recent research on child care

Child care isn't a topic that I usually spend a lot of time on, but I take an interest in it for two main reasons. One is that I completed a paper on the subject for a Canadian politics class a few years ago and so take interest in ongoing research relating to child care. The other reason is that this is a major area of policy debate in Canada these days.

The paper that I'm addressing is Universal Child Care, Maternal Labour Supply, and Family Well-Being, an empirical analysis of universal child care in Quebec, by Baker, Gruber & Milligian, published in the most recent issue of the Journal of Political Economy. The most notable, and surprising, results of the paper were that it claims that the increase in child care resulting from universal child care in Quebec (it had a flat rate of $5 and is now $7 per day) has resulted in an increase in aggression and an increase in health problems among young children. I will show that these results may be erroneous and are in need of further study with improved methodologies.

While I don't deny that there may be hidden social costs, as well as hidden social benefits of such programs, there are a few important points that I have about this paper, which was clearly carried out from a preconceived notion that the effects must be negative. The main outline I provide is to a) tackle the question of bias with respect to the variable 'aggression', b) refute the claim that the health effects are negative, then c) carry on my typical critique of estimated costs and benefits due to overly simplistic analysis of spinoff costs and benefits.

First, the authors don't consider the likelihood that selection bias overemphasized their results. For example, the offer of cheap child care will appeal most to those parents who are already stressed out from an overactive, anxious, or even aggressive child. The authors chose to show their results in terms of standard deviations rather than absolute amounts. While there was a significant rise in the use of child care over the period of the study, the paper does not mention that the absolute increase in 'reported aggression' was only from 1% to 1.8% of total users. If parents with problem toddlers are more likely to use the program, this could explain much of the claimed increase in aggression among children. This intuition is guided by my experience at summer camps, where parents would always offload their kids to camp for their Ritalin holiday. This tells me that some parents are more likely to take advantage of child care options outside the household when they have 'problem' kids.

One could also ask what leads parents to report aggression. The problem is that the economists who wrote the paper are dealing with reported aggression, rather than actual aggression. I can think of a simple explanation here - parents who are at work during the day are less likely to be as in tune with their children. Since they don't see their kids all the time, they may find certain behaviours more surprising and are thus more likely to report aggression. It could be relevant to say that I am not insensitive to the fact that it's not an inherent good for parents to spend less time with their kids, but it's also relevant to say that it could lend itself to being a cause of biased results in the empirical analysis.

Next, I ask, what exactly does this variable mean? What is this aggression? Do they demand more of their parents during the time that they do have with them? The operationalization of a variable is a key step in good empirical analysis, yet the researchers do not tell us in any satisfactory manner how they translated the survey results into their variable label of 'aggression. '

So, I have offered two causes that are intuitively likely to increase reported aggression without real increases in aggression among toddlers (the incentive of universal child care is more appealing to parents who are already stressed with over-anxious kids, and parents will be more surprised by such behaviours when spending less time with them). I have also raised an important question of what the variable 'aggression' means in reality. The overarching questions here are a) what causes parents to report aggression (is the increase a matter of perception or reality), and b) what the heck does this variable even mean? The paper doesn't provide any satisfying answers to either of these central questions.

Another question of interest that the authors touch on and then claim is not likely to be important, is, if the increase in aggression among toddlers participating in the program is real (not just perceived), then what does that mean? Let me start here by saying that I have already shown that the 0.8% is almost certainly an exaggeration.

The authors suggest that shifting the first points of increased social contact from kindergarten to a year or two earlier may simply shift this effect from the first year of school to the first year of child care. This would mean that the long term effect could be a) negative, due to such 'trauma' at a younger age when they could deal with the change less effectively, b) , nothing, or c) positive since children will engage in learning social norms at a younger age. I have no idea which is the case. Further research is needed. However, that research will also have similar difficulties of turning survey results into meaningful data, and they will have to tackle such methodological limitations on the head rather than ignoring them, as these researchers do.

The authors also draw attention to the negative health effects of child care. The primary variables that inform this conclusion are a greater incidence of the flu and colds among these children. There is one major reason to doubt this as a negative outcome. Simply put, exposure to pathogens is required in order for a child to develop a strong immune system. This is similar to the logic of a possible long term improvement resulting from social stresses, but is would be easier to verify empirically. This research should be left to epidemiological researchers, not economists. There are increased health care costs associated with visits to the doctor at a young age, but the likely long term improvement in the strength of the immune system means that we cannot consider an increased incidence of colds and flus among children at public facilities as an inherently negative thing. The long terms benefits may actually be substantial.

A common argument in favour of universal access to child care is that the benefits related to increased employment exceed the costs of the child care itself. The results of the empirical analysis clearly showed that more women are in the workforce as a result of greater participation in child care since the policy of universal access was implemented. We could look at this in terms of dollars and cents, or as an issue of equality and justice.

The authors take on the dollars and cents issue in an overly simplistic manner. The argument is summarized as follows - the increased tax revenues resulting from an increased labour supply do not match the costs of increased use of universal facilities. It's a common frustration for me that economists who are clearly against state provision of just about anything other than armies, police, education and property titles, continuously insist on carrying out an analysis that understates the benefits by ignoring any potential spinoff benefits. They do this by limiting their definition of spinoff benefits to only include the initial increase in taxes and corporate or individual profits, as though the effect dies off there. This may be supported by analyzing a useful piece of information called the marginal cost (which means only talking about the cost of providing the final unit ... a useful piece of information, but stems from a mysterious and misguided faith that economic efficiency is the raison d'être for homo economicus.) I might post a rather dense paper that illustrates that Darwinism, as applied to to the marketplace, played an unfortunate role in bankrupting the field of economics of its moral roots.

As was pointed out by Lucas, investment into human capital has strong potential for increasing returns. For those who aren't economists,that means that the efficiencies, or "positive externalities," associated with increased training, experience and education tend to increase as we get more of it. A women who has greater freedom to pursue education or maintain a successful career trajectory will have greater potential to reap the rewards of their investments in human capital. This adds efficiency to our economic system, may increases tax receipts, and offers a better chance for a society to realize the increasing returns that result (as happened in the Silicon Valley) when more people with similar expertise get together and grow their expertise and skills.

This leads to the simple point that women whose decision to use child care is based less on their financial position is a great bonus in terms of their ability to judge their interests. Incidentally, consumer choice theory, at the base of laissez-faire theory, states that the individual can best judge their pathway to happiness. Universal child care improves one's freedom of choice in this regard. The obvious question then, is, is public provision or private provision (funded by income transfers) the better way to go. My answer is simple. I have no idea. The most desirable solutions will be based on public dialogue, which is ultimately based on value judgments more so than 'economic efficiency.' Both analyses and both extremes have valid points. However, if the end result is increased provision of child care, then the results provided by this research in the Journal of Political Economy are equally relevant (and the questionable results are equally questionable).

To sum up, the conclusions of the paper need to be questioned with further research that a) addresses the inherent bias in the methods used in the research of interest, b) investigates the long term impacts of increased exposure to pathogens at a young age when quality public health care is available, and c) carries out a cost-benefit analysis that considers a broader definition of spinoff benefits rather than unrealistically focusing on hidden costs, while considering the inherent value of increasing the opportunity for women to pursue their personal development under a system with greater public provisions for child care.

That, all said with explicit acknowledgment that there is plenty of research to show that parent-child bonding (particularly as an infant) is important in early personal development. However, this study overemphasizes the hidden costs of universal child care while ignoring the hidden benefits.

The choice of the Journal of Political Economy, a respected academic journal, to publish this piece without acknowledging these flaws, does not bode well for those who investigate the possibility that public investments in child care can bring both economic and social benefits. I'm not saying that I know it's the underlying truth of the matter, but the recent study by Baker, Gruber & Milligan will offer ammunition to those who oppose any public forays into the field (and does an injustice to the principle of balanced, unbiased research).

On the positive side, one can be sure that more research will be forthcoming as a result of the contentious findings of their paper.