TH Javier E. Baez, German Caruso, and Chiyu Niu1 Thanks to strong economic growth over the last two decades, poverty in Mozambique has decreased and the average household is now more likely to access basic education, health, and housing. Yet, the country is still ravaged by intense and frequent weather disasters. To determine the scale and nature of the impacts of these shocks, this paper analyzes the vulnerability of rural livelihoods across three different extreme weather events: droughts, floods and cyclones. The study finds that per capita food and non-food consumption and asset ownership are reduced among households affected by any of the three weather shocks. Their children are less likely to attend school, have a higher probability of falling sick and show higher engagement in paid and unpaid work. What’s more, staple food prices are disrupted and remain affected nearly a year after the disaster. Helping households confront these events requires comprehensive risk management policies, including making agriculture more resilient to weather, improving the functioning of credit and insurance markets, facilitating economic diversification and market access, and increasing the availability of flexible safety nets – all before the shocks occur. Mozambique’s high levels of exposure and in 1991-2015 than from 1968-1990. These issues vulnerability to weather shocks threaten combined make low-income Mozambicans economic and social progress. Mozambique has especially vulnerable to more natural disasters in the recovered well after the end of its war in 1992. Its future. If Mozambique wants to end extreme poverty GDP expanded by 7.2 percent from 2000 to 2016 and by 2030 and achieve sustained growth, it must assess the share of Mozambicans living below the poverty and address the comprehensive impacts of these line decreased from 60.3 to 48.4 percent between shocks. 2002/03 and 2014/15. But the regular occurrence of severe weather events risks be putting a brake on this progress. Mozambique ranks high among African countries in terms of exposure to natural disasters like cyclones, floods, and droughts. Nearly a quarter of the population lives in areas extremely likely to This work investigates whether households in experience a natural disaster and climate change is Mozambique are systematically vulnerable to only expected to amplify this vulnerability. For major weather shocks, irrespective of the type of instance, tropical storms were four times more likely event. Households can be affected by shocks in to impact any given rural area during storm seasons 1 This note summarizes the key findings of a larger working paper: Baez, Javier; Caruso, German and Niu, Chiyu. 2018. “Extreme Weather and Poverty Risk: Evidence from Multiple Shocks in Mozambique”. Policy Research Working Paper; No. 8667. World Bank, Washington, DC. © World Bank. many ways: they can lose their assets, increase how systematically undermined among affected much work they do, eat less food, cut back on other households across the disasters. People cut back on basic consumption and decide to put their children the most basic of needs; for instance, people to work. They may also self-insure ex-ante, opting for consumed 25-30 percent less food due to the shocks, lower risk, lower return economic activities. Most as seen in Figure 1. Back-of-the-envelope studies of this kind measure the impact of a single calculations without accounting for possible event on some of these outcomes. However, by substitution effects suggest that affected individuals concurrently analyzing the effects of three reduced their average intake by nearly 150-200 independent weather shocks across time and calories per day. This is important in a context where location, the empirical analysis can document 43 percent of children under five show serious signs whether household welfare is more sensitive to some of height stunting for their age. Households also weather risks than others. consumed less of everything else. Cyclone Jokwe, for example, reduced household consumption by over The study investigates the impacts of three large half (54.8 percent), and the drought reduced weather shocks. (1) the floods that occurred consumption by nearly a quarter (21 percent) relative between late 2006 and early 2007 along the Zambezi to those unaffected. The impact even pushed some River Basin, (2) a category 3 cyclone (called Jokwe) of the households below the poverty line; both the that struck northeastern Mozambique between drought and the cyclone increased the poverty March 8 and 18, 2008, and (3) a drought that affected headcount by 12 and 17.5 percentage points parts of central and southern Mozambique between respectively. May 2005 to January 2006. The empirical analysis is Figure 1. Food consumption per capita fell significantly produced using multiple household surveys, weather among households in affected areas data from ground stations and satellite data and spatial data on main crop growing cycles. The study also performed multiple robustness tests to issues such as migration, alternative survey samples or different definitions of affected areas. The analysis also examined the effects of fourth, more recent event: the floods that affected Zambezia, Nampula and Niassa provinces in central and northern Mozambique in early 2015. Note: Parameter estimates of triple-difference models expressed as the effect on the dependent variable of a change of one standard deviation in the shock intensity measure. Asterisks denote statistical significance: *p < 0.1, **p < 0.05, ***p<0.01. Weather shocks defined in standardized z-scores of totals in the year of the event using the historical distribution from the reference period 2000-2012 for each district. Source: Baez, J., G. Caruso and C. Niu (2018) The findings show that households are highly Shocks also deeply impacted children’s welfare, vulnerable to all the shocks analyzed across across multiple measures. Overall, children from multiple measures of welfare. Food security is affected households were 8.3 percentage points less likely to attend school relative to unaffected children. Figure 2. Asset ownership fell among affected households They also became less healthy: child morbidity rates increased by fourfold in flooded districts. Intriguingly, child morbidity fell by nearly half in districts after the 2005 drought, so further study is needed to figure out what happened to cause the decline. A typical coping response of shock-affected households is to increase how much adults and children work. The evidence for Mozambique also confirms this behavior. 18 to 65-year-olds are 33 Note: Change in wealth index of households in affected areas relative percentage points more likely to be either working or to unaffected households. Parameter estimates of triple-difference actively looking for work, and children (ages 5 to 17) models expressed as the effect on the dependent variable of a change of one standard deviation in the shock intensity measure. Robust also increase their labor activities. The rise in child standard errors clustered at the district level. Asterisks denote labor attributed to the floods is very dramatic: a statistical significance: *p < 0.1, **p < 0.05, ***p<0.01. Weather shocks defined in standardized z-scores of totals in the year of the event fourfold increase relative to the level observed at using the historical distribution from the reference period 2000-2012 for each district. baseline. Source: Baez, J., G. Caruso and C. Niu (2018) Asset holdings, including productive assets, were found to decline among affected households. The study also found that households hit by cyclone Jokwe lost one-third of their belongings, and those The shocks translate into increased food price affected by the drought lost nearly one-fifth (Figure variability. Agriculture in Mozambique is largely 2). Given the localized nature of the shocks, it is traditional, practiced mostly by smallholder farmers possible that their assets were sold at lower prices that rely on rainfall for water and have low input than normal, since disasters pressure many utilization and technology adoption. The farmers households in affected areas to sell assets at roughly mainly consume what they grow, are isolated from the same time. The loss of assets hurts their ability to others, and the agricultural markets are not well generate income and grow out of poverty in the long integrated. Hence, extreme weather is expected to run. Affected households also obtained transfers disrupt crop yields, triggering supply- and demand- from other households at higher rates to cope with side effects in food markets. To analyze the effects of the disasters’ effects, an indication of the presence of the shocks on the prices of staples, the study tracked informal risk-sharing arrangements across maize prices for three reasons: it is the staple most households. For instance, after Cyclone Jokwe, widely produced and marketed, its yields are highly affected households were almost twice as likely to determined by the timing and quantity of rain, and it receive transfers compared to levels recorded from constitutes a significant part of the Mozambican diet. unaffected communities. Overall, the analysis finds that maize prices close to affected locations were deeply affected by weather shocks (Figure 3). However, weather shock impacts prices differently. For example, during the 2005-2006 drought, prices rose by 78.9 percent and remained 29.3 percent higher one year after the drought. remain functional to bolster communities already However, the floods in 2007 followed the opposite deeply affected by the shocks. Third, countries must pattern; we found maize prices in affected locations make human and physical capital more “resilient”. For decreased by over 40 percent for the first three instance, they should strengthen country-wide months. immunization, standardize building codes, build flood defenses, and incentivize household insurance. Figure 4. Maize prices in markets close to affected areas Fourth, countries must invest in learning more about are sensitive to the occurrence of weather shocks the impacts of risks. This is done by increasing the ability of organizations to collect statistical and satellite data, integrating weather forecasting and early warning systems, and supporting more applied research in this field. Finally, scalable safety net policies are necessary to ensure that individuals can experience minimally acceptable standards of living after shocks occur. Note: Graphs show point estimates of the effect on prices for an increase of a standard deviation in the disaster intensity. Bars show 10% confidence intervals. Impact estimates calculated for each month in a 12-month period following the weather shock. Weather shocks defined in standardized z-scores of totals in the year of the event using the historical distribution from the reference period 2000- 2012 for each district. ABOUT THE AUTHOR Source: Baez, J., G. Caruso and C. Niu (2018) Javier E. Baez is a Senior Economist at the World Bank’s Poverty and Equity Global Practice and a Research Fellow at the Institute of the Study of Labor (IZA). jbaez@worldbank.org Comprehensive disaster risk management German Caruso is an Economist at the World policies need to be in place to increase the Bank’s Poverty and Equity Global Practice. resilience of vulnerable households and keep gcaruso@worldbank.org natural shocks from turning into natural disasters. First, countries must provide basic public Chiyu Niu is a consultant at the World Bank’s goods before shocks occur. This is done by investing Poverty and Equity Global Practice. in goods and services like education, health, water cniu@worldbank.org and sanitation, and connectivity. Households with more human capital and access to better economic opportunities can better weather the effects of shocks. Second, countries must make sure essential markets, like agriculture, insurance, credit, and labor This note series is intended to summarize good practices and key policy findings on Poverty-related topics. The views expressed in the notes are those of the authors and do not necessarily reflect those of the World Bank, its board or its member countries. Copies of these notes series are available on www.worldbank.org/poverty