Genetically modified eggplants BOTEC (public)

	A	B	C	D	E	F
1		Best guess	Pessimistic	Optimistic	Upper bound effect size	Notes
2	Number of farmers receiving intervention with a hypothetical donation
3	Arbitrary donation size	$100,000	$100,000	$100,000	$100,000	Arbitrary
4	Cost per farmer, inputs aside from seeds	$25	$25	$25	$25	"All farmers received an input package worth approximately 25 USD. The package included fertilizer, netting (to prevent birds from eating the brinjal), and support posts for plants." Ahmed et al. 2020, p. 1189
5	Adjustment for seeds cost	10%	20%	5%	10%	Subjective guess
6	Adjustment for training and distribution cost	35%	100%	15%	35%	Subjective guess. "The intervention bundled together the following components: training; extension; an input package; and seedlings." Ahmed et al. 2020, p. 1189.
7	Cost per farmer	$36.25	$55.00	$30.00	$36.25	Calc
8	Total number of farmers covered in one year with hypothetical donation ($100,000)	2,759	1,818	3,333	2,759	Calc
9
10	Income effects
11	Percentage increase in profits per farmer	57.48%	28.95%	57.48%	128.00%	- "Bt brinjal farmers marketed more output, sold at a higher price, incurred lower input costs, and, consequently, had higher net revenues (by 128%)." Ahmed et al. 2020, p. 1187. - “This 13.9% increase in profits is due primarily to savings from decreased pesticide cost and increased the productivity of Bt brinjal.” Ahmed, Hoddinnott, Tanger, and Perlman, "Impact study demonstrates Bt brinjal eggplant variety helps farmers in Bangladesh earn more with less pesticide," 2019. - "Altogether we find that Bt eggplant adoption increases profits by 25%–63%." Ahsanuzzaman and Zilberman 2018. We are not sure which measure is most appropriate, nor what is specifically driving the differences between profits and net revenues in Ahmed et al 2019 and Ahmed et al 2020. We have not investigated further at this stage because under our best guess assumptions, the program would still be below our cost-effectiveness threshold even under optimistic effect size assumptions (see the "Upper bound effect size" column, which assumes the maximum effect size with our best guess assumptions for other parameters).
12	Percentage of agricultural production coming from eggplant production	10%	5%	20%	10%	"Virtually all farmers cultivate rice. At baseline, brinjal occupied only 10% of total cropped area for surveyed farmers (9.5% and 10.7% for treatment and control farmers, respectively)." Ahmed et al. 2020, p. 1191–92. We assume that this is proprotional to their income share.
13	Percentage of household consumption coming from agriculture	65%	65%	65%	65%	"In Bangladesh 35 percent of total income is generated by working in sectors other than agriculture (Figure 4.3)." Rapsomanikis 2015, p. 21.
14	Percentage increase in household consumption	3.74%	0.94%	7.47%	8.32%	Calc
15	% farmers who reuse seeds	20%	10%	30%	20%	Guess. We think that farmers may use seeds for more than one year, but are very unsure for how long. "The BADC is selling BT brinjal seeds through dealers. However, some farmers are preserving its seeds to cultivate it repeatedly." Business Standard, "Cultivation of BT brinjal increasing in Bangladesh: USAID."
16	How many years seeds can be re-used	3	1	5	3	Subjective guess
17	Average effect on ln(household consumption) per farmer treated	0.051	0.009	0.159	0.112	Calc, estimating 1 year of effects for all farmers, and longer-term effects for those farmers who are re-using seeds
18	IV/EV adjustment	0.7	0.5	0.8	0.7	Subjective guess that includes a 10% discount for the possibility that effects would be smaller if training wasn't included, and a 20% discount for the possibility that geographic areas in the study would not be representative of the effect at scale.
19	Excluded effects	1.1	1.1	1.2	1.1	Subjective guess that accounts for a decrease in pesticide poisoning. “Farmers growing Bt brinjal and who had pre-existing chronic conditions consistent with pesticide poisoning were 11.5% points less likely to report a symptom of pesticide poisoning and were less likely to incur cash medical expenses to treat these symptoms.” Ahmed et al. 2020, abstract.
20	Multiplier for resource sharing within households	6.7	6.7	6.7	6.7	From GiveWell's cost-effectiveness analysis for PxD
21
22	Adjusted long-term benefits per person receiving intervention in terms of ln(consumption) per farmer treated	0.26	0.03	0.98	0.58	Calc
23	Value assigned to increasing ln(consumption) by one unit for one person for one year	1.44	1.44	1.44	1.44	2022 GiveWell Cost-Effectiveness analysis — version 5 (public), "Moral weights and discount rate" tab
24	Units of value from intervention per farmer treated	0.38	0.05	1.41	0.83	Calc
25
26	Results
27
28	Total units of value generated with hypothetical donation ($100,000) (before accounting for leverage and funging)	1,052	90	4,690	2,293	Calc
29	Cost-effectiveness in multiples of cash	3.1	0.3	14.0	6.8	Calc using 335 as the number of units of value generated by unconditional cash transfers. See GiveWell, 2022 Cost-effectiveness analysis – version 5, "GiveDirectly" tab.
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
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