On Estimates of the Cost of Raising Children used in Child Support Tables
Roger F. Gay, Project for the Improvement of Child Support Litigation Technology
November 2, 1999
Page updated on July 20, 2001While surfing the web today, I ran across a rather disconcerting commentary at the Alabama Office of the Courts site. The page is entitled; CHILD SUPPORT INFORMATION. (Link updated on July 20, 2001 -- old link did not exist -- page moved.) Before revealing the details, I should note that Alabama is presenting the same misinformation as do many other states. The quote you find below is quite similar to a statement made in, for example, the Indiana statutes.
One would think that the Alabama Office of Courts (AOC) would do its best to provide the public with accurate information. Yet, I immediately found obvious misinformation and wrote to AOC to tell them about it.
The State of Alabama uses a child support guideline derived from a proposal developed by child support collection entrepreneur Robert G. Williams, of Policy Studies, Inc., known as the Income Shares model. The model has received continuous criticism since its adoption as one that is arbitrarily contrived, not based on any valid economic evidence, and because it does not correspond to any set of rational principles for making a child support award.
The Schedule of Basic Child Support Obligations, as it is referred to, is a table of numbers. In making a child support decision, the court is instructed to use the parents' combined income to look up a number in the table. The corresponding number is presumed to be the dollar amount that the parents should spend on their children. (It can then be adjusted for various reasons.) The "basic obligation" for the payer is part of that amount in proportion to the payer's income. AOC says the following to describe the table.
The National Center for State Courts has never sponsored development of a Schedule of Basic Child Support Obligations based on extensive economic research on the cost of supporting children at various income levels. As a matter of fact, they have never sponsored extensive research on any aspect of child support guideline design, with the exception perhaps of state by state reviews of political decisions. Certainly nothing however, of a technical or scientific nature. The Schedule of Basic Child Support Obligations was developed through research sponsored by the National Center for State Courts and is based on extensive economic research on the cost of supporting children at various income levels. The information being referred to is a proposal for design of guidelines by child support collection entrepreneur, Robert Williams. Williams did incorporate results from a "cost of raising children" study in his proposal. A far cry from "extensive economic research," Williams relied on a single study that was unrelated to the child support award question. The study was conducted over a summer by a sociology professor, Thomas Espenshade, who had no previous experience in making such estimates, and his work has not since received any positive recognition from the scientific community.
Espenshade has not himself received much criticism for the work. Other than its use by Williams in recommending values for child support tables, his work on estimating the cost of raising children has not received much attention at all. A far, far cry from "extensive economic research," Espenshade seemed to have had it in mind to get his feet wet in something new over a summer break. In point of fact, Espenshade isn't even an economist. His specialty isn't families or children either. It's immigration.
Espenshade's study was not about making child support awards. He did not have the data or methods necessary to calculate what parents at different income levels spend on children. He presented results that in quantitative detail have no validity, but his final purpose and point were quite simple. The fact that his study was not about child support awards, and that he did not push his numbers for such political use is probably the reason that more criticism has not been focused on him personally by advocacy groups.
Espenshade wanted to help train parents, and particularly those parents who had little to no experience dealing with their own financial issues. This would of course include immigrants from poor countries and even young parents in the United States. His study, although providing speculative inaccurate numbers, made a simple but important point. If you have children, it will cost you. He didn't really know how much.
Let's say for example that you get married and the two of you have enough to buy a sports car. You want the sports car but you are also considering having a child right away. Espenshade's results say that if you have the child, it may be impractical to buy the sports car. In concept, the difference between being able to afford the sports car if you don't have the child, and not if you do, is what Espenshade defines as the "cost of raising children."
In fact, Espenshade compared expenditure on food in proportion to the number of family members between couples with and without children. What he chose to view as the standard of living difference between couples with and without children represented his cost of raising children. This aspect of the method itself receives criticism even from those who believe any such comparisons can produce worthwhile results.
The data available for his study did not contain the information needed to determine what parents spend on children. In the end, Espenshade put a fudge factor into his equation to produce his result. The numeric result is not a statistical result based on the data. The result is not how much parents spend on children. The result is not accurately described as the cost of raising children. The result is primarily a result of his method and his fudge factor, not the data, and is not a reflection of anything special.
Again however, Espenshade's results have not received much criticism. His study has not really received much attention at all, except in the political context created by Robert Williams. Espenshade never claimed to have produced a set of numbers appropriate for use in child support guidelines.
The AOC also claims that the "extensive economic research" on which their table is based, is on the cost of supporting children at various income levels. Espenshade was not able to refine his results to differences in spending at different income levels. He did not select a different fudge factor for various income levels. He produced a table of percentages of the amount of total spending by families, which he attributed to the "cost of raising children."
The effect of the data on his results was so small that the difference in the percentage between low-income families and high-income families was insignificant. In other words, his detailed numeric results were saying that a wealthy family will suffer the same percentage standard of living loss to raise two children as a poor family. This in no way reflects reality, but only the fact that the same method and fudge factor were applied regardless of the income level. The result illustrates that the family spending data used had very little effect on the result.
The use of Espenshade's results in child support guidelines is apparently more the fault of Robert Williams than of Espenshade himself. However, both Espenshade and his publisher do deserve some measure of blame.
In his book on the study, Investing in Children, Dr. Espenshade goes to great lengths explaining what he would like to achieve in his study. It is clear that he wants to provide accurate information on what children cost parents. He discusses his methodology in somewhat philosophical ways, expressing the logic of what he is trying to do.
He does not however, explain that he has failed to do it. He does not mention the inclusion of a fudge factor in obtaining his final results. As unsatisfactory as his attempt was, he does not tell the reader. He does in fact mention the possibility of applying results of studies such as his to the problem of making appropriate child support awards.
We should also wonder why his publisher, The Urban Institute Press, published his study in the form of a book with a snappy title, and why it was promoted as though he had succeeded in obtaining the results he was looking for. A publisher caused similar confusion in the past by promoting a book called The Divorce Revolution with false information. The Urban Institute however, is a research organization. That being the impression they project to the public, shouldn't we expect a serious in-depth peer review of their publications?
As I have said though, the choice to recommend child support tables related to Espenshade's results was made by Robert Williams. It is clear that the results were not realistic enough for use in anything but casual conversation; to make more of a non-quantitative point. Or as University of Chicago economist Edward Lazear put it (less delicately);
. . . the presumption that underlies the focus of much of the empirical research and policy debate on income distribution [spending within families] seems born of ignorance and is supported by neither theory nor fact.
(Source: Edward P. Lazear and Robert T. Michael, Allocation of Income Within the Household, University of Chicago Press, 1988, page 25)
But let's pretend that Espenshade's results were exactly what Espenshade was looking for. Would they then be appropriate to use them as the basis for child support awards?
Let's continue with the conceptual explanation given above. The "cost of raising children" is that a particular family cannot afford to buy the sports car they wanted because they are providing for a child's needs instead. The result of having a child in that situation is that the parents, living together in an intact family, cannot afford the sports car.
But in application in child support guidelines, what Williams is saying is that the mother only needs to divorce. Her ex-husband will then owe her enough to take care of the child and his share of the cost of buying the sports car to boot -- to bring them to the standard of living they would have had together if they had not had children. That, in the application of Espenshade's conceptual theory is what it would take for one parent to reimburse the other for "the cost of raising children."
The Alabama Office of Courts as well as other government sources in other states misrepresent the numbers used in their child support tables. They are not is based on extensive economic research on the cost of supporting children at various income levels. Even if the statement were true, the goal it represents does not correspond to any rational view on calculating an appropriate child support award.