Montpetit, Beauregard, and Carrer (2026) quantify the welfare impact of universal childcare provision by calculating the MVPF of Québec’s 1997 reform, often deemed the most ambitious childcare reform in North America. The reform introduced reduced-fare daycare at $5/day/child and increased the number of spaces per preschooler from 0.15 in 1997 to approximately 0.45 by 2003. The reform significantly boosted mothers’ labor supply and childcare use, though it had no notable economic effects on children’s long-term outcomes or on fathers. The analysis leverages newly digitized archival data on daycare coverage rates by administrative region within Québec matched with longitudinal data on Canadian children and their parents.
The paper estimates two MVPF calculations. The first uses a sufficient-statistics approach with a non-linear difference-in-differences model to estimate mothers’ earnings gains along the income distribution. The second uses a structural model to account for the non-marginal change and to quantify non-pecuniary gains.
MVPF = 1.3
The net cost of the reform to the government is the amount of new subsidies allocated to the daycare market net of any additional income taxes levied from working mothers. The paper focuses on additional income taxes remitted by mothers in two-parent families because of (i) another simultaneous program affecting single parents and (ii) evidence of null impacts on children later in life and on fathers.
1. Total subsidies: The paper obtained the total subsidy amount from the Québec Treasury Board. As some subsidies existed before the reform, the paper deducted the estimated amount that would have been spent in the absence of the reform. To calculate these counterfactual subsidies, the paper applied a conservative growth rate of 3.68%, reflecting the pre-reform growth rate. The total upfront expenditure for the period analyzed (fiscal years 1997-1998 to 2002-2003) amounts to $2.741 billion.
2. Additional tax revenue and reduced transfers and benefits: Using the estimated labor income gains and the Canadian Tax and Credit Simulator (Milligan 2019), the paper calculates the net fiscal impact for each mother. This fiscal externality is estimated at $866 million using the reduced-form method and $760 million using the structural approach.
Thus, the net cost is $1.875 billion in the reduced-form and $1.981 billion under the structural approach.
The paper uses two approaches to estimate mothers’ WTP for the reform.
The sufficient-statistics approach uses a reduced-form estimate of the reform’s impact on mothers’ earnings. The paper finds average annual gross earnings gains of $3,003. Multiplying this estimate by the number of mothers in two-parent families with children in the relevant age ranges over the period considered yields $3.221 billion and thus $3.221 − $0.866 = $2.355 billion in net earnings gains.
For the structural estimator, the paper calculates the equivalent variation from the simulated reform. The authors show that, in their model, a mother’s equivalent variation is equal to the ratio between the difference in indirect utilities under the two policy states, and the inverse of her income net of childcare expenses in the status quo. Summing the WTP across the total number of mothers of preschoolers in Québec yields a total WTP of $5.599 billion.
The MVPF of the Québec childcare reform under the benchmark estimator is then the ratio of net earnings gains to the net cost, that is $2,355M/$1,875M = 1.256.
Under the structural approach, which includes non-pecuniary gains, the MVPF is the ratio of the simulated WTP to the net cost $5.599M/$1,981M = 2.826.
Confidence intervals are obtained by bootstrap. For the benchmark estimator, the paper uses the simulation method of Hendren and Sprung-Keyser (2020). The 90% confidence intervals under this estimator are [0.55, ∞]. For the structural estimator, the authors bootstrap over the full model, yielding 90% confidence bands of [2.02, 3.65].
The paper also performs counterfactual simulations using the structural model and computes counterfactual MVPFs with variation in the level of supply expansion and the price charged to families. The paper finds that the government could have achieved larger welfare gains by channelling more resources towards opening spots rather than to lowering prices. For instance, a reform that increased the daycare coverage rate by an additional 0.05, but doubled the fee (to $10/day/child) would yield an MVPF of 3.193.
Last, the paper also notes that, when accounting for the societal costs associated with the long-run increase in youth criminal activity reported by Baker, Gruber, and Milligan (2019), the MVPF slightly decreases to 1.233 under the benchmark approach.
Baker, Michael, Jonathan Gruber, and Kevin Milligan (2019). “The Long-Run Impacts of a Universal Child Care Program.” American Economic Journal: Economic Policy, 11(3):1–26. https://www.aeaweb.org/articles?id=10.1257/pol.20170603
Hendren, Nathaniel and Ben Sprung-Keyser (2020). “A Unified Welfare Analysis of Government Policies.” The Quarterly Journal of Economics, 135(3): 1209–1318. DOI: https://doi.org/10.1093/qje/qjaa006
Milligan, Kevin (2019). Canadian Tax and Credit Simulator. Database, Software and Documentation, Version 2019-1. https://sites.google.com/view/kevin-milligan/home/ctacs
Montpetit, Sébastien, Pierre-Loup Beauregard, and Luisa Carrer (2026). “A Welfare Analysis of Universal Childcare: Lessons From a Canadian Reform.” Centre for Household, Income, Labor, and Demographic Economics Child Working Paper Series No. 128. Current draft: https://sebastienmontpetit.github.io/WebsiteSM/MCB_QCchildcare.pdf