Frank Davenport

Combining insights from quantile and ordinal regression: Child malnutrition in Guatemala

Chronic child undernutrition is a persistent problem in developing countries and has been the focus of hundreds of studies where the primary intent is to improve targeting of public health and economic development policies. In national level cross-sectional studies undernutrition is measured as child stunting and the goal is to assess differences in prevalence among population subgroups. Several types of regression modeling frameworks have been used to study childhood stunting but the literature provides little guidance in terms of statistical properties and the ease with which the results can be communicated to the policy community. We compare the results from quantile regression and ordinal regression models. The two frameworks can be linked analytically and together yield complementary insights. We find that reflecting on interpretations from both models leads to a more thorough analysis and forces the analyst to consider the policy utility of the findings. Guatemala is used as the country focus for the study.

Using time series structural characteristics to analyze grain prices in food insecure countries

Two components of food security monitoring are accurate forecasts of local grain prices and the ability to identify unusual price behavior. We evaluated a method that can both facilitate forecasts of cross-country grain price data and identify dissimilarities in price behavior across multiple markets. This method, characteristic based clustering (CBC), identifies similarities in multiple time series based on structural characteristics in the data. Here, we conducted a simulation experiment to determine if CBC can be used to improve the accuracy of maize price forecasts. We then compared forecast accuracies among clustered and non-clustered price series over a rolling time horizon. We found that the accuracy of forecasts on clusters of time series were equal to or worse than forecasts based on individual time series. However, in the following experiment we found that CBC was still useful for price analysis. We used the clusters to explore the similarity of price behavior among Kenyan maize markets. We found that price behavior in the isolated markets of Mandera and Marsabit has become increasingly dissimilar from markets in other Kenyan cities, and that these dissimilarities could not be explained solely by geographic distance. The structural isolation of Mandera and Marsabit that we find in this paper is supported by field studies on food security and market integration in Kenya. Our results suggest that a market with a unique price series (as measured by structural characteristics that differ from neighboring markets) may lack market integration and food security.

Open Trade, Price Supports, and Regional Price Behavior in Mexican Maize Markets

Abstract We analyze wholesale maize prices in 12 Mexican markets from 1998 to 2010, a period when markets became more open to inter- and intranational trade. We ask how the influence of global and local forces on Mexican maize prices changed during this period. We also explore how the strength of global and local forces varies across maize-producing regions. In general, we expect the influence of global forces to increase and local forces to decrease as markets become more open. We find that the influence of global forces does vary over the study period and, counter to expectation, is the highest at the beginning and middle of the period rather than at the end. This result suggests that even under less open market conditions, buyers and sellers were still following global price signals. In contrast, the influence of local forces follows expectation and decreases over time. However, the estimated pattern of response is not uniform across various maize-producing regions. Taken together, our results suggest that opening agricultural markets can result in regionally distinct outcomes and counterintuitive price behavior.

Anthropogenic Enhancement of Moderate-to-Strong El Niño Events Likely Contributed to Drought and Poor Harvests in Southern Africa During 2016

Introduction. In December–February (DJF) of 2015/16, a strong El Niño (Niño‐3.4 SST >29°C) contributed to a severe drought over southern Africa (SA; Funk et al. 2016). A 9‐million ton cereal deficit resulted in 26 mil‐ lion people in need of humanitarian assistance (SADC 2016). While SA rainfall has a well‐documented nega‐ tive teleconnection with Niño‐3.4 SSTs (Hoell et al. 2015, 2017; Jury et al. 1994; Lindesay 1988; Misra 2003; Nicholson and Entekhabi 1987; Nicholson and Kim 1997; Reason et al. 2000; Rocha and Simmonds 1997), the link between climate change and El Niño remains unclear (Christensen et al. 2013) due to the large natural variability of ENSO SSTs (Wittenberg 2009), uncertainties surrounding measurements and trends (Solomon and Newman 2012), intermodel differences in ENSO representation and feedbacks (Guilyardi et al. 2012; Kim et al. 2014), and difficulties associated with quantifying ENSO strength (Cai et al. 2015). Figure 18.1a highlights observational uncertain‐ ties (Compo and Sardeshmukh 2010; Solomon and Newman 2012) using four datasets: ERSSTv4 (Huang et al. 2015), HadISST (Rayner et al. 2003), Kaplan SST (Kaplan et al. 1998), and Hurrell (Hurrell et al. 2008). These products differ substantially in their represen‐ tation of cool events and Niño‐3.4 variance. Two SST products indicate significant upward trends; two SST products do not. These data have been standardized to remove systematic differences in variance. Focusing just on the behavior of moderate–strong El Niño events (MSENEs), we can produce more ro‐ bust (first order) statistics by comparing the means of the top ten warmest Niño‐3.4 events between 1921–80 and the top six warmest events between 1981–2016. Rather than using a set SST threshold, MSENEs are defined as 1‐in‐6‐year warm events. This provides a simple nonparametric approach that takes advantage of the well understood quasi‐periodic nature of ENSO to identify MSENEs across multiple models and simulations. Modest changes in the number of events (say, 1‐in‐7 or 1‐in‐5) produced modest increases and decreases in El Niño temperatures, but did not sub‐ stantially change the results. We begin our analysis in 1921 (because ship data before 1921 is limited), and divide the remaining 96 years into two time periods with relatively weak and strong radiative forcing, respectively. Examining changes in MSENE means (horizontal lines in Fig. 18.1a), we find that all the observational datasets identify significant increases (Fig. ES18.1 examines ERSSTv4 errors). Note that we are not explicitly ex‐ amining changes in ENSO variance, ENSO means, or Niño‐3.4 SST trends, but only Niño‐3.4 magnitudes AFFILIATIONS: Funk—U.S. Geological Survey, Center for Earth Resources Observation and Science, and UC Santa Barbara Climate Hazards Group, Santa Barbara, California; Davenport, harrison, shukLa, pomposi, anD husak—UC Santa Barbara Climate Hazards Group, Santa Barbara, California; magaDzire, gaLu, anD koreCha—UC Santa Barbara Climate Hazards Group, Santa Barbara, California, and Famine Early Warning Systems Network; artan—Intergovernmental Authority on Development (IGAD) Climate Prediction & Applications Centre, Nairobi, Kenya; inDeje—IGAD USAID/Kenya and East Africa Planning for Resilience in East Africa Through Policy Adaptation, Research, and Economic Development, Nairobi, Kenya; maCharia—Regional Center for Mapping of Resources for Development, Nairobi, Kenya; nsaDisa—Director of the Southern African Development Community’s Climate Services Centre.

How will East African maize yields respond to climate change and can agricultural development mitigate this response

We analyze the response of Kenyan maize yields to near-term climate change and explore potential mitigation options. We model county level yields as a function of rainfall and temperature during a period of increased regional warming and drying (1989–2008). We then do a counter factual analysis by comparing existing maize yields from 2000 to 2008 to what yields might have been if observed warming and drying trends had not occurred. We also examine maize yields based on projected 2026–2040 climate trends. Without the observed warming and drying trends, Eastern Kenya would have had an 8% increase in maize yields, which in turn would have led to a net production increase of 500,000 metric tons. In Western Kenya, the magnitude of change is higher but the relative changes in predicted values are smaller. If warming and drying trends continue, we expect future maize yields to decline by 11% in Eastern Kenya (vs. 7% in Western Kenya). We also examine whether these future losses might be offset through agricultural development. For that analysis, we use a household panel dataset (2000, 2005) with measurements of individual farm plot yields, inputs, and outputs. We find that under a scenario of aggressive adoption of hybrid seeds and fertilizer usage coupled with warming and drying trends, yields in Western Kenya might increase by 6% while those in Eastern Kenya could increase by 14%. This increase in yields might be larger if there is a corresponding increase in usage of drought-tolerant hybrids. However, wide prediction intervals across models highlight the uncertainty in these outcomes and scenarios.