden but also because I guess one of the comparative advantages of Sweden
00:22:25
when it comes to Empirical research is that we have very good and sort of huge register data administrative data they can use for different purposes because we have this big welfare state and
00:22:38
people have these personal numbers personal identifiers and all the informations all the interactions with the welfare state are recorded and linked to these personal
00:22:51
numbers and a good thing is then that this can be used for research purposes because you have access to all this data and we have these unique personal identifiers which means that we can link
00:23:04
these different reges but obviously we need to do that in a safe way so we don't get direct access to the personal numbers they're all an anous by the statistic Sweden but it's still possible
00:23:17
to link a lot of interesting information across registers and for this purpose today one of the most important thing is that we can link almost all individuals born
00:23:28
1932 and later to the parents so because of that we have this entire population of sweds where we can link the parents and the children we can add various demographic and socioeconomic
00:23:42
characteristics from other registers and also historical censuses so we can go back to censuses from the 60s and 70s and link to this data and then we also have increasing availability of different measures of political
00:23:55
participation me and my colleague they have scanned and digitized the entire electoral roles for man elections and we also have complete data on all individuals appearing on party lists or
00:24:08
being nominated or elected to political office in Sweden between 1982 and 2022 So based on this data we thought we have
00:24:20
this idea that we should be able to improve the previous research in this area but then then obviously need to have some kind of methods in order to try to
00:24:31
disentangle sort of the social environment importance of the social environment and the genetic factors I already said the twin studies was a popular strategy but they have some
00:24:44
obvious problems but another strategy that were pioneered by some Economist especially in in Netherlands but also in Sweden they argue that we can to look at
00:24:57
adoptions because options in some way con construct a a natural experiment in this case what we often have when we're looking at the parental transmission is
00:25:09
that as a biological parent who also raise our children we're going to pass on both our genes and the social environment but the question is can we come up with an
00:25:21
experiment that this entangle these two things and that is what an adoption study can do in asserted assumptions so in these papers we have these very long
00:25:33
sections discussing these assumptions and trying to argue why we think they hold in a Swedish case but I will not bore you with all the details here but really sort of the is not that difficult we can what we can do in the registers
00:25:47
since in Sweden until the early 70s we have mainly domestic adoptions so International adoptions coming later before that was mainly domestic adoptions and that means that both the
00:26:00
biological and adopted parents were residing in Sweden so they can be found in our registers so what we can do then we can create two different samples first what we call an own birth sample
00:26:13
and that is what we usually do in these kind of studies here we have sort of the the outcome of an own child ey that going to depend on the
00:26:24
same outcome among his or her biological par and this raw parameter that's going to be the standard transmission coefficient that's going to be sort of the
00:26:36
correlation between the parental outcome and the child outcome so if this is voting this going to tell us how important the voting behavior of parents are for the children's voting behavior
00:26:51
the problem here is that the rule is going to sort of cover both these pre-birth and postbirth factors in order to try to disentangle them we need to go to the adoption sample and
00:27:02
utilize the fact that here we have two different kind of parents that contribute different things to the child first we have the birth parents the one
00:27:15
who created the child but then adopted it away for some reason they're going to contribute with the genes but if these adoptions is occurring very close to the birth of the child which was usually a case in
00:27:28
Sweden then they will have no social interaction with the child so they should not be any possibility for instance for social learning between the parent and the child on the other hand we have
00:27:41
these adoptive parents they are genetically unrelated to the child and as long as these assignment from adoptive parents to own birth
00:27:54
parents is sort of reasonable random then we can make the assumption that these beta two parameter are going to capture the impact of postbirth factors
00:28:09
or this kind of social learning so here we have this idea that since we have these two types of parents one contributing with the genes and the other one with the environment to make it simple we can try to disentangle and
00:28:22
see how important are these two type of factors for the political behavior of the side and under the assumption that this model is correctly specified so these
00:28:34
kind of terms are enter additively so with simple sum and we can separate them now we're also going to have this relationship that the overall
00:28:47
transmission row is going to equal beta 1 plus beta 2 so that's what we can try to see whether or not that makes sense in our data so that's what we did so we had this election
00:29:01
data for three different election I think when we did the studies 19782 1994 and 2010 there was four elections and we also have data on sort of whether a child was nominated to political office
00:29:14
at the municipality Regional or national level in Sweden we also have similar data on whether the parents both the biological parents and their adopted parents were ever nominated to political
00:29:27
office and then we try to see first of all do we find evidence of this kind of intergenerational persistence in political participation in in a rather equal
00:29:41
country like Sweden Sweden is also still see how long a democracy so that's also going to help us to see whether or not we can find it but next thing we're going to see if
00:29:53
there is such a persistence to what extent can it explained by these social environmental factors or this pre-birth sort of the postbirth factors or to what extent can
00:30:06
be explained by genetic factors and these pre-birth factors so that's what we did we have two different studies and to make a long story rather short this is what we
00:30:18
found the left graph is the main results for the vote voting study here a look at voting the right one is for the candidacy the Y AIS here is simply
00:30:32
matching transmission coefficient that is simply the regression coefficient saying that how much are children's turnout
00:30:43
increasing when parents turnout is increasing by one unit how much more likely are individuals to to run for political office when their parents run for
00:30:56
political office and what we see here is that in the own sample that transmission coefficient is about 23 means if you have two voting
00:31:07
parents you're going to turn out 23 percentage more in the election compared to an individual who both of the parents do not vote so that is measured that there is a correlation in voting
00:31:21
behavior between parents and children but another thing we try to do in this adoptive sample is try to distinguish between the importance of the pre-birth factors that is the green
00:31:36
part of the bar and the post birth factors that is the social environment measured by this reddish part of the bar and what we can see here is for voting both of them matter they are both
00:31:50
statistically significant I don't show that here sort of the pre-birth part counts for about 40% of this entire Transmission in the
00:32:02
adoptive sample whereas the postbirth account for about 60% so we say that if you're going to trust this estimate about 60% of the correlation is due to the fact of the social
00:32:15
environment that is the adoptive parents whereas 40% is due to the genetics that is the biological parents looking at that at candidacy we see that that's
00:32:29
even more even split that's about the 50/50 split so 50% is Du to genetics 50% due to the social environment you can also see another
00:32:43
thing here that is I said that if this model is completely correctly specified These Bars in the adoptive sample should add up to the transmission
00:32:55
in the omber sample we see that is not the case in either of the it come pretty clear close but not entirely there are many reason of course why this could be the case but one possibility is
00:33:07
that these model should not be additive in the sense that we have this genetic and social factors having independent effects on the outcom in question there
00:33:19
is this growing literature on nature nurture interactions obviously we have even more intricate things going on here so in the voting study where we have somebody data to
00:33:33
work with we try to see whether or not we could see that I not going into all the details here and this might be a bit more difficult to see for you but really what they saying is that the relative
00:33:48
importance of these genetic and social factors it varies a lot over different social conditions so for instance try to look how does it vary with age and then
00:34:01
we look at the age 18 then this is the relative important of the social environment the small white one is the relative important of genetic factors so we see when invia
00:34:15
very young in this case still living with their adoptive parents their genetic endowments seem to be rather unimportant for the political Behavior where whereas if you fast
00:34:29
forward the case to these individuals now being 50 years of age then we see that these social environment conditions are much smaller
00:34:40
whereas these biological factors are gaining importance the next one is instead comparing individuals who live with the parents and live without them we see the same when you live with your
00:34:53
adopted parents the social environment is much more important than the genetics but once you move out that's starting to change and here is a thing that they
00:35:06
focused a lot on on this in social learning theory in the 70s they said that if you have really consistent cues then the social learning should be more successful so if you get really clear
00:35:19
message from your parents what you should do then it's more likely they going to get the message and act on it and that what you also see that is if they're
00:35:30
consistent Behavior or the pattern so they always vote and never vote both the mother and father vote then the social learning seems to be much more effective and much more important than it is if
00:35:43
you get these kind of inconsistent cues so this sort of another thing adds on that is not that easy so we have just is 404 46 split or 56 split it could also
00:35:56
vary across the live CLE cycle it could vary across the context that you are in so that's why I think even more research is needed obviously because these are really complex things that we need to
00:36:07
sort of dig into more but at least I would like to think that it was somewhat of an important contribution to show that this kind of intergenerational political in
00:36:23
equality has both biological and social roots and also suggest that this kind of literature that we historically had on this topic they need to sort of work
00:36:36
more together so we have this traditional political socialization literature we have this newer behavioral genetics literature and I think if you only follow one of these routs you're
00:36:48
going to miss out of a lot of important things going on there in order to really understand how political inequality in participation is reproduced across Generations we need sort of take the
00:37:02
sort of the the insights from both of these literature into account and I think our studies were a small first step to actually do that even though I didn't say at the
00:37:14
time when I published these articles I can say it now that they're still very limited in many respects one of the things that are limited in as you saw is the way they
00:37:27
use this simple decomposition into nature and nurture obviously in order to really understand this we need to get more along try to understand this interaction more than they actually did
00:37:39
we sort of scrap the surface on that I would argue so we need to understand that better but another way on which I'm going to spend the rest of this lecture on another limitation that both of these studies
00:37:52
had that they also shared with a lot of previous literature or actually say all the previous literature in the field was that he was stuck he was we can call it
00:38:05
the two generation Paradigm that is we only talked about parent and children okay we spoke about biological adopted parents and children but we
00:38:18
still had this idea that political transmission is something that only occur from parents to children we completely ignored and disregarded the role played by more
00:38:31
distant ancestors couldn't it for instance be the case that grandparents or great grandparents could also matter for how the children turn out to
00:38:44
what extent they participate in politics or not and if that is the case it open UPS even more and perhaps even more disturbing opportunities for this kind
00:38:55
of persistence in political inequality across Generations because then it will not only matter who our parents are it could also matter how our grandparents are or what their parents are and their
00:39:09
parents are and the question is where should we stop so before asking a question where we should stop I guess we should ask the
00:39:21
question do they really have any good reasons for trying to pursue in this kind of research actually are this some reasons for trying to get out of these
00:39:34
two generational Paradigm and consider multiple Generations when doing these kind of exercises and I guess they are both I would argue theoretical and methodical reasons to do that
00:39:48
theoretically there is a quote from sort of political scientist probably the most well-known political science nowadays Robert patnam he wrote a book on inequality in J in the 2016 and there he
00:40:01
raised the possibility that Grandparents were part of the story and his argument for that was that grandparents today he said they're often more important in the grandchildren's life than the
00:40:14
counterparts were half a century ago because grandparents are healthier and wealthier than it used to be so his argument was that since the grandparents are healthier they live longer they can spend more time with the grandchildren
00:40:27
and they're also more wealthier so they open up these new possibilities that they can also impact the political behavior of the children but there are also these
00:40:39
methodical reasons to try to consider additional generation and this might be a bit more technical but if you have this idea that political inequality is
00:40:52
only transmission transmitted from one generation to another you're going to have what in statistical Palance is called the first order Mark of chain so that is the only thing going to matter
00:41:05
is the state in the previous the previous state is going to matter for the current state so that's going to mean that even if we have a rather strong correlation between parents and the
00:41:17
children we can still have rather High political mobility in the long run because you have this kind of geometric Decay where this changes quickly so we have a stupid example we can assume that
00:41:30
correlation between the parent and child is.3 then according to this two generation Paradigm expected correlation in voting
00:41:42
between the child and his or her grandparents going to be3 raised to the^ of two so they are going to be 0.9 that correlation and then we can do that one
00:41:56
more step and look at the correlation between great grandparents and grandchild great grandchildren that correlation going to be 3 raised to the power of three that's
00:42:09
only going to be 0 0.027 so that's just saying the fact that even if you have strong correlation with parent and children that correlation will not last for very long
00:42:23
and that good news if you think that you should have high political mobility in the country the problem is if grandparents or great parents can have this kind of direct influence on the
00:42:37
great grandchildren or grandchildren then it can be shown that these two generation view are going to underestimate the overall persistence in political inequality over time so
00:42:51
there's going to be excess persistence for instance what we're going to see then if grandparents are important we can to see that the actual correlation we can measure in data between grandparents and grandchildren that
00:43:03
should be much higher than the one we should expect on this standard two generation model so that's what we set out to do in a paper that we recently
00:43:16
published and the question is as I said why should it matter obviously state of transmission could work between grandparents and grandchildren you can sort of you can transfer your status to
00:43:29
your grandchildren so it can happen social learning if you spend a lot of time your grandchildren you can act as a teacher and role model in the same way as parents can do so in terms of these
00:43:42
two mechanisms I would argue that the difference between grandparents and parents is not one of kind it's more one of the degree they spend less time with the children it's traditional art we should
00:43:57
not sort of need to consider genetic Transmission in this case because obviously transmission of genes are always passing the parents so once we control for the identity of parents
00:44:10
there should no longer be any relationship between grandparents voting and the children's voting but there is this influential argument by
00:44:24
economy economic historian Gregory Clark will argue that that is not completely true he argued that even if we have this kind of genetic transmission but we have genetic
00:44:36
Transmission in things that we cannot observe then that can give rise to this kind of exist persistence in different kind of Mobility outcomes so this is the model I look at
00:44:50
my watch and I see should not get into the details but if you're interested you should read Gregory Clark you can read the paper try to explain this but really what they saying is that if we have this kind of model where we actually control
00:45:04
for parents Behavior then we could find an effect of grandparents behavior on the grandchildren's due to genetic
00:45:17
transmission if the genetic transmission follow this particular form that Gregory Clark portrays so there is really no way and I guess that the problem there's no way if you just have date on these three
00:45:30
generations if you run the regression the control for parents we see that the vter turnout of the grandparents that matters for the grandchildren there is no way according to the data to say
00:45:43
whether or not that depends on the fact that the grandparents are actually influencing their grandchildren for instance through social learning or if it's due to this
00:45:55
kind of sort of genetic inheritance of latent traits so what we try to do in this study was to first of all ask the question is there evidence of this kind
00:46:10
of exist persistence so is the sort of for inst the correlation in voting between grandchildren and grandparents is that larger than what should expected
00:46:21
from this simple two generation model of voting transmission and given that we found that we were also interested in trying to say can we say something about a machanism giving
00:46:35
rise to that exist persistence and since coming from Sweden have the St have used the Swedish administrative registers we have turna data for in this case five election 1970
00:46:46
82 94 2000 2 2010 2018 we link all individuals born 1970 2000 to the parents Grand parents and in some cases even great grandparents so we
00:47:00
can for about 1 million children we can link them to three previous generations we get this large panel of four generations or we have two 7 million Unique Individuals who we can L link to
00:47:14
their parents and grandparents and then we try to measure the voter turnout and see to what exent is that this kind of exist persistence so the first way we do that is to estimate the standard how to
00:47:27
regress a model it's nothing fancy here that is just the the turnout of the child this is the turnout of the parents this is the turnout of grandparents
00:47:39
these are socioeconomic controls trying to get away later on for this status transmission and stuff like that so if we just look at so first the B variant transmission coefficient and
00:47:52
here we have standardized them you can think of these as sort of partial correlation coefficients so here we see that the sort of the transmission coefficient from parents to children is about
00:48:06
0.25 so if this were the entire story then we should then expect that the correlation between grandparents and grandchildren should be about 25 raised
00:48:18
to the^ two that's about 0.625 looking at actual correlation we see that is about1 so that is the first indication that this kind of standard
00:48:33
two generation model is going to underestimate the long run stability in political participation across these different Generations so he's just saying that if you just looking at
00:48:45
parents and children you're going to underestimate the overall persistence in political participation in the long run then we see what happens if we enter these two in the same model and
00:48:58
obviously that some of this correlation between G grandparents and grandchildren is due so of the parents so what we see here is once we do this
00:49:10
control the transmission coefficient from grandparents or grandchildren is decreasing from1 to about 0.05 so about cut in half so half is going through the parents but the other half is still sort
00:49:23
of going directly in this case from grandparents to grandchildren here we control for the socioeconomic status of all generations and then this correlation coefficient
00:49:37
reduces by about a quarter so about a quarter then could be attributed to this kind of social transmission whereas the other sort of
00:49:48
remain some kind of direct result of this kind of exist persistence in political participation across Generations then we did the same for our
00:50:00
larger sample including four generations there we see that in the B variat this blue one is the transmission confient from great
00:50:13
grandparents to their great great grandchildren just looking at the standard here 25 raised to the
00:50:27
power of raised to the power of three that's going to be a rather small number I think it's 0.1 or something the actual ceration here is 0.4 so sort of the
00:50:41
persistance that we see across these four generations is about three to four time as big as the one we should expected for this standard two generation model so we can still say
00:50:53
that this impact is not huge this it's not that whether or not you vote is completely determined whether or not your grand great grandparents used to vote but we still see that this
00:51:06
persistence is much higher than the one we should expect from this standard two generation model and it also sort of say that these more distant Generations can be important once we control for this
00:51:19
other stuff most of it goes away and it could be due to the fact that social learning is not the main mechanism because very few of us spend a lot of time with our
00:51:31
great grandparents so they could be more about other things but what we see here is that and what we argue here is sort of this stand that two generation model that have
00:51:43
dominated much previous research including my own research on this topic is a problematic from the stand point they tend to underestimate the overall persistence in political inequality over
00:51:57
time so if a generation is hit by a shock that make them much more likely to participate that shock are staying on for much longer than we should expect
00:52:11
that based on this standard two generation model so that's another thing I would argue that in order to understand how political inequality is reproduced and maintained in democracies
00:52:25
we should also allow for the the fact that this kind of inequality can span multiple generations and in that sense it make it even more difficult problem to combat because it's not off enough to
00:52:38
Target just the parents in order to do something you might also need to Target the grandparents this recording great grandparents even though I thought this result were interesting I should also
00:52:50
argue they are less important so it's not that great grandparents once you really look at the data they do do not look super important but the grandparents are I would argue the
00:53:03
problem with this results is that they do not say very much about what this excess persistent come froms does it come from this kind of genetic transmission of the Clark type
00:53:17
where parents transfer this lat and trades to the children or are they more due to social learning and sort of direct influence the fact that we as Grandparents can influence our
00:53:29
grandchild grandchildren to behave in a certain way the problem is with this kind of data we have we cannot distinguish these two both of them going to give rise to this positive second
00:53:43
order alter regressive coefficient so we need to find some other way to try to disentangle that we need to similar to before we need to shortcut either the genetic pathway or or we need to
00:53:57
shortcut the social pathway and see what happens and I guess you can already guessed one of the things we do we're going to reuse the adopted children
00:54:09
because then we can shortcut the genetic pathway now we're also going to use the natural adoptions because we have more date on that so they're simply going to sort of shortcut the genetic link between grandparents and
00:54:22
grandchildren but we also going to do something else we're going to study sh whose grandparents died before they were born because then they can't interact so if we think that interaction is important then you should not then we
00:54:36
should see a bigger effect for the grandparents who are actually alive and spend time with a grandchilden than if you're only looking at grandparents as such and if you do
00:54:50
that we have a different sample these are adop so we see first of all we see that the the correlation transmission coefficient parents is about cut in half similar to before just single the fact that so the
00:55:03
transmission app parents is not only due to the social environment it's also due to some kind of genetic factors and we see that this green one that is the the
00:55:15
impact of grandparents in the adoptive sample and that reduces somewhat here in the B variat goes from it was 11 now is 3 03
00:55:29
something so it's 23 2/3 of a reduction in that coefficient here we see this is the blue one is the difference in transmission between
00:55:42
living and dead grandparents so this is another way to try to make it comparable so this is the extra boost to get from having living grandparents rather than just having dead ones so
00:55:55
that's blue bar should be to due to this kind of social transmission and sort of interaction with the grand shield and we see this these green and blue ones are very similar in all models so there
00:56:09
are two different ways to try to get at the direct influence of grandparents on grandchildren but they provide very similar results and the overall message is that there's
00:56:24
still remaining in impact of grandparents on the grandchildren's voting but that is smaller than it was before it's about a third to 40% of the
00:56:36
initial estimate we had on the previous slides and one reason for that is that part of that again are due to these kind of sort of biolog biological pre-birth
00:56:49
factors rather than social factors but the social environment is also important somewhat less important for grandparents than it was for parents but again that is reasonable because as parents we
00:57:01
spend more time with our children than we do as Grandparents so I think that makes sense that that's is the fact so looking at the time I see a type to to to wrap up so that is what I do
00:57:15
but as I said in the beginning or might not have said that but in pre-democratic societies the identity of one's parents were the key to political power
00:57:26
that is the chances to participate and influence politics was mainly an accidental birth the promise of democracy when it was first introduced 100 years ago in countries like Sweden
00:57:40
and Finland was that democracy going to put an end to that family background was no longer to be the key to political power and influence but if you look at
00:57:55
the data the of data presented for you here today we see that that promise has yet to be fulfilled we can see the bush the gandhis the trors and the leens there
00:58:09
are signs that family background are still matter for political influence and political participation and to some extent which we also seen during this lecture and try
00:58:21
to understand that is not that surprising because these are very complex phenomen that depends on a lot of things I've tried to show that it both involves multiple Generations we
00:58:34
have seen that this kind of persistence can have both social and biological Roots so obviously going to take a lot to change this for real but just because
00:58:46
something is difficult doesn't mean that something is impossible and that's what we're going to learn tomorrow for tomorrow I said I promise it's going to be a bit more constructive then we're going to try to see are the ways we can
00:58:59
try to mitigate the importance of family background in modern democracy and I can just tell you that the answer going to be a partly a yes
00:59:11
not completely yes but partly yes but if you want to know more about that then you have to be here tomorrow as well so for those of you that are interested see you tomorrow
00:59:22
[Applause] [Music]
End of transcript