Skip to main content
. 2018 Nov;62:147–164. doi: 10.1016/j.jhealeco.2018.09.010

Table 5.

Habit formation model with buyer random effects for households that participated in promotional sessions. The ‘pre-post promotion’ column includes only households that had purchased SQ-LNS before participating in the promotional sessions that launched phase 3 and uses purchases before and after the promotion to estimate the loyalty card effect on demand.

Phase 3 Only
Phase 2 & 3
All Phase 1 price
New buyers Pre-existing buyers
Low High
Constant 2.45 4.91** −1.95 −1.14 2.16**
(1.83) (2.21) (2.47)× (2.16) (0.96)
Loyalty card 0.91*** 0.78*** 1.06*** 0.60** 0.30
(0.18) (0.15) (0.37) (0.25) (0.30)
[0.015] [0.11] [0.0435] [0.12] [0.41]
Avg sachets (t-1) 0.36*** 0.36*** 0.36*** 0.29*** 0.38***
(0.032) (0.036) (0.069) (0.034) (0.043)
Loyalty card × Avg sachets (t-1) −0.0090 −0.025 −0.0097 0.011 0.044
(0.044) (0.041) (0.089) (0.044) (0.091)
[0.96] [0.91] [0.96] [0.96] [0.86]
Post-promo 0.48**
(0.19)
Loyalty card × Post-promo 1.12***
(0.36)
[0.23]
…× Post-period 2 −0.49** −0.39*** −0.60* −0.13 −1.03**
(0.20) (0.13) (0.35) (0.19) (0.50)
[0.29] [0.53] [0.39] [0.77] [0.34]
…× Post-period 3 −0.41* −0.12 −0.70*** −0.17 −0.66*
(0.22) (0.33) (0.21) (0.30) (0.34)
[0.40] [0.86] [0.23] [0.67] [0.57]
…× Post-period 4 −0.58** −0.28 −0.91** −0.28 −0.96*
(0.26) (0.18) (0.45) (0.22) (0.53)
[0.20] [0.61] [0.14] [0.52] [0.35]
…× Post-period 5 −0.56*** −0.45** −0.68*** −0.28 −0.91**
(0.15) (0.18) (0.22) (0.23) (0.36)
[0.18] [0.43] [0.28] [0.53] [0.39]
…× Post-period 6 −0.38 0.072 −0.85*** −0.016 −0.89**
(0.28) (0.34) (0.28) (0.38) (0.44)
[0.35] [0.90] [0.14] [0.98] [0.37]
…× Post-period 7 −0.83*** −0.77*** −0.91*** −0.38* −1.48***
(0.15) (0.14) (0.28) (0.21) (0.33)
[0.048] [0.17] [0.12] [0.36] [0.13]
Male won loyalty card 0.012 0.036 0.019 −0.14** 0.43
(0.12) (0.12) (0.21) (0.065) (0.26)
[0.98] [0.97] [0.98] [0.75] [0.82]
Village FE YES YES YES YES YES
Observations 13,994 7,683 6,311 9,672 6,349
Number of voucher 481 263 218 336 145

*** p < 0.01, ** p < 0.05, * p < 0.1. Robust standard errors clustered by village in parentheses. Randomization Inference-based p-values reported in []. Unreported controls include week and week squared and dummies for post-promo periods, other promotion prices, couple at promotion, and male at promotion.