Title: The Intertemporal Effects of Seeing Prices: Evidence from Garbled Offers in Malawi (with Maite Deambrosi and Jiajing Feng)

ABSTRACT: Preventive health care products are often sold at heavily discounted prices in developing countries. Does that hurt future demand for these products, by inducing subjects to anchor on low prices? Previous research on discounts for anti-malaria bed nets (Dupas, Ecma 2014) documents that experience and learning effects brought about by discounts dominate anchoring effects (if any). Bed nets are, however, an experience good sold at low frequency. Many essential preventive health care products – such as chlorine – are credence goods and have to be purchased at much higher frequency. Moreover, if discounts were perceived as a bargain by study participants, higher future demand might not generalize outside of the experiment. In a field experiment in Malawi, we investigate whether discounts generate anchoring and transaction utility effects across different types of preventive health care products. We do so by cross-randomizing actual discounts and implicit deals. Concretely, we decompose transaction prices into a sticker price and a randomly assigned delivery fee – revealed regardless of the purchasing decision. We find that demand decreases with discounts but increases with the size of the implicit deal. Such intertemporal effects of seeing prices are consistent with anchoring and transaction utility effects. Anchoring decays with time, but lasts up to 2 months after seeing prices. Using impulse-response functions, we simulate counterfactual demand under different pricing policies, showing that universal discounts are dominated by targeted (garbled) offers to those most responsive to each instrument.