sobremexico-ibero.knit

class: center, middle
# Back to work or back to the maternity ward? the effects of extending the school-day on fertility.
### Francisco Cabrera-Hernandez `$^a$`  Fernanda Marquez-Padilla `$^b$` 
#### `$^a$` CIDE and HSE `$^b$` COLMEX 
May, 2024
###ICEEF, Brunel University, London UK.

---

# Motivation
- Increase in LFP; **large decrease in fertility** in middle-income countries **vs. "New Fertility Era"**

- Determined simultaneously with childcare as a mediator.

- **Childcare reduces costs of rising children**: increases fertility

-  Childcare **increase opportunity cost of staying at home**

- **Empirical question** depends of baseline FLFP and fertility.

---

# This Paper

- **We exploit the full-time schools' (FTS) availability to study fertility rates** (implicit childcare subsidy).

- Causal effects of a large school day extension by three-and-a-half hours  in pre and primary school (staggered roll out).

- We use the **universe of births and program roll out for a panel of municipalities (2008-2018)**

- 22 million births across 27,000 municipality-year.

- We provide evidence at the intensive and extensive margin (higher parity births) **not common in the literature**.

---
# What do we know?
- Recent evidence that increase access to childcare may increase fertility in more developed countries (Doepke et. al. 2022)

- **Previous empirical evidence offering mixed results** on the effects of childcare on fertility (see e.g. Bick, 2016)

- Effects may be different in the context of a less developed country.

- **Short school days had a direct negative effect on women’s (mothers and grandmothers) LFP** (Cabrera-Hernández and Padilla-Romo, 2020)

- Scarce evidence from LMIC countries. But **FTS reduces teenage pregnancy in Chile** (Berthelon and Krueger, 2021).

---
#What do we learn?
- FTS availability **reduces fertility by 4.4% on average over 8 years**.

- Consistent with **returning to work after having a child** and delaying/stopping fertility.

- The effect is driven by reductions at the intensive margin (mothers who already have one child).

- **Strong effect for teenage pregnancy** (direct effect of staying in school).

- Effects stronger for less educated women in municipalities with stronger labor markets.

- Robust to selection into treated municipalities (migration).

---

# The FTS programme

-   It started in 2007/8 in 500 schools, it reached **25,000 by 2018 or 60% of eligible.**

-   80% of municipalities are treated.

-   For from preschool to 6th grade, it extends school-day **from 4.5 hours to 8.**

-   Aiming to improve learning and **support working mothers**

-   Increased inputs, money, staff, teachers, and infrastructure (e.g. kitchens).

---
# Adoption
<img src="data:image/png;base64,#Imagen3.png" width="75%" style="display: block; margin: auto;" />
---

# Data

- FTS: **Administrative data (2007/8—2017/8)**. Identifying treated schools per academic-year

- Birth data: **universe of birth registries in Mexico**, SINAC (2008–2018)

- Unique birth registry issued by all health institutions at the time of birth **(required later on for birth certificate)**

- We link FTS adoption to data from births’ at the municipality level.

- CONAPO: Municipal population (to construct birth rates) and Marginality index (poverty measure)

- Employment data: ENOE (survey data) for more **dynamic labour markets**

---

# Identification

-   We allow our model to capture dynamic effects:

`$$Y_{mt}=\nu_m+\theta_{t}+\sum_{k=-6,k\neq -1}\delta_kFTS_{m,t-k}+\beta X{mt}+\lambda_{mt}+u_{mt}$$`

- `$Y_{mt}$` is log of births (+1) in municipality `$m$` in academic year `$t$`.

-   `$FTS_{m,t-k}$`, indicates municipality degree of exposure, in academic year `$t$` to full-time schooling in `$k+1$` years.

-  `$k\geq 0$` is equal to one `$k+1$` years after opening its first FTS.

-  `$\delta_k$` is the intent-to-treat effect of the FTS program on crime under standard parallel trend assumptions.

- We add `$\delta_{mt}$` municipality trend

---
# Dynamic Effects II

- We compute estimators **robust to dynamic and heterogeneous effects** developed by Chaisemartin and D'Haultfoeuille (C&D 2020).

- Adequate as 58% of weight in Goodman-Bacon (2021) decomposition comes from always treated and timing groups.

- Never treated group is small. E.g. IW estimator (Sun and Abraham, 2021) rely on these.

---

#Average Results
<img src="data:image/png;base64,#Imagen4a.png" width="85%" style="display: block; margin: auto;" />

---

#Dynamic Effects
<img src="data:image/png;base64,#Imagen5.png" width="120%" style="display: block; margin: auto;" />

---

##Driven by the intensive margin
<img src="data:image/png;base64,#Imagen7.png" width="100%" style="display: block; margin: auto;" />

---

##Reducing higher order births
<img src="data:image/png;base64,#Imagen7a.png" width="80%" style="display: block; margin: auto;" />
---

##Driven by teenagers and women 30-34.
<img src="data:image/png;base64,#Imagen6.png" width="70%" style="display: block; margin: auto;" />

---
##Consistent with staying in school (Black et al. 2008)
 
- Reduction of 19% in teenage pregnancy after 8 years.  
<img src="data:image/png;base64,#Imagen8.png" width="100%" style="display: block; margin: auto;" />

---
##Stronger effects on poorer and less educated women

<img src="data:image/png;base64,#Imagen9.png" width="100%" style="display: block; margin: auto;" />
---
##...and in muncipalities with dynamic labour markets

---
##Robustness: Migration

---
##Robustness: Alternative Transformations

---
##Robustness: Other Estimation Methods
- Sun and Abraham (2021) /  Borusyak and Jaravel (2018)
<img src="data:image/png;base64,#Imagen13.png" width="70%" style="display: block; margin: auto;" />

---
#Conclusions and discussion
- Results differ and complement findings in contexts of higher economic development.

- Informal institutions (childcare).

- Less flexible labor markets (i.e. part time work).

- Low Female LFP at baseline.

**Overall lack of childcare-related policies may not be a binding constraint for fertility but for increasing FLFP**