Matthew Harding is an Econometrician who develops cutting edge statistical methods for the analysis of Big Data to answer crucial economic questions related to individual consumption and choices in areas such as health and energy. As a Data Scientist he focuses on the analysis of “Deep Data”, large and information-rich data sets derived from many seemingly unrelated sources but linked across individuals to provide novel behavioral insights. He is particularly interested in the role of technology and automation to induce behavior change and help individuals live happier and healthier lives. At the same time his research emphasizes solutions for achieving triple-win strategies. These are solutions that not only benefit individual consumers, but are profitable for firms, and have a large positive impact on society at large.
As part of a visit at the Sant'Anna School, he will hold a seminar titled "A Structural Approach to Dynamic Energy Pricing and Consumer Welfare" on May 23, 2019 (2:00 to 3:30pm, Aula 3, Sede Centrale).
Abstract: With the proliferation of smart meters and smart homes, there is growing interest from utilities in policies that exert control over demand in the electricity market. Time of use pricing is an approach to demand side management that charges a different marginal price for electricity during the peak usage hours of the day. We analyze a panel of high frequency observations taken from a randomized control trial in the summer of 2011. Experimental treatments include two price structures over the peak of the day, as well as four technologies for information provision about the prevailing marginal cost of electricity. We estimate a market demand curve for each hour of the day using a structural demand system. Relative to an untreated control group, treated households face periods of both elevated and reduced prices. This pricing structure potentially contributes to consumer welfare by providing the opportunity to shift consumption from high to low price periods. We investigate the degree to which consumer demand response varies relative to the household’s location within the distribution of hourly demand. This allows for a naive estimation of average consumer welfare. Additionally, we compute the change in total welfare across households when heterogeneity is introduced across consumer demographics.
He will also hold a Q&A Session for all interested students and researchers titled "Big Data, Machine Learning, and Deep Learning: Opportunities and Costs for Economists", researchers and faculty interesed in his research on May 24 (2.00 to 5.00 pm, Aula 6, Sede Centrale). Please join us for these events!