Promoting alternatives to substitute animal-based proteins is an important strategy to mitigate the environmental, animal welfare, and health impacts of animal agriculture. Given the essential role of consumer preference and marketing success in food promotion, in this dissertation I assess consumer preferences for alternative proteins and market dynamic in plant-based meat alternatives industry. In the first chapter, I conduct a meta-analysis to provide evidence on consumer preferences for plant-based meat alternatives (PBMA) and lab-grown meat not conditional on research context, utilizing machine-learning techniques in both the data collection and the data analysis phases to improve the efficiency of the meta-analysis. I demonstrate that machine-learning reduces the workload in the manual title-abstract screen phase by 69% accounting for 24% of total workload in data collection. Besides, machine learning improves out-of-sample of sample prediction accuracy by 48-78 percentage points when compared to econometric model. Empirically, the findings further reveal that demand for meat alternatives is higher among younger consumers, especially when the products displayed benefit information. Food value theory can explain consumers’ heterogenous demand for alternative proteins. In the second chapter, I utilize consumers’ food values to identify the drivers of demand for alternative meat and milk products in China, one of the world’s largest consumer markets. I find that public food values, such as environmental impacts and animal welfare, drive consumers’ demand for alternative meat and milk. Results show that approximately 35% of urban food shoppers constitute the potential market for these products. I estimate that modest consumption of alternative meat and milk products in these markets can improve food system sustainability by lowering China's animal production greenhouse gas emissions. The PBMA market has garnered substantial investment, with numerous new product developments underway. In the third chapter, I evaluate the effects of a new brand entry using store-level scanner data from IRI. I employ three empirical approaches: the two-way fixed effect approach, which allows to evaluate average effects, and the extended two-way fixed effects approach and the rolling approach with double machine learning which account for dynamic effects. The results suggest that entry effects vary across geographical locations, entry waves, and post-entry times. From methodological perspective, I show that the TWFE estimates could be biased when the staggered entry effects are not homogenous across entry waves and post-entry times. Notably, I also found that, compared to the other models, the rolling approach integrated with DML controls for selection bias by including high-dimensional covariates, leading to an improved model precision ranging from 24% to 45%. In sum, findings from this dissertation can be used to inform policymakers and industry to better understand the consumer demand and market dynamic in alternative protein industry. Also, this dissertation provides insights for applied economists in utilizing diverse methodologies, including econometric models, machine learning techniques, and/or the combination of them, to provide robust and valid empirical evidence in the field of agricultural and food economics.
Read
