Rough model of spillover effects in GD's program

	A	B	C	D	E	F	G	H	I	J	K	L	M	N	O	P	Q	R	S	T	U	V	W	X	Y	Z	AA	AB	AC
1	User inputs	Amar	Andrew	Caitlin	Chris	Dan	Elie	Holden	Isabel	James	Josh	Natalie	Nicole	Olivia	Sophie	Stephan	Anyone	Conventional	1	2	3	4	5	1	2	3	4	5

2	Shared
3	Discount rate	6.1%	6.0%	6.0%	4.0%	4.2%	4.2%	3.0%	5.5%	4.0%	4.0%	4.0%	5.0%	4%	6.0%	4.0%		4.2%	5.0%	1.5%	0.0%			Based on general returns to investment	Low-end of commonly used discount rates	No discount			Former Research Analyst Emma Trefethen wrote an internal document about discount rates that many staff found very helpful. You can view it here.
4
5	Deworming — General
6	Replicability adjustment for deworming	16%	11%	11%	7%	11%	11%	25%	11%	11%	11%	11%	11%	11%	9%	11%		11%	50%	23%	85%			Rough midpoint of next two options	Lower end of Ioannidis 2005's theoretical internal validity (Table 4 in http://www.plosmedicine.org/article/info:doi/10.1371/journal.pmed.0020124): "Underpowered but well-performed Phase I/II RCT"	High end of Ioannidis 2005's theoretical internal validity (Table 4 in http://www.plosmedicine.org/article/info:doi/10.1371/journal.pmed.0020124): "Adequately powered RCT with little bias and 1:1 pre-study odds"
7	Adjustment for years of treatment in Miguel and Kremer vs. years of treatment in charities' programs	90%	90%	90%	90%	90%	90%	90%	90%	90%	90%	90%	90%	90%	90%	90%		90%	100%	80%	24.1%			Assumes that every year of deworming is as beneficial as each year in the study	Arbitrary middle assumption - assumes that the average year of deworming is 80% as beneficial as each year in the study	Assumes that our charities deworm each child for 10 years, but that only 2.41 (number of years in study) are needed to realize benefits (2.41 years from Baird et al 2015, Pg 7) and future years will not result in additional benefits.
8	Adjustment for El Niño	65%	65%	65%	65%	65%	65%	54%	65%	65%	65%	65%	65%	65%	65%	65%		65%	71.15%	54.21%				We estimate this magnitude by taking the ratio between the baseline prevalence of any moderate-heavy infection in 1998 (.37 from p. 168) and 1999 (.52 from pg 173), without any adjustment for potential externalities affecting the 1999 prevalence (See Miguel and Kremer 2004).	We adjust our first option by using an odds ratio, which we believe may be appropriate given that this adjustment is based on prevalence and the appropriate adjustment for intensity may be more extreme.				GiveWell's discussion of the El Niño event during Miguel and Kremer's trial
9	Additional years of treatment assigned to the treatment group from Miguel and Kremer	1.69	1.69	1.69	1.69	1.69	1.69	2.00	1.69	1.69	1.69	1.69	1.69	169%	1.69	1.69		1.69	2.41					Baird et al. 2016 Supplementary Appendix, Table 2, p. S20					Baird et al. 2016 Supplementary Appendix
10
11	Vitamin A supplementation (VAS) — Helen Keller International (HKI)
12	Value of development benefits from a year of VAS coverage relative to a year of deworming in Miguel and Kremer	4%	5%	7%	7%	6%	6.3%	6.8%	6%	7%	7%	7%	7%	7%	7%	7%		7%
13
14	Cash — GiveDirectly
15	Percentage of transfers invested	39%	39%	39%	39%	39%	39%	50%	39%	39%	39%	39%	39%	39%	39%	39%		39%	39%	50%	25%			Best guess based on the increase in asset value relative to the size of the large-arm transfer in GiveDirectly's RCT.	Based on balancing self reported data on spending from Give Directly follow-up surveys (http://www.givewell.org/node/2361#Datafromfollowupsurveys) in which investments in buildings, land, livestock, and businesses typically account for ~70% of spending with the RCT results which have investments as $505.94 (USD PPP) (within villages, or $601.88 across villages) on a transfer of $1,525 USDPPP (i.e. about 1/3 of the total: p 117, http://www.princeton.edu/haushofer/publications/Haushofer_Shapiro_UCT_Online_Appendix_2016.04.25.pdf see p 1, https://www.princeton.edu/~joha/publications/Haushofer_Shapiro_UCT_2016.04.25.pdf for total transfer size)	Moderately conservative			Haushofer & Shapiro 2016
16	Return on investment	10%	10%	10%	8%	10%	10%	10%	10%	10%	10%	10%	10%	10%	10%	10%		10%	10%	15%	0%			Conservative guess, adjusted downward from the roof ROI calculation (see below on this sheet) due to not knowing survey methodology and inconsistent results, and the possibility that iron roofs may be a particularly good investment. See: http://www.givewell.org/node/2360#AvariantofGiveDirectly	Estimated ROI on an iron roof from the RCT of GiveDirectly's program. See Haushofer & Shapiro 2016, p. 33.	Very conservative			Cash Transfers — What return on investment do cash-transfer recipients earn?
17	Duration of investment benefits (in years)	15	15	10	8	15	30	10	15	15	10	20	10	15	10	10		15	5	15	20	40	80	Low end estimate. Investing for many years is hard for poor individuals that regularly experience income and expense shocks.	Slightly lower than GiveDirectly's estimate of the lifespan of a roof (http://blog.givewell.org/2012/12/19/cost-effectiveness-of-nets-vs-deworming-vs-cash-transfers/)	GiveDirectly's estimate of the lifespan of a roof.	Same as deworming	Arbitrary high upper estimate, since Haushofer and Shapiro 2013 reports that "most respondents were unable to put an upper bound on the durability of metal roofs." (http://files.givewell.org/files/DWDA%202009/Interventions/Cash%20Transfers/haushofer_shapiro_uct_2013.11.16.pdf, pg. 34)
18
19
20
21
22
23
24
25
26
27
28
29
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