top of page

Forecasting the Impact of Abortion Law Changes on State Foster Care Systems

Updated: Oct 7, 2022

The social and cultural impacts stemming from changes to abortion laws following the 2022 Supreme Court decision in the case of Dobbs vs. the Jackson Women’s Health Organization will play out over the course of the next several generations. There are a host of data projects that could be conducted investigating the likely effects of this ruling, and this project is just one example. Here, I investigate the likely impact of the Dobbs decision on state foster care systems in states that have, as of the time of this writing (October 1, 2022), effectively outlawed abortions.


Overview

This forecast is based on the core premise that abortion restrictions will result in more births meaning child populations in states that have restricted abortions will grow much faster in the coming years than they would have otherwise. Assuming that at least the same proportion of this added child population will require foster care services as state child populations did pre-Dobbs, the need for foster care services will be meaningfully higher than would have been the case if abortion was still legal.


I estimated the magnitude of this increase using the following publicly available data:

I ran the forecast to the year 2040, which provides insight into approximately one generation of children born post Dobbs. There were a few key components of this forecast that had to be built separately and then combined to produce the end result. I proceeded with the analysis as follows:


1. Forecast the overall child population to 2040

This was the easiest part of the analysis. The Weldon Cooper Center already provided population estimates for the years 2030 and 2040 separated out into age categories every 5 years (e.g., 0 to 4, 5 to 9, etc.) and included a similar projection from 2020 that was validated against actual data from the 2020 census. I simply used these forecasts and assumed the changes over the course of each 10-year chunk were distributed evenly across the decade.


2. Estimate the proportion of each state’s overall child population that required foster care services

I combined data from the Annie E. Casey foundation about the historical need for foster care services in each state with Kaiser Family Foundation data about the historical population in each state (again, segmented by age groups) to estimate the total proportion of each state’s child population that required foster care services in the recent past. This also generated an estimate of the amount of variability there was in this proportion from year to year.


3. Forecast the pre-Dobbs need for foster care

Based on the information from steps 1 & 2, I built a baseline forecast of what the foster care need in each state would likely have been had there not been any changes to abortion law, out to the year 2040 (See this related post for an in-depth exploration of incorporating uncertainty in the forecast through simulation). Using Texas as an example, you can see that foster care need was projected to rise between now and 2040 simply based on forecasted population growth and trends in the proportion of the state’s children in foster care up to the year 2019:

4. Forecast annual abortions - if they were legal

Reliable data about abortions is hard to come by, and I was only able to track down state counts of abortions for 2019 and 2020. This data was based on voluntary reporting, so it almost certainly represents an undercount of the true number of abortions. In order to incorporate an appropriate amount of uncertainty into the data, I combined the abortion counts with birth counts from each state to get an estimate of total pregnancies, and then divided this total pregnancy count by the total population of adults aged 19-34 to arrive at a (very) rough estimate of the overall rate of pregnancy per adult of child-bearing age in the state. A more accurate measure of the pregnancy rate in each state would have been great, but what really matters is that we have a way to convert changes in the adult population of child-bearing age over time into projected changes in the number of pregnancies. The above data lets us do that pretty well.


Using this estimate of the pregnancy rate, the proportion of abortions to overall pregnancies, and the forecasted adult population (aged 19-34), I was able to forecast the number of abortions each year in every state through the year 2040 (again, with uncertainty incorporated). This forecast turned into the estimate of the number of additional children that are likely to be born each year in states that have outlawed abortion, over and above what was originally forecasted based on the Weldon Cooper data, simply by assuming the forecasted abortions would become births. Here is the plot of the forecast for Texas:

5. Forecast additional foster care need based on additional births due to abortion restriction

I used two different approaches to arrive at this estimate: 1) I assumed that the likelihood that any child would require foster care services was equal across all ages of children within a state, and 2) I assumed that the likelihood of requiring foster care services was age-dependent. I think the latter scenario is more likely to be the most accurate, because historical records tend to show that state foster care populations have higher concentrations of younger children, but I ran both for the sake of comparison (the difference is subtle, but meaningful).


The plots below show the baseline forecast of foster care need in Texas as well as the updated forecast, accounting for additional births due to abortion restrictions, across both of these scenarios:


Forecasted Need for Foster Care Services in Texas

In the first scenario (the likelihood of requiring foster care services does not depend on age), more than 43,000 children are projected to need foster care services by the year 2040. This represents an increase of almost 4,400 children (+11%) from the 2040 projection prior to the change in abortion law. The second scenario projects about 42,200 children in the Texas foster care system by the year 2040 (an increase of ~ 8% over the baseline forecast). The age-dependent forecast produces an increase in need that rises earlier and starts to level off closer to 2040 relative to the non-age-dependent forecast (plot on the left). This is because the Texas foster care population has historically had a higher concentration of younger children, so the need for foster care services will likely increase most rapidly in the years immediately following the Dobbs decision in Texas.


Conclusions

Any increase in the need for foster care services is important, and this forecasted level of change represents a significant rise in the need for services over the next 18 years. At this point, is important to consider two things: 1) The available data on the number of abortions performed is likely to be an undercount, and 2) the cohort of children born to mothers who would have had an abortion, if it was legal, may very well have a higher likelihood of needing foster care services compared to children born pre-Dobbs. There is no data that can provide guidance with respect to these possibilities, but it is noteworthy that the forecasts above are surrounded by an 80% confidence interval, representing the uncertainty about the forecast. If we think that there is a fair chance that our main estimate is low, then we should focus attention on the upper end of this confidence interval, where we can see that it wouldn’t be out of the question for the foster care population in Texas to approach 50,000 children by 2040 because of the change to abortion laws.


The table below summarizes this same information for all states that have enacted total or near total abortion restrictions as of October 1, 2022:

In all cases, the need for foster care services is projected to be higher than it would have been if abortion had not been restricted. The effect of abortion restrictions on the number of children needing foster services is generally somewhere between 3%-10%, depending on the state (although see Georgia, with a projected increase between 14% and 21%). This highlights a significant need for increases in state funding if the demand for foster care services is to be adequately met. States imposing abortion restrictions should begin planning for this increased need now.


Assumptions and Caveats

Forecasting out almost 20 years into the future carries considerable uncertainty, and I’ve attempted to incorporate that uncertainty into the forecast appropriately while at the same time, producing a forecast that is meaningful and actionable. There are a few underlying assumptions that should be stated explicitly because they have an important influence on the forecast:

  • I am assuming that recent, state-level data about births, abortions, and foster care populations provide valuable insight about what will happen in the future.

  • I assume that changes to abortion law will not have a meaningful impact on human behavior. More specifically, I am assuming that the pregnancy rate will not change meaningfully as a result of changes to abortion law, and that the number of children born each year following the Dobbs decision will represent the number of children that would have been born prior to the change in abortion law plus the number of abortions that would have occurred prior to the change in the law.

  • There are several things that will influence this analysis that can't be estimated because there is no historical data that would suggest what kind of effect they will have. Some women who have the means will travel out of state to have abortions, some women will have illegal abortions, and as we get to the end of the forecast range above, some girls who were born as a result of the Dobbs ruling will have children of their own. All of these factors matter, but without data to inform their likelihood and effect, I cannot meaningfully incorporate them into the forecast without a considerable amount of guessing. So I left them out.

  • It is worth reiterating that there is a fair chance that children born to parents/mothers who otherwise would have elected for abortion (had it been legal) will have a higher likelihood of requiring foster care services than children born before abortion was restricted. However, there is no historical data that would inform a better estimate, so I produced the forecast assuming a continuation of the historical foster care need. Because of this, I think it would be wise to focus attention on the upper end of the confidence interval in this estimate.

Restricted access to abortion will have a number of social consequences. In this project, I have focused solely on changes in the need for foster care services, though a similar forecast could be built investigating need for pediatric care, public school resources, food assistance, and a host of other public service programs. All are likely to see near-term rises in demand in states where abortion has been restricted.


Need is forecasted to increase in all states that have enacted abortion restrictions above what would have been expected prior to abortion law changes. It is important to underscore the point that it is the NEED FOR SERVICES that is forecasted to increase. In order for the observed foster care population to actually increase, the supply of resources necessary to meet this need (licensed foster care homes, state funding, etc.) must also exist. The foster care systems in many states tend to be chronically under-resourced, and the number of children receiving services can be limited by the available capacity. It will be important for all state foster care systems to have a robust mechanism for reporting the extent to which the need for services is not being met. That said, states should plan now for increased need in order to provide appropriate support for the children who require it most.


I hope you found this project insightful. As I mentioned, there are a number of similar meaningful forecasts that could be developed to assess changes in demand for state resources as a result of abortion law changes. If you are a state agency and are interested in partnering for a project in your state, please contact CommonWealth Data Solutions (info@cwdatasolutions.com). We would be proud to help.


Thank you,


Russ Clay, PhD

Founder and Principal Data Scientist

148 views0 comments

Recent Posts

See All
bottom of page