Arne Henningsen

Most research questions in agricultural and applied economics are causal in nature: they study how changes in one or more variables (such as policies, prices or weather) affect one or more other variables (e.g., income, crop yields or pollution). Only a minority of these research questions can be studied with experimental methods, so most empirical studies in agricultural and applied economics rely on observational data. However, estimating causal effects with observational data requires an appropriate research design and a transparent discussion of all identifying assumptions, together with a critical discussion of how plausible they are. This paper provides an overview of approaches that are frequently used in agricultural and applied economics to estimate causal effects with observational data. It then provides advice and guidelines for agricultural and applied economists seeking to estimate causal effects with observational data, including how to assess and discuss the identification strategies adopted in their analysis.

We provide a ‘big picture’ of what Open Science is and what benefits, limitations, and risks it entails for agricultural economists. We show that Open Science comprises various aspects, such as the accessibility of science, transparency of scientific processes, open and findable, accessible, interoperable, and reusable (FAIR) research data and code, and openness in teaching and education. We identify potential benefits of Open Science for individual researchers and the public, as well as adoption barriers. We highlight that public benefits of a widespread uptake of Open Science practices still remain unexplored. We share best practice examples for key aspects of agricultural economic research, i.e. primary data collection and analysis, optimization and simulation models, use of replication packages, and an Open Science Community. Assessing the author guidelines of twelve agricultural economics journals for their Open Science practices, we find that data citation and transparency are considered important in many journals already, whereas replication, pre-registration, or results-blind reviews are encouraged but rarely enforced. It also becomes evident that the journals differ in terms of how strictly they enforce their open science guidelines. We close by providing recommendations for researchers, journal editors, policymakers, universities, research institutes, and funding agencies to better align public benefits with private incentives.

Yield-enhancing agricultural technologies, like improved crop varieties, are widely promoted in developing countries to improve the food security, income, and welfare of farm households. Nonetheless, farm households show low adoption rates of these technologies. To gain more insight into the drivers of and the barriers to the adoption of improved crop varieties, we study the adoption of improved sunflower varieties by smallholder farmers in Tanzania. Unlike most earlier studies, we distinguish between the initial adoption and the extent of adoption. Additionally, we investigate the roles of market constraints and liquidity constraints, which are largely ignored in previous adoption studies. We find that risk aversion and liquidity constraints are barriers to adoption, whilst radios and extension service are important information channels for new technologies. Our results can help to improve policies, development programmes, and business decisions and finally to enhance agricultural productivity and farm household welfare.

Animal-level responses to weather variability in US dairy systems are well described, but the potential of housing and other farm management practices (for example, fans and sprinklers) to moderate the impacts of weather remains uncertain. Here we assess the influence of historical variation in the temperature–humidity index (THI) on milk yields using monthly state-level yield data and high-resolution daily weather data over 1981–2018. We find that milk yields are compromised by exposure to both extreme heat (>79 THI) and cold (<39 THI), causing average daily yield decreases of around 3.7% and 6.1%, respectively, relative to optimal conditions (65–69 THI). Colder regions are more sensitive to heat extremes, and warm regions are more sensitive to cold extremes. Sensitivity to THI has reduced dramatically over time. Climate trends contributed modestly (around 0.1% over 38 years) to rising yields in most states via alleviating cold stress, although more extreme future conditions may negate these benefits.

Summary: The inverse hyperbolic sine (IHS) transformation is frequently applied in econometric studies to transform right-skewed variables that include zero or negative values. We show that regression results can heavily depend on the units of measurement of IHS-transformed variables. Hence, arbitrary choices regarding the units of measurement for these variables can have a considerable effect on recommendations for policies or business decisions. In order to address this problem, we suggest a procedure for choosing units of measurement for IHS-transformed variables. A Monte Carlo simulation assesses this procedure under various scenarios, and an empirical illustration shows the relevance and applicability of our suggested procedure.