that the municipality made the EC pill available 2 years prior. Thus, if we wish to approximately estimate the impact in terms of total avoided cases, this would be −3.92/100, 000 × (13, 016.5 × 346) = −176.5 cases of haemorrhage, or −1.59/100, 000 × (13, 016.5 × 346) = −71.6 cases of abortion related morbidity. If we consider such a calculation for each of the lag terms indicated in Figure 3, this suggests that in total over all years, based on these estimates, the passage of EC pill reform has accounted for 2806 fewer cases of haemorrhage, and 889 fewer cases of abortion related morbidity (full calculations are provided in Appendix Table A4). When these total cases are expressed in terms of the total numbers of pills disbursed following the implementation of the EC reform (112,962), this suggests that for every 40 pills disbursed a single case of haemorrhage was avoided, and for every 127 pills disbursed, a single case of abortion related morbidity was avoided.16,17 15This value is calculated as 14, 120.3/60.3 = 234.2. 16We note that the values estimated for haemorrhage range between 1 fewer case per 234 pills in the case of DD models to 1 fewer case per 40 pills in the event study model. It is important to note that both of these estimates are based on coefficients only. These estimates are broadly comparable when taking into account the reported confidence intervals. 17We can undertake similar calculations with event studies presented separately by the intensity of EC pill disbursements in Figure 4. We document similar figures for high intensity EC pill municipalities only in Appendix Table A5. In the case of haemorrhage this suggests an estimated 1,446 fewer cases in these municipalities only, and in the case of abortion related morbidity, 261 fewer cases. In these municipalities, 64,991 EC pills were disbursed, implying that for every 44.9 pills disbursed a single case of haemorrhage was avoided, and for every 249 EC pills disbursed a single 18 In Appendix Figure A4 we document potential combinations of the impact of EC pills on (a) reduced conceptions/pill, and (b) reduced morbidities/conception that could theoretically explain the estimated effect sizes. These rough estimates can be used to provide a back of the envelope calculation of the cost savings implied by the EC reform. We collated the cost of the EC pill as paid by the MoH from public government records.18 The unitary price for a single EC dose varied from as low as 886 Chilean Pesos (CLP) to as high as 5,676 CLP (or 1.50 USD to 8.22 USD based on exchange rates at the time of purchase), with an average price of 3,379 CLP (4.75 USD). While it is difficult to estimate the full costs of an inpatient visit to the health system, estimates presented by the Chilean Budgetary Department suggest that a single night of hospitalization on average costs 43,842 CLP (Instituto de Administración de Salud, 2016). The administrative data used in this paper shows that the average hospitalization for haemorrhage resulted in 2.3 nights, and the average stay for abortion related morbidity resulted in 3.4 nights. Bringing this all together, we can determine a back of the envelope estimate of total costs for pills disbursed as 112, 962 pills × 3379 CLP/pill = 381.83 million CLP, and total savings in hospitalizations as 2806 cases of haemorrhage × 2.3 nights per case × 43, 842 per night plus 889 cases of abortion related morbidity × 3.4 nights per case × 43, 842 per night = 415.46 million CLP. These rough estimates, which only include the direct savings in terms of hospitalization suggest that the government may have saved 33.6 million pesos due to the reform.19 case of abortion-related morbidity was avoided. 18These records were compiled from CENABAST, (the National Centre for Medical Provision), and are available for 9 different bulk purchases of the EC pill from 5 different suppliers from between 2015–2020. We searched for all purchases using the generic name Levonorgestrel, which covers 3 different brands supplied (Cerciora, Escapel-2, and Pregnon). 19This equates to around 50,000 USD based on prevailing exchange rates in 2017. This is a clear lower bound, given that the EC pill has generated other cost savings for the health system, such as fewer births. 19 4.6 Placebo Outcomes, Alternative Explanations and Additional Tests Despite the event-study evidence and lack of pre-trends, we still may be concerned that these results are capturing systematic differences in prevailing health outcomes within municipalities. If for example, at the same time the EC pill was adopted municipalities engaged in general healthpromoting policies, our estimates may capture this, rather than a true EC pill effect. We examine this in a number of ways. First, we document that results are virtually unchanged when including controls for a mayor’s party, gender, and vote share upon election. These results are provided as online Appendix Figures A5, A6, A7 and A8 (corresponding, respectively, to Figures 3a, 3b, 4a and 4b). Second, to provide a test of the idea that these results may be capturing general improvements in health rather than anything related to the EC pill, we conduct a number of placebo tests. These tests consist of estimating identical specifications following equation 1, however now using health outcomes which do not plausibly depend on EC pill availability. The first of these is a pure placebo test where we