Six admission design challenges for expanding free childcare in New York City - and what to do about it!

Oct 17

Admission design

If childcare becomes free through a system of universal vouchers, we need to rethink how admissions work. Talking about the design of the admission system forces us see the childcare system as a whole. It also helps us value simplicity over complexity. Otherwise accountability, transparency, flexibility, and agency will suffer. Politicians might learn a thing or two about how to put vision into practice, and we might be better equipped to ask them more pointed questions. The worst idea is to leave childcare admission design to the experts — there is no telling what will happen then. But admission design is also a technical subject that requires system thinking.

A case for optimism

Personally, I’m an optimist about the power of competition and entrepreneurship in childcare markets with a healthy mix of private and public providers — especially in the United States where charter schools and alternative methods of learning and education for children are increasingly the norm. It’s easy to imagine a policy that makes childcare free for all ages: a city can simply provide every family with a voucher for each infant and toddler, redeemable whenever they need care. With strong regulation and oversight, combined with America’s remarkably adaptive colleges, labor market, and entrepreneurial spirit, the system would evolve to meet demand. In short, ‘‘just do it’’.

A case for pessimism

However, I can’t help but feel pessimistic about how decentralized institutions manage childcare admissions — and whether they can effectively decide, on their own, who should be admitted and when. Childcare systems are inherently messy. Centers are small, facilities vary in quality, staff differ in experience, and the needs of children and families can differ dramatically. Even well-intentioned efforts by individual providers to “triage” admissions, as if in an emergency room, might make things worse. We’ve seen how chaotic similar systems can be in higher education: Why must applicants apply to a hundred programs, and how do we decide who most deserves a spot? Now imagine a system in which government funds flow through countless small actors, each making politically charged triage decisions about who receives childcare and when. It’s enough to bring to mind Ronald Reagan’s famous line about the most terrifying words in any language: “I’m from the government, and I’m here to help.” Yikes.

Speaking from Denmark

My mix of optimism and pessimism might make me sound like a typical Republican voter, even though I live in Bernie Sanders’ favorite model country — Denmark — and firmly believe that universal free childcare is a basic right every family should have. As a Copenhagen resident, I’ve learned to reconcile these conflicting views by studying and witnessing firsthand how the city promotes family agency and genuine childcare choice — an approach that U.S. advocates could champion more effectively.

Six challenges for childcare admission system design

Childcare sits at the intersection of economics, human capital, emotion, and trust. Admissions systems are the primary gateway through which access to care for infants and toddlers is negotiated. Ask anyone who runs a childcare center—or any parent who has had to find one—and they’ll confirm how complex this process can be. These challenges don’t disappear simply because parents pay with a childcare voucher instead of their own cash. Yet free childcare introduces new complexities of its own. Below are six key issues to consider when designing admissions systems for universal, free childcare for infants and toddlers.

Challenge 1. The ‘‘continuity of care’’ problem

The best childcare systems provide children with continuity. Most parents would prefer their child to remain with the same caregiver for several years, especially when care begins at 12 or 14 months of age. Yet many childcare systems fail to ensure continuous care. Disruptions occur when children are moved between facilities, or when staff and children are reassigned within the same center.

An effective admissions system should minimize these disruptions by emphasizing stability from the start. That means giving careful attention to the child’s initial entry date and ensuring that the first placement is the right one for the long term — so infants and toddlers can grow alongside the same caregivers, classmates, and families.

Early Childhood Education and Care (ECEC) specialists have recently started to recognize the fundamental importance of continuity in childcare systems as public funding for childcare increases for children of younger and younger ages. For example, the European Union looks to the integrated childcare centers in Denmark, which host children from ages 6 months to 6 years, as a solution to the clashing program offerings in most other European countries. For example, see what Italy is trying to do to solve this problem:

Childcare entrepreneurs in the United States also recognize the importance of continuity in their business models. Parents own understanding of the importance of continuity explains why so many private childcare facilities in the United States spend a lot of effort to keep children in childcare once they are admitted. This includes giving discounted or flat rates for infants even though the cost of providing infant care is more expensive (lower teacher to child ratios) than toddler care . Continuity also supports trust and childcare reputations. Having parents vouch for a childcare by continuing to use it is an important and effective performance metric for both public and private childcares.

Challenge 2. The ‘‘public and private delivery’’ mismatch problem

The “public–private delivery” mismatch arises from the fact that infants and toddlers are typically cared for in privately operated childcare facilities, while older children receive care and education through public providers.

The “public–private delivery” mismatch amplifies the continuity of care problem facing many families in the United States. Most middle-class families in the United States, for example, rely on private childcare providers for infants and toddlers rather than public programs. This is a pattern that is observed in many childcare systems around the world. In the United States, publicly funded options — like Early Head Start or state-subsidized centers — serve only a small fraction of children under age three, and are typically targeted to low-income families. As a result, the vast majority of infants and toddlers, especially in middle-income households, are cared for in private daycare centers, family childcare homes, or by nannies and relatives.

Public childcare in the United States kicks in as children grow older and become eligible for publicly funded pre-kindergarten (Pre-K)programs. At that stage, many families—across income levels—shift from private childcare to public or mixed-delivery Pre-K options, which are often free or heavily subsidized. The result is a sharp increase in public enrollment beginning around age three or four, marking a major shift from a largely private childcare market for infants and toddlers to a more publicly supported system for preschool-age children.

The “public–private delivery” mismatch and the ‘‘continuity of care’’ problem create an obvious question for childcare systems. If we want to solve the ‘‘continuity of care’’ problem, do we expand private or public care. Or, do we find better ways for private and public childcare facilities to cooperate. A discussion of these opportunities must account for admissions and continued admissions across both public and private childcare facilities. Addressing these challenges requires a conversation about the possible conflicts between private and public childcare, but also about the potential opportunities for partnerships between private and public systems.

Challenge 3. The ‘‘admission date mismatch’’ problem

Nobody looking for childcare for an infant in the United States imagines a system where childcare starts on the “back-to-school” day in September. If we did, parents of a child born in January would be stuck deciding between starting at nine months or waiting until twenty-one months — both might be bad choices.

The “admission date mismatch” problem comes up when we expand free childcare for younger and younger ages. Do we let parents choose when their child starts — at a time that works for their family— or do we set fixed admission dates — like the first week in September at most schools — that everyone must follow? If your child were born in January, when would you want them to start childcare? How much freedom would you have to decide?

Childcare experts often warn parents not to start too early, saying it can be tough on very young children. Others argue it depends on the child. The best childcare systems understand this and give parents real choice—to start when it feels right for their family. Childcare for infants and toddlers works best when parents decide when to start—based on what’s right for their family and their child’s development. Some parents may choose to wait, while others, with work deadlines or few alternatives, need to start sooner. Things change as children get older. By the time they reach pre-K age, it makes sense to align childcare with school schedules, and a September start can work well for everyone.

The ‘‘admission date mismatch’’ problem is what do we do about our choices of ‘‘rolling’’ versus ‘‘batched’’ admission systems for childcare. New Yorkers and their schools have learned to accept new methods of running ‘‘batched’’ admissions.

Applying to N.Y.C. Public Schools Can Feel Daunting. Here’s What to Know.

But how might we modify these solutions for infants and toddlers when we choose ‘‘rolling’’ admission systems? This is a whole different kettle of fish.

Challenge 4. The ‘‘prioritization of applicants’’ mismatch problem

Most education and childcare systems determine who gets what in one of two very different ways. Childcare centers typically rely on waitlists, while schools use catchment areas. But if childcare becomes available to everyone from an earlier age—and the system seeks to maintain continuity of care across those ages—how should priorities be set? Should waitlists or catchment areas take precedence? Some may even advocate for a more complex triage system, one that ranks families by need or other criteria, similar to the model used in Head Start programs.

Any systems that makes scarce childcare services free needs to have a way of communicating who gets what and when. A very effective way for childcare institutions to manage needs is to give older children priority. When faced with two applicants a parent with a 12 month old child who has waited 12 months and a 3 month old child that has waited 3 months, most childcare institutions will admit the older child when a spot comes up.

The methods of prioritizing applicants if all children start at the same age gives us a lot of other options to consider. For example, if childcare is provided by schools, then a catchment area might be useful since traffic guards and school buses can be coordinated. However, for children before the age of six, none of these arguments really hold, because parents take children to childcare facilities themselves.

The admission system that you choose has vastly different effects on the range of choices available to parents. If you admit using age priority, it is natural to make all childcares available to all parents. This means that a parent in New York might have 1000s of choices if childcare is made free.

However, if the system evolves in the Pre_K direction, with admissions on a common date, a natural method of prioritization might be to give priority by distance. This would help daycares coordinate so that families are bunched around the same institution, but this would effectively eliminate choice.

Other sources of prioritization can start to sound like lotteries. For example, we might order priority by an invisible triage system. This means that parents won’t know exactly who is admitted to each childcare. However, this system has a trade-off. Parents have to apply to a lot of daycares. Some systems might ask parents to rank 10 daycares. But that is a lot of daycares to vet for each parent. And the amount of contact needed to do that vetting is daunting.

The point is the childcare is not a triage like an emergency room - even though this is how many head-start child cares see admissions. Parents should be able to effectively choose their childcare and have a good sense of what their chance is of getting it. Absent the actual of lotteries, most AI systems will effectively return a number for each parent. You can get into these daycares and not these daycares. A good system should be read to defend this choice set to every parent.

Challenge 5. The ‘‘childcare vetting’’ problem

Admission systems can also support parents in the process of “childcare vetting.” For example, if a family’s only option is the nearest childcare center, parents are motivated to visit, assess its quality, and decide when to begin care. If the public option proves unsatisfactory, they can choose to wait.

In systems with broader choice, parents may be required to submit long lists of acceptable childcare centers—greatly expanding the scope of vetting and compelling each family to investigate multiple providers. To impose some order, certain systems introduce small costs that discourage parents from joining too many waitlists; private centers, for instance, often charge fees to be placed on a list.

Some advocates of free childcare argue that robust regulation could make parental choice unnecessary. Yet there are clear limits to what regulation alone can achieve. Childcare is a deeply personal decision, and parents often value systems that preserve genuine choice while motivating providers to improve through healthy competition.

At present childcare systems in the United States provide a lot of different childcare options. The current culture suggest that parents need to be active participants in vetting of child cares. While childcare is highly regulated, it remains a market place where parents have a lot of choices. The federal agency, childcare.gov provides detailed recommendations to parents on what to look for when selecting a center, underscoring the personal weight of these decisions and the uneven quality of care available.

US.gov advice on how to vet a childcare institution

Childcare admission systems are part of the important end game for childcare regulations. Are we going to move to a system where parents chooses their preferred childcare from a large set of choices. A system where each parent is responsible for vetting the child care that they use. Or a system where the state plays a role by administering which child care each parent is expected to get. We can also ask if there is a middle ground?

Challenge 6. The ‘‘elimination of waitlists’’ problem

Here is a conversation I saw on Reddit.

How do I explain that we aren’t first come first serve but we also aren’t just income qualified? I don’t think it’s appropriate to make a family hear me drone on about the complex process but when there is a waitlist, I want to explain it accurately. There isn’t a way for me to say ``you’re 17th on the list'' because anybody can apply today and get put on spot 1. Do you just say ``families are ranked by need''? How do you explain it so you are accurate? Head Start Administrator.
Ooh, I’m an intake specialist for a head start program. I got you. I usually use some variation of ``families are assessed according to need, as well as availability in the center. We will review your application and give you a call back if approved.'' I’ve never had anyone push for more detail than that. Reply

The question above highlights that Head Start administrators are unable to provide parents with the information they truly need — what childcare options are available to me, and when? Parents need this information not only to evaluate childcare programs but also to plan ahead. In essence, the parent’s question is: I want to prepare now so I can transition smoothly back to work in six months (or whatever timeframe is realistic). I want to make decisions based on clear, reliable information so I can choose childcare today from a concrete set of options.

Americans currently have a hard time seeing the possibility of certainty in childcare assignment. A childcare provider attempting to guarantee a spot is like an individual farmer promising consumers regular vegetable deliveries for the entire year, despite the uncertainties of weather, finding workers to pick the crops, and the unpredictable retention of existing customers. A more sustainable solution is a cooperative that coordinates and manages these deliveries across multiple farmers. The difficulty in childcare, however, is that the service is deeply personal. Parents are unlikely to be satisfied with just any option; they want a provider who engenders trust that aligns with their child’s specific needs.

A big problem with waitlist admissions is that parents sign up to too many waitlists. There is needless coordination of applications. Of course, a childcare could charge a fee for parents to be on their waitlist. However, this seems unfair. It would also be regressive in that some families might win by paying fees for multiple waitlists.

But if we want to make waitlists disappear, what do we replace them with. Centralized systems give us a lot of ideas. How to introduce guaranteed admissions. How to make childcare institutions predict open spots in advance of placements. How to make offers to parents well in advance of making placements.

A system where all childcare costs the same - free - creates a lot of opportunities for coordination, prediction, choice, and placement guarentees.

What to do about it? How Copenhagen addresses the challenges

Copenhagen uses a coordinated waitlist system that allows parents to start free childcare from ages 6 months and onwards. Here are some things to reflect on about the Copenhagen childcare admission system:

Applicants are expected by the time the child is 4 months of age. Applying earlier gives no advantage.
Copenhagen offers generous parental leaves. So most parents start childcare from ages 6 months to 18 months. Some might start even later. It helps childcare systems if parents don’t need free childcare before age 6 months. But even in Copenhagen there are a lot of different views about when is the best age to start childcare.
There is no batching of admissions for infants and toddlers. Admissions for childcare before the age of three is rolling. Older children have priority.
Guaranteed placement times are available. But these admissions do not let parents jump the queue. I have posted about the guaranteed placement time option here and here.
There is a sibling priority so that parents are never stuck with different childcare for their children. This means that there are two ordered waitlists at each childcare. The upshot is that parents with a second child can effectively choose when to start. For other parents, more popular childcares might require a later starting date. In practice, the data suggests that this can be several months or more.
Parents have a lot of choice. They can apply to any childcare in the city. All children are treated equally.
Many childcares are integrated with Pre-K. Some Pre-Ks are famously not integrated. For example, Copenhagen famous forest Kindergartens cannot offer care to infants and toddlers.
Parents can apply to a maximum of two childcare institutions, but have a reasonable expectation of when to start at each.
All childcare institutions offer visits during regular hours. It helps that parents cannot apply to a lot of waitlists.
Copenhagen gives private and public childcare similar subsidized support. Vouchers pay the same if private or public.
Copenhagen city offers inspections and enforces rules on staffing ratios at private and public childcares. Toddler and infant care, for example, requires a maximum 3 to one ratio of children to teachers. Older childcare requires a 6 to one ratio.
Copenhagen does not regulate how each childcares moves children between rooms in the age 6 months to three years categories. See my earlier post on room policies here.
Childcare admission is based on the ‘‘They say when, we say who’’ principle. See my earlier post here.
Copenhagen offers a simple effective central portal to apply to all public and many private childcares. See here. Is this what you imagine for New York?
There is no so-called ‘justified envy’. No one is ever given a spot ahead of you if they do not have priority. A parent would have justified envy if someone jumps the queue. Copenhagen strives to avoid this outcome, even though the system offers guaranteed placement dates to those parents who ask.
Predictability. Childcare facilities have lean waitlists. So can offer reasonable predictions about when spots are coming available.
Public accountability. An expensive popular public childcares earns its keep by being popular. If everyone faces the same waitlist pressures, we also have clear signals of popularity. In practice, closures of childcare facilities is not decided by so-called experts, but by accountability to parents. Sometimes the most costly childcares are kept open because these facilities have better facilities that parents value.
Public accountability. The guaranteed placement time provision forces the system to keep adequate supply of spots.
Coordination with childcare institutions. The administrators have a lot of input and communication with each provider. This power is used to encourage childcare institutions to predict and account for future openings well in advance of when parents will use them. This means there is a healthy communication of the value of predictions, which encourages investment and utilization of AI and data.

Digging deeper

My list of six challenges is only a starting point. Many key issues in childcare admissions require insights from mathematics and game theory to be fully understood. For example, a central concern in many systems is how to encourage parents to report their true preferences for when and where their child should start, rather than trying to game the system. Ensuring truthful reporting allows the system to make informed decisions, after which we can ask a more fundamental question: what outcomes should society aim to achieve?

Nobel prize winner, Al Roth, maintains a useful website that talks about these sorts of issues. He even mentions work we were doing on childcare admission design as early as 2011, showing how long this topic has been on my mind.

School assignment as viewed by families over time, when sibs are given priority

In joint research with Daniel Monte and Norovsambuu Tumennasan, we use game theory to illustrate some of the challenges in integrating different types of priorities in childcare admissions. These papers examine the so-called prioritization mismatch problem, which arises when admissions are dynamic—parents can start at different dates, admission designers use multiple methods to prioritize applicants, and parents care about continuity of care. Our studies shed light on the complex trade-offs and strategic considerations inherent in designing fair and efficient childcare systems.

The Day Care Assignment: A Dynamic Matching Problem

Strategic Performance of Deferred Acceptance in Dynamic Matching Problems

The impossibility results in these papers help explain why I favor simplicity in childcare admission design. I could go on to discuss how Copenhagen addresses the six challenges I have outlined. My goal is to make the rules of these systems better understood, so that other cities and countries can draw inspiration from them.

I could also discuss how other places address these challenges differently. You will find varied solutions across the Nordic countries and around the world. There is a valuable community of researchers, led by people like Chris Nielson, working to build better childcare admission systems. I am fortunate to have many insightful conversations with Chris, which help me understand the challenges of designing effective systems. We are also collaborating on projects aimed at creating insights to improve implementation of these systems as early childcare becomes available to more families.

John Kennes