ABCDEFGHIJKLMNOPQRSTUVWXYZAA
1
CountryPhilippinesBangladeshNigeriaRationaleReferences
2
Demographics
3
% of population living rurally52.59%61.82%48.04%The number of children of interest are the number born each year, times by the fraction living rurally. This is a conservative assumption, since there is evidence of significant organophosphate exposure in urban pregnant women (e.g. Balalian et al. 2021, Sinai cohort summarised in Engel et al. 2016).https://www.statista.com/statistics/760934/bangladesh-share-of-rural-population/#:~:text=In%202020%2C%20approximately%2061.82%20percent,were%20residing%20in%20rural%20areas.
https://tradingeconomics.com/nigeria/rural-population-percent-of-total-population-wb-data.html#:~:text=Rural%20population%20(%25%20of%20total,compiled%20from%20officially%20recognized%20sources.
https://tradingeconomics.com/philippines/rural-population-percent-of-total-population-wb-data.html#:~:text=Rural%20population%20(%25%20of%20total,compiled%20from%20officially%20recognized%20sources.

4
Number of children born each year (2020)2,182,5002,889,9607,639,490https://knoema.com/atlas/Nigeria/topics/Demographics/Fertility/Number-of-births#:~:text=Number%20of%20births%20of%20Nigeria,by%20our%20digital%20data%20assistant.
https://knoema.com/atlas/Bangladesh/topics/Demographics/Fertility/Number-of-births
https://knoema.com/atlas/Philippines/topics/Demographics/Fertility/Number-of-births#:~:text=Number%20of%20births%20of%20Philippines,slipped%20by%200.39%20%25%20in%202020.
5
6
Organophosphate burdenExplanation in Forum post.
7
Median DAP conc. for 50-75th % (nmol/L)909090
8
Median DAP conc. for 75-90th % (nmol/L)167167167
9
Median DAP conc. for 90-100th % (nmol/L)365365365
10
Proportion of population above 60 nmol/L70%70%70%
11
12
Organophosphate effect
13
Uncertainty of effect75%75%75%I use Bellinger et al.’s (2012) estimate that a ten-fold increase in DAPs causes a loss of 4.25 FSIQ points. I adjust the effect size by 75% (in line with Givewell practice). I make a further adjustment of 75% for the uncertainty of the effect existing, given it is less convincing than the evidence for lead’s DNT. I use Givewell’s figure of 0.22 ‘units of value’ per IQ point gained.
14
Effect size adjustment 75%75%75%
15
IQ points increase per 10-fold decrease in maternal urinary DAP4.254.254.25
16
Units of value from increasing IQ by 1 point0.220.220.22
17
18
Probability of regulatory outcomes
19
No effect50%40%75.0%I make probability estimates of different regulatory outcomes for each case. These are particularly arbitrary, but are slightly more pessimistic than Givewell’s estimates of CPSP’s success. I assume an upper limit of reduction at 20%, although bans in other countries have resulted in decreases of specific metabolites up to 50%. I used relative rather than absolute reductions in exposure because longitudinal U.S. data suggest declines due to regulation of organophosphates are proportional at different exposure levels (Gillezeau et al. 2019).
20
5% reduction in exposure30%35%15.0%
21
10% reduction in exposure 15%20%8.0%
22
20% reduction in exposure5%5%2.0%
23
Sum reduction in exposure0.040.04750.0195
24
25
Effect of regulation
26
Number of years regulation brought forward999I use the same number of years that regulation is brought forward as Givewell does for CPSP.
27
Expected IQ points gained per child0.0743750.08832031250.0362578125
28
Sum of expected IQ points gained each year85366157791133067
29
Sum of expected IQ points gained due to regulation76829314201161197599
30
Sum of units of value gained due to regulation169024312426263472
31
32
Cost
33
Cost of program ($)450000390000450000Givewell’s estimate of the CPSP budget says adding 3 years of work in Guyana and Tanzania would cost $219,717 each. I make the generous assumption that it costs more than twice as much to perform such work for an expanded list of pesticides that include organophosphates with a high DNT burden. Rounding up, total costs would be $450K per country. I assume Bangladesh is ‘only’ $390,000 total, due to pre-existing activities, networks, and government precedent.
34
35
Other adjustmentsI used the same adjustments for government/monitoring costs as the Givewell CPSP BOTEC. Because the number of pesticides banned for DNT is larger than CPSP’s usual work, the adjustment for agricultural harm is 40% rather than 70%, even though effective alternatives for organophosphates are likely available (Hertz-Picciotto et al. 2018).
36
Government cost of enforcing regulation95%95%95%
37
Additional monitoring costs90%90%90%
38
Risk of agricultural harm40%40%40%
39
Reduced ADHD, ASD102%102%102%Organophosphates do have convincing links to ADHD and other neurodevelopmental disorders, but I keep the adjustment to 102%. Gaylord et al. (2020) estimates that intellectual disability adds an additional 24% of costs to that from IQ losses, but I use only a 110% adjustment.
40
Reduced intellectual disability110%110%110%
41
Total adjustments0.3837240.3837240.383724
42
43
Cost-effectiveness
44
Units of value per $1,000 (before qualitative adjustments)375.6099414801.0912953585.4929105
45
Units of value per $1,000 for GiveDirectly's cash transfer program3.43.43.4
46
Cost-effectiveness in multiples of cash transfers (before qualitative adjustments)110.4735122235.6150869172.2037972
47
Cost-effectiveness in multiples of cash transfers (after qualitative adjustments)42.3913379990.4111635966.07872988
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100