Rivero et al
With One Child Here and One Child There: Is There Specialization and Complementarity in Children’s Support Related to Their Place of Residence? The Mexican Case
Estela Rivero (El Colegio De Mexico), Anairis Hernandez Jabalera (El Colegio de Mexico), Maria Dolores Puga (CSIC)
Hi,
I have couple of doubts regarding the paper of Estela Rivero (El Colegio De Mexico), Anairis Hernandez Jabalera (El Colegio de Mexico), Maria Dolores Puga (CSIC)named “With One Child Here and One Child There: Is There Specialization and Complementarity in Children’s Support Related to Their Place of Residence? The Mexican Case”
It the beginning I must appreciate the uniqueness of the idea and the relevance of the study in understanding the situation of elderly. A few doubts which I have are:
1. As per mentioned in your presentation, the data you have used is of 2001, which is more than a decade old. So don’t you think after the global recession there has been some affect on the trend of migration of young adults for employment and hence in the structure of specialization of support for elderly, which actually is a very important phenomenon not captured by your study. It would be better if you used some more recent data.
2. More over in your analysis you have shown that “having migrant children, children living in the same city, or a combination of both increases the chances
of receiving economic support”, which is very rational ad understandable. But this variation in receiving economic support can vary vastly based on the type of employment the children are part of (i.e. salaried, casual, informal), which if studies in details can give a more insightful outcome from the study.
It is indeed a wonderful paper and thank you for sharing such quality work. Wish you many more success.
Regards
Suchandrima Chakraborty
India
Dear Suchandrima,
Thank you very much for reading our paper and your very interesting comments, which are really useful. We agree with you that the data we used is quite old and migration dynamics may have changed in a decade. This has indeed happenned in Mexico, where international migration has decreased and return migration increased, among other things.
I see your comment as having two main implications for how changes in migration could affect our results:
1) How the volume of migration (and hence the number of migrant children per family has increased or decreased in this period), and as I mentioned earlier, this has indeed changed. But not as importantly as to contrarrest our results.
2) How migrants’ selectivity or the economic situation of migrants has changed. For example, if now the migrants have less chance of finding a job than before because of their characteristics, they may be less able to compensate their siblings for the time they spend with their parents. This has also happened, and is one of the reasons behind return migration.
How much does this affect our results? We can’t really know. Unfortunately the data we use was, at the time we wrote the paper, the only data that allowed us to make this analyisis. However, there is now a new dataset (a 2011 wave of the same survey) that will allow us to see how things have changed.
As for incorporating the work that children do: I think that is a great idea, thank you very much!
Thanks again,
Estela
Nice paper and great to see data that allow you to develop measures of residence beyond the household.. a couple of questions:
1) is there anyway to conduct within sibling set analyses to lend more robust support to your conclusion that children organize to support parents…you show aggregate results but wouldn’t it be more effective to examine within sibling sets?
2) you need to be careful about endogeneity issues given you are dealing with cross sectional data
3) I presume the 4 types of help are not mutually exclusive (though this is what the modeling suggests) so would it be useful to also examine “intensity of help” (on a scale of 0-4 with 0 being no help and 4 being all types of support?)
Just some suggestions..
Dear Sangeetha,
Thank you for thanking an interest in our paper and for your comments. Here are our responses:
1) Yes, we can conduct within sibling set analysis, as we can identify who are the children of the same household. We have thought about doing this as a next step and from the perspective of the children. However, one of the problems we are encountering when thinking of this alternative is how to structure or construct the sibling-network analysis. We thought about doing match-ups of sibling-sets of the same size, but different conditions, but this is almost impossible!
2) You are right, thank you for the advice. The MHAS is a panel survey and has a follow-up wave in 2003 (and now another one in 2011). We have been thinking about doing a longitudinal analysis. However, this introduces another level of complexity to the analysis, as children who were migrants in one wave may not be migrants in the 2nd and the such. This is important in itself, as represents lack of stability in the networks, but complicates the analysis. We are thinking of this as a first-step, calling attention paper, and would like to explore the rest of these issues later.
3) Effectively, the four types of helps are not exclusive and we analyze them independently (in separate logit models). I just realized that this may be confusing in the paper and the presentation. The suggestion of having a general indicator of help is great, thanks.
Thanks again for your comments,
Estela
Dear Estela Rivero and colleagues:
I am pleased that you chose to present this research. I am a doctoral candidate and I am studying similar topics in the the context of Latin America and the Caribbean using the SABE dataset, which includes Mexico City. I like the direction of your paper and I think this is a very important area of research that will be given even more attention in the coming years. I have a few questions for calrification.
1) Why limit the sample to older adults with at least one disability as opposed to examining the enitre sample and choosing disability as a mediating or moderating factor? I am more curious on how children support their parents overall, according to their distance, and then how this can be adjusted by parental vulnerability (in the case of disability for this paper)?
2) Does the MHAS provide any information on the economic and demographic circumstances of children? For instance, employment, income, marital status, whether they have children? These can all influence children’s likelihood of providing support to thier parents (money and time) and hence parents’ receipts of support with or without disabilities.
3) The MHAS is a longitudinal study, can you explain why you opted for just one cross-section?
4) Does the MHAS indicate the specific country in which migrant children live? This question is jolted by the finding that parents with migrant children only have lower odds than those with children in multuple locations to receive economic support. It makes sense that having children in multiple locations is associated with higher odds of receiving economic support as there is more variation. I wonder though if this can be based on children/siblings assisting each other based on the conditions in thier respective locations.
5) Regarding the complementary or specialization objective, I think you would it more useful to analyse the data from the children’s perspective and assess whether having siblings within a given location affects the odds of support that children provide/parents receive?
6) Other suggestions, can include examining the wealth of the household, as older Mexicans rely more on assets, and how this influences support transfers. Also including parents’ provision of support to children, regardless of thier location.
I hope this is not too overwhelming. I will love the opportunity to talk to you more about your work as we are studying similar topics.
Estela and Anairis, nice paper! I agree with comment 5. I think that carrying out the analysis from the adult children’s perspective could give a very useful insight into the matter and it would complement the existing analysis. My suggestion would be to try to link both perspectives i.e. household & adult children.
Dear Gabriela:
Thank you very much for your suggestion. We intent to do this as one of our next steps, as we think that you can’t understand specialization or complementarity without seeing what the children are doing. As you say correctly, both perspectives complement each other. Here, by taking the perspective of the parents, we wanted to show whether parents with children in different locations –or children who are migrants- may be more vulnerable. The following step would be to analyze the children and analyze whether children who have less siblings, or who have less sibling in the same location do more caregiving work.
Thanks,
Estela and Anairis
Dear Nekehia,
We are very glad that you enjoyed our research, and more than pleased to learn that we are studying similar topics. We would certainly like to communicate more with you if you want to. I am sure the cyberseminar organizers can give you our emails.
Here are our responses to your questions, we hope they are useful:
1) Why limit the sample to older adults with at least one disability as opposed to examining the enitre sample and choosing disability as a mediating or moderating factor?
We made this choice because older adults with at least one disability are the ones that are in most need of receiving help. In addition, while it would great to see how children organize to take care of their parents in general, it is older adults with at least one ADL limitation who receive help in this aspect, and the same happens with IADL. If we had not considered this limitation, we would be left with economic and household chore help, and we would not be getting the whole picture. Now, introducing limitations as a mediating factor is an option, but limiting the sample to this group from the beginning is a better control, making the group more equal.
2) Does the MHAS provide any information on the economic and demographic circumstances of children? For instance, employment, income, marital status, whether they have children? These can all influence children’s likelihood of providing support to thier parents (money and time) and hence parents’ receipts of support with or without disabilities
This is a good question, and yes, the MHAS provides some information on the economic and demographic circumstances of the children. For non-resident children this is limited, but would allow one to characterize what life-cycle stage they are in, their education level and whether they are employed. However, since our analysis is not at the children level, we do not use this information and including it in our analysis would complicate it too much. We limit our analysis to adult children who are in working age, which is enough control, we think.
3) The MHAS is a longitudinal study, can you explain why you opted for just one cross-section?
Certainly, the fact that the MHAS is a longitudinal study opens multiple possibilities. In this particular case, however, we thought that using the second wave (and the third wave which became available when we were finishing the study) would only complicate this part of the study. This is because to analyze the question we are trying to solve would require finding sibling sets that do not change in anything, but in having one sibling migrating and being in another location long enough to send money. This is extremely complicated to find. We would end up with a very small sample. As I answered to Sangeetha before, if anything, using the panels gets mixed with sibling-sets instability.
4) Does the MHAS indicate the specific country in which migrant children live? This question is jolted by the finding that parents with migrant children only have lower odds than those with children in multuple locations to receive economic support. It makes sense that having children in multiple locations is associated with higher odds of receiving economic support as there is more variation. I wonder though if this can be based on children/siblings assisting each other based on the conditions in thier respective locations.
Yes, we know whether children migrated to the US –the main destination for international migrants in Mexico-, or whether they live in Mexico in a different city than their parents. If they are internal migrants, however, we do not know where they live.
Might this be that siblings are helping each other? We do not know, but this is money that is flowing to the parental home.
5) Regarding the complementary or specialization objective, I think you would it more useful to analyse the data from the children’s perspective and assess whether having siblings within a given location affects the odds of support that children provide/parents receive?
Yes, we are planning on doing this as well, thanks!
6) Other suggestions, can include examining the wealth of the household, as older Mexicans rely more on assets, and how this influences support transfers. Also including parents’ provision of support to children, regardless of thier location.
You are right, we did not include parental wealth. This is important.
Thank you, Nekeia, and best of lucks with your dissertation,
Estela
Estela and colleagues:
Thanks for your responses. I will contact the organizers to get your emails. I do want to keep in touch and continue such conversations. Take care. Nekehia
Dear Nekehia,
(I am posting this line here because I made a mistake and posted the same response in another discussant’s comments. The website was too intelligent and detected the duplication, prohibiting me the second posting.)
We are very glad that you enjoyed our research, and more than pleased to learn that we are studying similar topics. We would certainly like to communicate more with you if you want to. I am sure the cyberseminar organizers can give you our emails.
Here are our responses to your questions, we hope they are useful:
1) Why limit the sample to older adults with at least one disability as opposed to examining the enitre sample and choosing disability as a mediating or moderating factor?
We made this choice because older adults with at least one disability are the ones that are in most need of receiving help. In addition, while it would great to see how children organize to take care of their parents in general, it is older adults with at least one ADL limitation who receive help in this aspect, and the same happens with IADL. If we had not considered this limitation, we would be left with economic and household chore help, and we would not be getting the whole picture. Now, introducing limitations as a mediating factor is an option, but limiting the sample to this group from the beginning is a better control, making the group more equal.
2) Does the MHAS provide any information on the economic and demographic circumstances of children? For instance, employment, income, marital status, whether they have children? These can all influence children’s likelihood of providing support to thier parents (money and time) and hence parents’ receipts of support with or without disabilities
This is a good question, and yes, the MHAS provides some information on the economic and demographic circumstances of the children. For non-resident children this is limited, but would allow one to characterize what life-cycle stage they are in, their education level and whether they are employed. However, since our analysis is not at the children level, we do not use this information and including it in our analysis would complicate it too much. We limit our analysis to adult children who are in working age, which is enough control, we think.
3) The MHAS is a longitudinal study, can you explain why you opted for just one cross-section?
Certainly, the fact that the MHAS is a longitudinal study opens multiple possibilities. In this particular case, however, we thought that using the second wave (and the third wave which became available when we were finishing the study) would only complicate this part of the study. This is because to analyze the question we are trying to solve would require finding sibling sets that do not change in anything, but in having one sibling migrating and being in another location long enough to send money. This is extremely complicated to find. We would end up with a very small sample. As I answered to Sangeetha before, if anything, using the panels gets mixed with sibling-sets instability.
4) Does the MHAS indicate the specific country in which migrant children live? This question is jolted by the finding that parents with migrant children only have lower odds than those with children in multuple locations to receive economic support. It makes sense that having children in multiple locations is associated with higher odds of receiving economic support as there is more variation. I wonder though if this can be based on children/siblings assisting each other based on the conditions in thier respective locations.
Yes, we know whether children migrated to the US –the main destination for international migrants in Mexico-, or whether they live in Mexico in a different city than their parents. If they are internal migrants, however, we do not know where they live.
Might this be that siblings are helping each other? We do not know, but this is money that is flowing to the parental home.
5) Regarding the complementary or specialization objective, I think you would it more useful to analyse the data from the children’s perspective and assess whether having siblings within a given location affects the odds of support that children provide/parents receive?
Yes, we are planning on doing this as well, thanks!
6) Other suggestions, can include examining the wealth of the household, as older Mexicans rely more on assets, and how this influences support transfers. Also including parents’ provision of support to children, regardless of thier location.
You are right, we did not include parental wealth. This is important.
Thank you, Nekeia, and best of lucks with your dissertation,
Estela
Hi,
I have a few questions on this interesting study:
1. Why did you restrict the sample to “household where at least one of the individuals had at least one basic disability” and “households with at least four adult children”? I did not see any reason which was mentioned in your paper about this sample selection. And you mentioned in the paper that “If we consider migration and the decrease of fertility, many of these will depend on fewer children and many families will be geographically disperse”. In my opinion, it will be also very interesting in looking at the elderly caring in the households with only one migrant child and no disabled parents.
2. I’m not sure whether the dataset you are using in this paper including information on the migrant children’s siblings and other relatives. But elderly care can be provided by other relatives instead of offspring. Was support from other relatives taken into account in considering the relationship between children’s resident location and elderly care?
Thanks!
Hi Fei,
Thank you for your questions. They will be very useful when reviewing our paper. Here are our answers:
1. In my opinion, it will be also very interesting in looking at the elderly caring in the households with only one migrant child and no disabled parents.
We restricted the sample to households were with at least one disability because these are the ones that are in most need of receiving help. In addition, while it would great to see how children organize to take care of their parents in general, it is older adults with at least one ADL limitation who receive help in this aspect, and the same happens with IADL. If we had not considered this limitation, we would be left with economic and household chore help, and we would not be getting the whole picture. Now, introducing limitations as a mediating factor is an option, but limiting the sample to this group from the beginning is a better control, making the group more equal.
With respect to the other part of the question, we restricted the sample to families with at least four children, so we could compare families in the same conditions and that could be in any of the living-arrangement categories. The minimum number of children that we need for that is three children (one children to live in the parental household, one to live in the same city as the parents, and one to be migrant). Originally we had divided between internal and international migrants, but we did not find any difference between these two groups, and the division held.
2. You are right that elderly care can be provided by other relatives instead of offspring. It can also be provided by friends and neighbors. Other studies in Mexico (Rabell 2009) show that friends and neighbors provide monetary support and eventual help, while our own work shows that partners may provide help with IADLs and ADLs. However, when present, children are commonly one of the main providers of help, especially as parents grow even older (Rivero 2011). Our past research has also shown that families rely on multiple hands or sources of help (commonly from children –Rivero 2011). That is why we are interested in seeing what happens when the children are geographically disperse.
To control for other external factors, we did not consider other sources of help. But now that we mention it, in the models controlled for whether the parents are widowed or divorced or not.
Thanks again for your useful comments,
Estela and Anairis
Dear Gabriela: (Repply to discussant)
Thank you for your suggestions.
The idea of being more specific about the concepts we used is great. For example. Help with activities of daily life and help with instrumental activities of daily life, are commonly-used terms, especially in the literature about aging. But you are right, explaining how we operationalize these, specifically with the data provided by the survey would be very useful.
With respect to why we limited the sample to families with at least four children, we did this so we could compare families in the same conditions and that could be in any of the living-arrangement categories. The minimum number of children that we need for that is three children (one children to live in the parental household, one to live in the same city as the parents, and one to be migrant). Originally we had divided between internal and international migrants, but we did not find any difference between these two groups, and the division held.
In order to give the reader an overview of Mexican families with disable parents, it would be
informative to present the proportion of households in each family type in 2001. This is a very good suggestion. We will present descriptive statistics, and describe them.
I would have liked to see the results from the full model where authors controlled for
possible cofounding variables on figure 2 (as an Appendix). You are right, we will present this in the paper.
The suggestion of presenting a research agenda is excellent, as the implications of this paper go well beyond Mexico. We will do this in the conclusions.
Again, thank you very much for your kind suggestions. I am sure they will help us to improve the paper.
Estela and Anairis