Relating rainfall patterns to agricultural income: Implications for rural development in Mozambique


Material Information

Relating rainfall patterns to agricultural income: Implications for rural development in Mozambique
Series Title:
Weather, Climate and Society
Physical Description:
Journal Article
Matyas, Corene
Publication Date:


Rural farmers in Mozambique rely on rainfed agriculture for food and income. Yet they 39 experience high rainfall variability ranging from extreme drought to flooding rainfall 40 from tropical cyclone systems. To explore linkages between rainfall and agriculture, we 41 regress changes in annual household per capita agricultural income on reliance on staple 42 food crops, agricultural and demographic characteristics, and rainfall patterns using 43 longitudinal data for rural households for 2002 and 2005. We characterize rainfall 44 patterns by defining nine rainfall zones using the percent of normal rainfall received in 45 each month of three agricultural growing seasons and rainfall from two tropical cyclones 46 that occurred during the study period. Results show that in a period where monthly 47 rainfall seldom occurred in normal amounts, most households experienced decreases in 48 agricultural income. Even after controlling for rainfall patterns, we find that greater 49 household dependency on staple crop agriculture is associated with declining annual 50 agricultural income. We also find that areas affected by both wet and dry rainfall 51 extremes in the first year of the study had decreases in the well-being of rural households 52 when measured two years later. Taken together, our findings suggest that anti-poverty 53 policies focused on increasing agricultural income seem likely to fail in countries 54 characterized by highly variable rainfall and exposure to extreme events, particularly 55 when coupled with high levels of poverty and widespread dependence on rainfed 56 agriculture.
Collected for University of Florida's Institutional Repository by the UFIR Self-Submittal tool. Submitted by Corene Matyas.
Publication Status:

Record Information

Source Institution:
University of Florida Institutional Repository
Holding Location:
University of Florida
Rights Management:
All rights reserved by the submitter.
System ID:

This item is only available as the following downloads:

Full Text


Weather, Climate, and Society EARLY ONLINE RELEASE This is a preliminary PD F of the author-produced manuscript that has been peer-reviewed and accepted for publication. Since it is being posted so soon after acceptance, it has not yet been copyedited, formatted, or processed by AMS Publications. This preliminary version of the manuscript may be downloaded, distributed, and cited, but please be aware that there will be visual differences and possibly some content differences between this version and the final published version. The DOI for this manuscrip t is doi: 10.1175/WCAS-D-13-00012.1 The final published version of this manuscript will replace the preliminary version at the above DOI once it is available. If you would like to cite this EOR in a separate work, please use the following full citation: Silva, J., and C. Matyas, 2013: Relating Ra infall Patterns to Agricultural Income: Implications for Rural Development in Mozambique. Wea. Climate Soc. doi:10.1175/WCAS-D-13-00012.1, in press. 2013 American Meteorological Society AMERICAN METEOROLOGICAL SOCIETY


1 Relating Rainfall Patterns to Agricultural 1 Income : Implications for Rural Development in 2 Mozambique 3 4 5 Julie A Silva 1 6 Department of Geographical Sciences, 7 University of Maryland, College Park 8 9 10 Corene J Matyas 11 Department of Geography, University of Flori da 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 1 Corresponding author address : Julie Silva Department of Geographical Sciences, 31 University of Maryland, College Park, 2181 Lefrak Hall, College Park, MD 20742 USA. 32 E mail: 33 34 \000DQXVFULSW\003\013QRQ\020/DH;\014 &OLFN\003KHUH\003WR\003GRZQORDG\003\000DQXVFULSW\003\013QRQ\020/DH;\014\035\003$\0006BUHVXEPLVVLRQB\025\023\024\026B\023\032B\024\024BY\025\021GRF\003


2 35 Abstract 36 37 Rural farmers in Mozambique rely on rainfed agriculture for food and income. Yet they 38 experience high rainfall variability ranging from extreme drought to flooding rainfall 39 from tropical cyclone systems To explore linkages between rainfall and agriculture, we 40 regress changes in annual h ousehold per capita agricultural income on reliance on staple 41 food crops, agricultural and demographic characteristics, and rainfall patterns u sing 42 longitudinal data for rural households for 2002 and 2005. We characterize rainfall 43 patterns by defining nine rainfall zones using the percent of normal rainfall received in 44 each month of three agricultural growing seasons and rainfall from two tropical cyclones 45 that occurred during the study period. Results show that in a period where monthly 46 rainfall seldom occ urred in normal amounts, most households experienced decreases in 47 agricultural income. Even after controlling for rainfall patterns, we find that greater 48 household dependency on staple crop agriculture is associated with declining annual 49 agricultural incom e We also find that areas affected by both wet and dry rainfall 50 extremes in the first year of the study had decreases in the well being of rural households 51 when measured two years later. Taken together, our findings suggest that anti poverty 52 policies focu sed on increasing agricultural income seem likely to fail in countries 53 characterized by highly variable rainfall and exposure to extreme events, particularly 54 when coupled with high levels of poverty and widespread dependence on rainfed 55 agriculture. 56 57


3 1. Introd uction 58 The endemic poverty of rural agriculturalists in many less developed countries (LDCs) 59 together with growing evidence of climate change has led to increasing interest in the 60 effects of weather on the well being of these populations. Since the 1960s the dominant 61 approach to rural development in low income countries has focused on strengthening the 62 small farm agricultural sector in order to alleviate rural poverty ( IFPRI 2002; Ellis and 63 Biggs 2001 ; Ellis 2000 ). Yet, t he agricultural sector across sub Saharan Africa is highly 64 vulnerable to climate variability and extremes ( Easterling et al. 2007; Thomas et al. 2007; 65 Parry et al. 2004 ; Rosenzweig and Parry 1994; Molua 2002 ) d ue to its high dependence 66 on rainfed agriculture Research suggests that the re gion will experience decreased annual 67 rainfall and increased warming overall and exposure to extreme weather events is 68 predicted to increase locally ( Easterling et al 2007 ). A need exists to collectively 69 examine rainfall patterns dependence on crop agri culture and rural well being to inform 70 policies aimed at mitigating climate vulnerability. 71 Mozambique provides an illustrative region to examine vulnerability to climate 72 variability and extreme events in southern Africa. The country exemplifies many of th e 73 challenges facing LDCs, including extreme poverty, low levels of human and physical 74 capital, heavy aid dependency, and limited capacity to cope with climate change. 75 Mozambique is arguably one of the poorest nations in the world, and its low rankings on 76 s tandard development indicators such as GDP per capita (ranked 209 out of 226 77 countries) and the Human Development Index (ranked 165 out of 169 countries) reveal 78 that the vast majority of Mozambicans experience a very poor quality of li fe (CIA 2013, 79 UNDP 20 12 ) Nearly 70 per cent of the population lives in rural areas and most 80


4 households depend on rainfed subsistence agriculture (INE 2008; Mather et al 2008). 81 Thus, weather related shocks to agricultural households have considerable potential to 82 exacerbate e xisting poverty. A better understanding of how economic vulnerability of 83 rural households is driven by agricultural dependency and how this varies across regions 84 characterized by different rainfall patterns could have a direct impact on poverty 85 reduction improved food security and the ability for rural people to maintain social 86 norms and traditions that they have reason to value in developing countries such as 87 Mozambique. 88 Recent empirical work finds that the weather is one of the most significant factors 89 determining the well being of rural Mozambican households. Mather (2012), Cunguara 90 (2008), and Mather et al. (2008) find that a high number of drought days in the growing 91 season ha ve a strong, negative effect on agricultural income, primarily through decre ases 92 in maize and cassava production. Cunguara and Darnhoefer (2011) find that the use of 93 improved agricultural technologies by rural Mozambican farmers had no statistically 94 significant impact on household income due to drought conditions during their study 95 period. However, the se studies did not report on the effects of high rainfall events. 96 Cunguara and Kelly (2009) find only minimal statistical significance for the majority of 97 monthly rainfall totals during the growing season on successful outcomes for hou seholds 98 farming maize Their weather related findings could be due to the fact that examining 99 discreet monthly totals does not account for the complexity of the rainfall patterns over 100 space and time In addition, all four studies call for more in depth res earch linking 101 weather to agriculture. 102


5 The goal of our study is to relate distinctive rainfall patterns to changes in crop 103 income for farmers in Mozambique by testing two main hypotheses. First, we 104 hypothesize that the differing economic outcomes for rura l agriculturalists will be 105 explicable, in part, by the varying rainfall patterns that households experience over the 106 study period. Second, w e expect that high er dependency on income from staple food 107 crops will be associated with declining agricultural inco mes. We group households into 108 zones based on intraseasonal rainfall patterns before we test the first hypothesis. This 109 method allows us to consider both wet and dry events as well as identify households that 110 experienced similar rainfall patterns. Using the most recently available longitudinal data 111 on rural households from surveys conducted in 2002 and 2005 we utilize regression 112 analysis to examine the effects of distinctive rainfall patterns on crop income while 113 controlling for agricultural and demographic characteristics of households. After 114 discuss ing the variability of rainfall throughout the study period we demonstrate that 115 both wet and dry events are associated with the largest negative changes in agricultural 116 income across the study period. By measur ing the effects of staple crop dependency on 117 agricultural incomes in a period characterized by highly variable and extreme weather 118 this study also has implications regarding 119 strategy 120 121 2. Characteristics of the Study R egion 122 There are marked differences between the northern, central, and southern regions of 123 Mozambique with regard to poverty and development stemming from both the 124 Portuguese colonial legacy and post independence government policies (Isaacman 1996 ; 125


6 Wuyts 19 96, 2003 ; Pitcher 2000; Sheldon 2002 ). Colonial policies channeled 126 development into the south of the country near the capital city Maputo In addition, South 127 African mining companies heavily recruited workers from the southern provinces 128 allowing households in this area more access to remittance income and financial capital 129 to invest in farming (Hermele 1992) This trend of southern economic dominance has 130 continued, in part because of infrastructure damage resulting from two protracting armed 131 conflicts: the war for independence ( 1961 1974 ) and the civil war ( 1977 1992 ). Thus the 132 southern regions are characterized by higher household incomes, better market access, 133 and greater linkages to the global economy (Silva 2007, 2008). The northern and central 134 regions of the country have better agricultural potential than the south, but poor 135 infrastructure and great distance limits access to markets (Cunguara and 136 Darnhofer, 2011) and contributes to endemic poverty in those areas (Silva 2007) 137 Insert Figure 1 138 In terms of climate, bad weather can occur in different parts of the country during 139 the same year. For example, during the three year period of this study during which 140 drought conditions have been documented ( Reason and Phaladi 2005 ; Rouault and 141 Richard 2005 ) two tropical cyclones made landfall in Mozambique and produced more 142 than 150 mm of rainfall during their p assage in different portions of the country (Matyas 143 and Silva 2013) 1 The coastline of Mozambique lies between 11 27 S latitude within the 144 Southern Indian Ocean basin, a latitudinal zone which favors the tracks of tropical 145 cyclones (Ash and Matyas 2012) (Fig 1) Mozambique experiences approximately two 146 landfalling storms every three years (Matyas and Silva 2013) which can result in flooding 147 rainfall and wind related damage 148 1 The time period of each cyclone is discussed in detail in the section describing the study region.


7 R ainfall in Mozambique is hig hly variable from year to year ( Washington and 149 Preston 2006 ; Shongwe et al. 2009 ) On average, the northern and cen tral regions receive 150 100 0 and 1100 mm of ra infall during the main growing season months of November 151 March while only 600 mm is received during the October March growing season near 152 Maputo (FAO 2003). Multi day rainfall events occur when tropical temperature troughs 153 (TTTs) develop over the regio n ( Todd and Washington 1999 ; Usman and Reason 2004 ; 154 Manhique et al. 2011 ) and the frequency and locations of these events often correspond 155 with the phase of the E l Nio S outhern O scillation (ENSO) ) La Nia conditions are 156 associated with high rainfall due to moisture advec tion from the South Indian Ocean and 157 a stronger than normal Angola low which leads to the development of the TTTs. D ry 158 conditions occur in conjunction with El Nio events as convection shifts offshore east of 159 Madagascar ( Jury et al. 2004 ; Usman and Reason 2004 ; Reason et al. 2005 ; Alemaw and 160 Chaoka 2006 ; Kruger 2007 ; Manhique et al. 2011 ) These weather patterns impact 161 agricultural production by causing delays in planting forcing crops to be replanted 162 decreasing crop yields, and forcing farmers to alter traditional agricultural practices in 163 response to altered conditions (Silva et al 2010). Previous studies have found 164 Mozambican farmers to be highly vulnerable to such w eather extremes because of 165 among other reasons, their limited use of irrigation and high levels of poverty ( Arndt and 166 Baou 2000 ; Patt and Schroter 2008; Hahn et al 2009 ; Sietz et al. 2011 ). 167 Due to the heavy reliance on rainfed agriculture and high rain fall variability, this 168 study analyze s rainfall patterns across Mozambique during the three summer growing 169 seasons occurring between the TIA surveys in 2002 and 2005. As few complete long 170 term rain gauge rec ords exist for Mozambique the paucity of ground b ased 171


8 measurements would lead to highly inaccurate in terpolation at the household level 172 Instead, we utilize satellite based estimates of monthly rainfall from the Tropical Rainfall 173 Measuring Mission (TRMM) 3B43 product (Huffman et al. 2007) Adeyewa and 174 Ka kamura ( 2003 ) found that these TRMM estimates closely matched those of the rain 175 gauges over southern Africa. O ne pixel spans 0.25 latitude and longitude so that 1075 176 pixel s facilitate the analysis of mesoscale rainfall patterns over Mozambique. We obtain 177 data from 1998 2009 using the GES DISC Interactive Online Visualization ANd 178 179 Data and Information Services Center (DISC). A GIS interpolate s the data using ordinary 180 spherical kriging across the area 9 30 S and 26 46 E As locational coordinates are 181 only available for village s monthly rainfall totals are assigned to all households within 182 each village and a 12 year climatology is then calculated f or each month. As monthly 183 rainfall var ies widely across the country, we calculate the percentage of the 12 year 184 normal rainfall received in each study month at each village and perform our analyses 185 with these variables instead of the actual monthly rainfal l totals. Thus, our study expands 186 the rainfall time series utilized by Cunguara and Kelly (2009) that examined monthly 187 rainfall totals over six years. 188 Our goal is to group households together that experienced similar rainfall patterns 189 during the primary su mmer growing season months. Thus, we analyze data for November 190 March over the entire country and include October for the three southern provinces of 191 Gaza, Inhambane, and Maputo as their growing season typically begins in this month 192 Although some crops a re harvested in April (FAO 2004) we omit rainfall from this 193 month due to inaccuracies in the TRMM data (Matyas and Silva 2013). The growing 194


9 seasons included in the current study are : 1 ) October/November 2002 March 2003 2) 195 October /November 2003 March 2004 and 3) October/November 2004 March 2005. We 196 also calculate storm total rainfall from three tropical cyclones using the TRMM 3B42 197 daily rainfall product. Storm coordinates from the Joint Typhoon Warning Center ( JTWC 198 2003 ) are utilized to determine the days upon which any tropical cyclone could have 199 produced rainfall over Mozambique. Daily rainfall was sum med for November 11 12 for 200 Cyclone Atang, December 31, 2002 January 6, 2003 for Cyclone Delfina and March 1 201 5, 2003 for Cyclone Japhet. 202 To measure the extent to which regional rainfall patterns can affect household 203 income, we subdivide the 53 6 villages into rainfall zones. This approach differs from 204 those utilized by previous studies in that we identify patterns of rainfall ranging from 205 extreme wet events to extremely dry periods so as to set each group of households apart 206 from others across the country. Villages receiving at least 150 mm of rainfall from a 207 tropical cyclone as these are the two most 208 extreme wet events to affect at least 30 villages Atang only produced high rainfall over 209 12 villages making it too small of an event to warrant its own group Two separate 210 hierarchical cluster analyses emphasizing between groups linkages are then performed 211 with the remaining 402 villages. Fifteen study months are analyzed to produce four 212 rainfall zones in the nort hern and central portions of the country, while eighteen months 213 are entered into the second analysis for 115 villages located in the three southern 214 provinces. T hree rainfall zones are identified in this region. Thus, we identify nine zones 215 that characteriz e different patterns of intraseasonal rainfall variability. 216


10 a. Rainfall Patterns 217 Our 12 year climatology approximates the totals reported by the FAO (2003), showing 218 that o n average, rainfall is highest in the central portion of the country while areas in th e 219 south receive less than 600 mm during the growing season (Fig. 2a) Although previous 220 research has identified all three seasons as drought years, rainfall deficits did not occur in 221 every study month, and intraseasonal variability in rainfall was high thr oughout the 222 country As a result n one of the three seasons closely resembled the 12 ye ar average 223 (Fig 2b d ). Rainfall totals varied the most during Season 1 as the three tropical cyclones 224 and several multi day rainfall events during January not related t o Delfina kept the 225 northern and central regions above normal in terms of seasonal rainfall (Fig. 3a) In the 226 south however, rainfall was less than 50% of normal overall and most locations did not 227 receive rainfall that was above normal after October Season 2 rainfall patterns were 228 closer to normal, but many areas still experienced rainfall deficits (Fig. 3b) Season 3 229 started wetter than normal in the north and center of Mozambique, but by February, zone 230 averaged deficits ranged from 40 80% of normal nation wide and these continued 231 through March (Fig. 3c) 232 We next provide a brief description of the nine rainfal l zones We discuss the 233 combine d results of our TRMM based analysis with data from the Food and Agriculture 234 Organization of the United Nations (FAO) on rainfall patterns, main crops planted, and 235 projected yields for each province (FAO 2003, 2004, 2005). In most cases our rainfall 236 zones cross multiple provinces as the cluster analysis considered the percentage of 237 normal monthly rainfall for each village without regard for administrative boundaries. 238 Zones are numbered from north to south (Fig. 1) 239


11 Insert Figure 2 240 Located in Cabo Delgado Province near the border with Tanzania, Zone 1 is the 241 third smallest zone and includes 33 villages The most distinguish ing rainfall 242 characteristic for this zone was that Atang made landfall in Tanzania (JTWC 2003) and 243 brought more than 100 mm of rainfall to 12 villages during November 11 12 Seventeen 244 villages received more than 75% of their normal monthly rainfall from th is tropical 245 cyclone, which represents the most extreme rainfall event in the zone. Although total 246 accumulated rainfall was lower than normal during Season 2 due in large part to 247 November 2003 good crop yields were reported as well as in Season 3 when the seasonal 248 average was near normal. Cassava is the main crop planted within the fertile river valley 249 soils, and cereal crop production is normally very high in this region. 250 Insert Figure 3 251 Most Zone 2 villages (66) are located in the provinces of Nampula (47 %) and 252 Cabo Delgado (45%) Some areas in Nampula received high rainfall from Delfina and 253 suffered crop losses howeve r, yields were generally good in Cabo Delg ado. Rainfall was 254 near normal in the other months of Season 1. Rainfall was below normal each mon th 255 during Season 2, but a steady distribution allowed for good crop yields similar to Season 256 1 A bove normal during November and December 2004 allow ed for a good yield in the 257 first planting but fell to only 50% of normal by March. Maize and cassava are th e main 258 crops grown in Zone 2. 259 The 63 v illages in Zone 3 are located within 100 km of Malawi in the higher 260 altitude regions of Niassa (44%) Zambezia (17%) and Tete (38%) Provinces. The most 261 distinctive pattern for this zone was the contrast of Seasons 1 and 3 (Fig. 3) A pattern of 262


12 l ower than normal totals during the first two months and higher than normal during the 263 last three months for Season 1 was reversed in Season 3. Above normal rainfall occurred 264 in November of Season 2 followed by slightly below n ormal conditions for the rest of 265 this season. C rop yields were good for the first two years, but the irregular rains in 266 Season 3 favored crops that were planted early. Maize is the main crop planted. 267 The 104 villages receiving at least 150 mm of rainfall f rom Delfina comprise 268 Zone 4 and are located in Nampula (43%) and Zambezia (55%) It extend s from the 269 coastline to 350 km inland and covers approximately 109 000 sq. km, making it the third 270 largest zone. The wettest of all months in all zones was January 20 03 when an average of 271 550 m m of rainfall occurred during the passage of Delfina (Fig. 3) This represented more 272 than 200% of normal January rainfall. In no other month did the percentage of normal 273 rainfall surpass 140% and s erious drought was not reported during any of the study 274 months In fact, this region along with Zone 5 experienced the most normal of rainfall 275 patterns during Season 2 while other zones experienced monthly averages below 50% of 276 ther conditions to Zone 4, 277 where the majority 278 The cluster analysis determined that 125 villages within portions of Zambezia 279 (18%) Tete (23%) Manica (28%) and Sofala (31%) Pr ovinces had simila r rainfall 280 patterns. Zone 5 is approximately 280, 000 sq. km spanning from the coast of the 281 Mozambique Channel inland to the border with Zambia. Overall, the rainfall pattern here 282 was the most similar to normal of the nine zones. The wettest month at 168% o f normal 283 was March 2003 when some areas received rainfall from Japhet which the FAO reported 284 to be beneficial to crops 285


13 months T he lowlands experienced irregular rainfall pa tterns and lower crop yields in 286 Season 2 but r ainfall was more abundant in the interior highlands so that yields were 287 mixed across the zone Good yields were also reported in Season 3. The majority crop in 288 this region is maize, with these four provinces totaling about hal f of the maize planted 289 nation wide. 290 Households within the 30 villages that received at least 150 mm of rainfall from 291 Japhet comprise Zone 6 which straddles the southern portions of Manica (50%) and 292 Sofala (50%) Provinces. The second wettest month in the study was March 2003 when 293 Japhet produced an average of 4 00 mm of rainfall over the region, bringing the monthly 294 total to 230% of normal In no other month did rainfall occur that was more t han 110% of 295 normal within Zone 6 Early season droughts did occur during Seasons 1 and 2 when 296 rainfall was only 26% of normal in both Decembers Beneficial rains from Japhet 297 (Kadomura 2005; JTWC 2003) allowed for crops damaged by early season drought to be 298 replanted and eventually harvested. The current study does not i nclude any villages 299 located near the landfall location of Japhet, where strong winds were reported to cause 300 damage ( Kadomura 2005 ) A 301 planted within these two provinces each study year. 302 In Zone 7, 89% of the 45 villages are located within In hambane Prov ince and 303 most are within 30 km of the coastline where conditions are moister in general than 304 farther inland This zone has the highest percentage of villages located near the coast and 305 because the TRMM 3B43 product is known to underestimate rainfall along the coast 306 (Chen et al. 2013), rainfall totals could have been higher than our study shows. Although 307 drought was the major weather extreme within all three rainfall zones in southern 308


14 Mozambique during the study period Zone 7 was the least dry overall and received 309 normal rainfall in two months of Season 1. After receiving only 50 and 22% of normal 310 rainfall in November and December of Season 2, rainfall increased thereafter and the 311 FAO reports secondary harvests improved as some farmers planted cassava inste ad of 312 maize. In Season 3 the FAO noted that cassava production increased from previous 313 years. This province features a mixture of crops with maize, ground nuts, and cassava 314 comprising 30 % 22 % and 19% of the crops planted, respectively 315 Due to their ran ge of distances from the coastline, rainfall variability is higher 316 among the 36 villages of Zone 8 than Zone 7. The majority of villages (94%) in Zone 8 317 are located in Gaza Province. T he FAO reported severe drought and seeds lost during 318 Season 1, yet rainf all patterns were better suited for good yields in Season 2. Our study 319 shows that an average of 46% of normal rainfall was received during Season 1, which 320 increased to 74% in Season 2. During Season 3, f our dry months followed the good 321 rainfall in October with March being the only other month when rainfall was more than 322 56% above normal. The FAO predicted a 45% reduction in maize harvest for Season 3 323 which is the main crop planted in the zone. 324 Zone 9 is the smallest in the study covering approximately 13 500 sq. km, and all 34 325 villages are located with Maputo Province. Season 1 rainfall was less than 42% of normal 326 in three of the six study months and our finding that the lowest cumulative rainfall total 327 for Season 1 occurred here is confirmed by the FAO, which remarks that rainfall was the 328 lowest in 50 years resulting in a 50% reduction in maize crops from the previous year. 329 We find that an extremely dry October Decembe r was followed by a wet January and 330 March yielding a near normal cumulative rainfall for Season 2. Early maize crops failed, 331


15 but later yields were twice as good as during the previous year Although Season 3 332 rainfall was approximately 75% of normal for all months, the FAO reported that after 333 good rains in October, rainfall was erratic and not good for high crop yields throughout 334 the rest of the season. 335 b. Socio economic Characteristics 336 Now that we have described the nine distinctive rainfall patterns experienced across 337 Mozambique in the study period we examine socio economic characteristics a cross the 338 zones. We use longitudinal household level socio economic data from the National 339 Agricultural Survey of Mozambique (Trabalho de Inqurito Agrcola, TIA) to calculate 340 changes in annual crop income participation in different crop types, income sha res from 341 different crop types land area under cultivation, and other characteristics of households 342 over the 2002 2005 time period. 2 The survey sample is nationally representative of small 343 and medium scale farm households. 3 After eliminating extreme outlie rs and households 344 that did not receive any crop income in either 2002 or 2005, we use data from 3,859 345 households located in 536 villages across the country. 346 Using the TIA data set, Mather et al. (2008) and Cunguara (2008) do an extensive 347 analysis of chang ing income source components by administrative provinces and by 348 income quartiles. They find that income from staple food crop s (grains, pulses, roots, and 349 tubers) is particularly important for the most economically vulnerable households as it 350 accounts for 68% or more of the income for the poorest 20% of rural Mozambican 351 households. Since the vast majority of Mozambican small and medium scale farmers 352 2 The TIA surveys were also conducted in 2006, 2008 and 2012 but the same households were not interviewed, thus preventing our longitudinal study from including these data. 3 Small farms consist of less than 10 hectares, and medium scale farms consist of 10 50 hectares (World Bank 2006).


16 practice rainfed agriculture and very few use agricultural inputs such as agricultural 353 traction or irrigatio n ( Walker et al. 2004; Mather 2008 et al. ; Cunguara and Kelly 2009), 354 they are extremely vulnerable to bad weather. Their findings provide support for our 355 on crop income. 356 Since rural households acquire the bulk of their food from their own pro duction, 357 we calculate all crop income as the annual sum of all crop sales plus imputed income 358 from food retained and consumed by the household. Thus crop income includes the 359 retained and sold value of staple crops and sales from cash crops, vegetable s fru its 360 cashew, and coconuts 4 Costs of seeds fertilizers, and pesticides are subtracted from 361 gross crop income. In the TIA dataset, the prices used to value retained quantities of food 362 crops consumed by the household are the annual average retail price of e ach product from 363 the nearest rural retail market as reported by the Mozambican Sistema de Inforamao de 364 Mercados database (for more detail on the calculation of rural crop prices for 2002 and 365 2005 see Mather et al. 2008). The TIA uses the survey responden 366 sales from staple crops, cash crops, vegetable, fruit cashew, and coconut. 367 As evidenced by Table 1, median annual agricultural per capita incomes for rural 368 househo lds were extremely low in both 2002 and 2005 Our findings correspond to other 369 accounts of poverty in the country, which report that 75% of Mozambicans lived on less 370 than $1.25 per day in 2003 (World Bank 2013) These levels of poverty underscore the 371 need for rapid policy interventions to improve rural incomes and overall w ell being. In 372 addition to examining median annual household per capita income from all agricultural 373 sources, we also look at median income from staple crops, since most rural households 374 4 Table 1 provides a detailed accou nt of crops categorized as staples and cash crops in the TIA dataset. The full TIA survey instruments for 2002 and 2005 can be accessed at


17 derive the majority of their agricultural income from that source (Fig 4). W e find that 375 median income for both staples and all crops declined between 2002 and 2005 in every 376 area except Zone 6 (Japhet). Also, decreases in median staple crop income were either 377 greater or equal to reductions in income from all crop agriculture 378 Insert Table 1 379 Although southern Mozambique is widely acknowledged to be more 380 economically developed than other part s of the country, we find that median incomes 381 from staple crops or all crops did not follow any north south geographic pattern in either 382 year Zone s 4, 8 and 9 experienced the greatest declines in both types of agricultural 383 income and, by 2005, were the poorest in absolute terms according to both our measures. 384 Zone 9, our region with the most severe drought conditions, saw median agricultur al 385 incomes decline by more than 50%, and Zone 4 (Delfina) experienced decreases of only a 386 slightly lower magnitude. By 2005, Zone 8, which also experienced drought conditions, 387 had the lowest median incomes of all the zones. Zones 1 and 7 experienced the sm allest 388 declines in income from both staples and all crops, although some villages in Zone 1 389 experienced an extreme wet event (Atang) and Zone 7 was characterized by dr y 390 conditions ( although coastal locations may have been less dry ) Although Zone 6 (Japhet ) 391 was the only region with an absolute increase in median agricultural incomes according 392 to both measures, it was the poorest region, in absolute terms, at the start of the study 393 period. 394 Insert Figure 4 395


18 W e next report the breakdown of crop types by rainf all zone (unlike previous studies that 396 employed administrative districts ) to regionally summarize the composition of crop types 397 that comprise total crop income 398 Insert Table 2 399 Staple food crops make up the largest single income source for agricultural 400 inc ome in every zone in both 2002 ranging from 78 96%, and 2005 ranging from 71 401 90% (Fig. 4). With the exception of Zone 1, the percentage of in come from staple crops 402 decreased by an average of 7% Zone 8 experienced the largest decline (19%) in share of 403 s taple crop income. In general, household participation in staple crop production 404 decreased much less than agricultural income share (Table 2 ) For example, in Zone 4, 405 100% of households grew staple crops in 2002, and that number declined by only 1% in 406 2005 This rate of initital participation and decrease was similar to that seen in Zones 1, 407 2, 3, and 5. In contrast, Zones 8 experienced the greatest decrease in households growing 408 staple crops, dropping from 97% to 77%. Participation in staple crop agricultu re also 409 experienced substantial decreases in Zones 7 and 9, the other two regions that 410 experienced low rainfall, by 1 7 % and 1 2%, respectively 411 Increases in income share from agricultural sources that were not staple crops 412 tended to be very small in all zon es. For example, Zone 3 experienced the greatest 413 increase in cash crop income share, rising by 3 %. Most zones did experience a slight 414 increase in cash crop income share, with the average being 1.2%. Zone 2 had the highest 415 rate of household s participating i n cash crop production in both years (Table 2), and its 416 income share from this source was also highest. Zones 3, 4, and 5 also had relatively high 417 degree s of income coming from cash crops, although rates of household part icipation in 418


19 this activity fell by 2 4% in these zones. Zone 6 (Japhet) witnessed the greatest decrease 419 in households participating in cash crops (17%), but its income share from that source 420 changed by less than 1%. 421 Income share from vegetable s remained relatively steady in all zones, altho ugh 422 participation levels decreased in all but Zone 8. Vegetables accounted for less than 3% of 423 total agricultural income in any zone. Participation rates in fruit crops were relatively 424 similar across all zones, and income shares from this source changed le ss than 2%. One 425 exception was Zone 3, which had 34% of households participating in this activity in 426 2002, but saw the greatest participation decrease (12%). Fruit income share was highest 427 in Zones 6 in both years, and at least one fifth of households parti cipated in the activity. 428 There are two exceptions to the pattern of small changes in income composition 429 noted above O ne exception is the 17% increase in income from cashew groups in Zone 430 8 which largely offset the declines in staple crop income share an d contributing more 431 that 20% to overall agricultural income in 2005 Across the other zones, cashew crop 432 income share increased by 2.8%. Household participation in cashew production increased 433 the most in Zones 8 and 9, 12% and 10%, respectively. Zone 7 is unique in that more 434 than 16% of its agricultural income is derived from coconut in 2002 and 2005 while no 435 other zone exceeded 4%. The highest participation rates for coconut production were also 436 in Zone 7 in both years, 51% and 56% respectively, more than double those in any other 437 zone. 438 Taken together, our descriptive analysis illustrates that changes in household rates 439 of participation in crop types rarely correspond to changes in income shares from those 440 sources. Unfortunately, the TIA dataset does not p rovide information on the time 441


20 households devote to any particular activity, so it is difficult to determine if the 442 discrepancies between participation rates and income shares are the result of household 443 labor allocation. But the case of staple crops, wher e income shares generally declined 444 much more than household participation levels, suggests downward pressure on income 445 from these crops. Cunguara and Kelly (2009) find that staple food crop prices were stable 446 in Mozambique over the 2002 2005 time period, w ith the exception of maize which 447 experienced a slight price hike in 2005. This suggests that the income declines associated 448 with staple crop reliance are not due to a drop in prices and we hypothesize can be 449 attributed to weather related factors, something we explore in our regression analysis. 450 Insert Table 3 451 In terms of our demographic variables, the zones are generally similar by several 452 key measures (Table 3) Household size ranges roughly from 4 6 members and increased 453 slightly in all zones over the stu dy period. Per capita land under cultivation averaged 0 .70 454 hectares with relatively small changes in the zones over time. On average, households in 455 the three southern most zones (7, 8, and 9) have older heads, higher levels of education 456 and higher shares of female headed households These three zones, as well as Zone 1, 457 had lower dependency ratios, averaging 0 .72 as compared with 1.05 for the other zones. 458 Except for Zone 9, d ependency ratios increased slightly in every zone The percentage of 459 households us ing fertilizer or pesticides tended to be low with the exceptions of Zones 2 460 and 3, and changes were generally small. All zones experienced a substantial increase in 461 the share of households with at least one chronically ill member ranging from 10% in 462 Zon es 6 and 9 to 24% in Zone 2. 5 463 5 The very big differences in illness may be explained by a change in the wording of the questionnaire. In the 2002 TIA, the definition of serious illness was inability to work, while the question was phrased


21 464 3. Regression Analysis 465 In the following section we extend our analysis by using a regression approach to 466 examine the relationship between agricultural income chang e and complex rainfall 467 patterns. This method allows us to contro l for initial levels of wealth staple crop income 468 share, and other household characteristics in a way that is not possible through the 469 calculation of simple descriptive statistics. 470 The Mozambican government has increasingly encouraged small and medium 471 s cale farmers to produce export crops as a means of alleviating rural poverty (GOM 472 2001, 2006; World Bank 2006; IMF 2007). Development policies aim to increase food 473 security and foster economic growth by assisting all agricultural producers, ranging from 474 ag ribusinesses to small scale farmers, in export ing more higher value agricultural goods 475 (GOM 2006) 476 exports, that farmers relying less on staple crops and thus more on agricultural pro ducts 477 for which an international market exists would fare better economically. Thus w e 478 hypothesize that high er dependency on income from staple food crops in the initial year 479 of our study will be associated with declining total crop incomes. 480 We also hypot hesize that the differing economic outcomes for rural 481 agriculturalists will be related to rainfall patterns. More specifically, we predict 482 households experiencing extreme rainfall from a tropical cyclone that produces high 483 rainfall over a large area will h ave negative agricultural outcomes similar to those of 484 differently in the 2005 instrument. Thus a higher percentage of sick children was likely captured in the T IA 2005 survey.


22 households in regions experienc ing prolonged dry conditions. We expect this because 485 seasonal rainfall forecasts are regularly issued (Arndt and Bacou 2000) so that farmers 486 may anticipate drought condi tions. In contrast, rainfall from tropical cyclones is more 487 difficult to predict (Langousis and Veneziano 2009) and farmers would not have enough 488 lead times to adjust planting and harvesting schedules. In addition, tropical cyclones can 489 seriously damage cr ops as well as the infrastructure necessary for farmers to access 490 inputs or agricultural markets, as was the case with Cyclone Eline which contributed to 491 widespread flooding and damage in southern Mozambique in 2000 (Christie and Hanlon 492 2001). 493 a. Modeling Ap proach 494 We estimate a generalized linear regression model using a generalized es timating 495 equation (GEE) with an exchangeable working correlation matrix to explore the 496 relationship between changes in rural household per capita agricultural income and 497 rainfal l patterns This regression approach is commonly applied when using survey data 498 with stratified sampling designs where the observations are clustered ( Mathanga et al. 499 2010; Ghosh et al 2008; Cole 2001; Wen 2001 ; Zorn 2001 ). In the TIA data panel, 500 multiple households were surveyed within each of the sampled villages, resulting in 501 geographically clustered observations. However, even when clustered data observations 502 are correlated, the GEE estimation generates regression coefficient estimates and 503 empirically c orrected standard errors that account for clustering and are robust to 504 hereoskedasticity 6 This modeling approach allows us to pool all households in the 505 6 One limitation of using the GEE approach is that it is not possible to calculate an r 2 for the model (since the estimating procedure does not use a likelihood function). We also estimated our model via OLS and the signs and magnitude of r egression parameter coefficients were all similar to the results of our GEE


23 sample in a single regression and control for geographical effects, such as elevation, soil 506 types, and other unobservable characteristics that may be shared by households located in 507 the same village. We then test for the effects of distinct rainfall patterns using dummy 508 variables for each rainfall zone. 509 Table 4 includes details on the construction of each variable used in the analysis. 510 The dependent variable (lnY i ) is the nat ural log of annual household per capita change in 511 total crop income (including staple crops, cash crops vegetables, fruits, cashew, and 512 coconuts) between 2002 and 2005. 513 Insert Table 4 : Construction of Independent Variables 514 All independent variables are for the year 2002, thus measuring the initial conditions in 515 the household before the rainfall patterns occurred and avoiding issues with possible 516 endogeneity if measures of change are u sed as explanatory variables in the model We 517 include a variable (lnPCI i 02 ) measuring annual per capita household total crop income 518 logged, in the regression to control for initial levels of wealth from all crop related 519 sources We do this because poorer households will likely experience higher rates of 520 income change given the ir lower initial income base (i.e., even small increases in income 521 can result in high rates of change when initial incomes are very low). In contrast, 522 wealthier households will likely have lower rates of increase (due to their higher initial 523 income base). 524 estimation (although the significance levels were much higher, most likely resulting from correlation among clustered observations) Due to space constraints, we only report the re gression results from the GEE. Results from the OLS model are available from the lead author upon request.


24 Seven variables in our analysis measure agricultural characteristics of 525 households 7 SFC i 02 is percentage of total crop income derived from staple food crops 526 The majority of rural f armers earn ed some staple food crop income in 2002, allowing for 527 this measure to be constructed as a continuous variable. We also include five dummy 528 variables that measure whether or not a household received any income from cash crops 529 (DCC i 02 ), vegetable s (DV i 02 ), fruits (DF i 02 ), cashew (DCA i 02 ), or coconuts (DCO i 02 ). We 530 construct all five measures as dummy variables because of the limited number of 531 observations for crops other than staples and high correlation between different types of 532 agricultural produc tion. We include these crop variables to measure the effect of 533 agricultural income composition on changes in agricultural income over time. As stated 534 previously, w e hypothesize that high er dependency on income from staple food crops 535 will be associated with declining agricultural incomes for rural households In order to 536 control for the effects of farm size on income change, o ur model also includes a variable 537 (lnPCLand i 02 ), which measures the per capita land area cultivated by households, logged, 538 to control for the effect of farm size on income change. We expect a positive relationship 539 between farm size and crop income based on previous empirical studies (Mather 2012) 540 We also include demographic variables in the analysis that theory and previous 541 empirical evi dence suggest will impact agricultural income. These variables are age of 542 household head ( HAGE i 02 ); a dummy variable if the household includes one or more 543 chronically ill members ( DILL i 02 ), the household dependency ratio (HDR i 02 ) a dummy 544 variable for fema le headed households (D FEM i 02 ) and the maximum years of education 545 7 We omit any measures of agricultural input use, such as fertilizer or pesticide, because of high correlation with cash cropping. In addition, we om it any measure for use of animal traction and irrigation, which is largely confined to southern regions of the country where the disease burdens on livestock are lower and some irrigation infrastructure is present.


25 attained by any household member (EDU i 02 ) Older household heads, higher dependency 546 ratios, and chronic illness are expected to negatively affect income as these condition s 547 can constrain la bor endowments and reduce agricultural productivity. Households having 548 more educated members are expected to experience better economic outcomes, since 549 more schooling has been found to foster innovation among farmers in some rural African 550 societies (Knight et al. 2010) Female headed households should fare worse than those 551 headed by males. For example, Walker et al. (2004) found that households that are 552 headed by women are much poorer than are other Mozambican households, since women 553 do not have equal acces s to land and other resources. 554 W e include a series of categorical dummy variables to measure the effect of 555 household location in a particular rainfall zone to test our hypothesis that the unique 556 rainfall patterns had differing effects on household agricult ural income (DZ1 i DZ2 i 557 DZ3 i DZ4 i DZ5 i DZ6 i DZ7 i DZ8 i DZ9 i ). To avoid a situation with perfect multi 558 collinearity, one categorical dummy variable must be excluded from the regression when 559 an intercept is estimated. The omitted category becomes the r eference group against 560 which the effects of the other catego ries are assessed. The results of the paramete r 561 coefficients on the eight zone dummy variable included in the analysis are interpreted as 562 the expected difference in mean of household income change in that particular zone as 563 compared to the omitted category holding all other predi c tors in the regression constant 564 Although both droughts and floods can negatively impact agriculture, farmers may have 565 more time to adjust planting and harvesting schedul es when climate forecasts warn of 566 drought conditions. Thus, w e choose to omit a cyclone affected region rather than the 567 region most severely affected by the drought (Zone 9) Although three tropical cyclones 568


26 brought rainfall to Mozambique, w e chose to omi t Delfina (DZ4 i ) over Atang or Japhet 569 for two reasons. First, rainfall from Delfina was the most wide spread and intense of the 570 cyclone events Second, previous research determined that the some areas were reported 571 to benefit from the rainfall brought by J aphet (Kadomura 2005, Matyas and Silva 2013). 572 We expect households in all other zones to fare better in terms of agricultural income 573 change than those in the zone affected by Delfina. 574 b. Regression Results 575 The parameter coefficients and the statistical test r esults of independent variables 576 obtained from the regression analysis are presented in Table 5 8 The parameter coefficient 577 on household per capita agricultural income in 2002 (lnPCI i 02 ) is positive and significant 578 Thus, households with higher initial leve ls of agricultural income experience lower rates 579 of income change than initially poorer households 580 Insert Table 5 581 While variables measuring agricultural characteristics and regional rainfall 582 patterns are significant none of the parameter coefficients f or the demographic variables 583 are significant in the model. Regarding our agricultural variables, w e find a negative and 584 statistically significant relationship between income share from staple food crops ( SFC i 02 ) 585 and income change. This indicates the more r eliant a household was on staple food crop 586 income in 2002, the worse it fared economically over the study period supporting our 587 second hypothesis. The parameter coefficient on household per capita land area under 588 cultivation in 2002 ( lnPCLand i 02 ) is also positive and significant, suggesting that 589 8 The variables included in the analysis have correlation coefficients of less than +/ 0.35, with the majority being correlated at less than +/ 0.10. The one exception is our coconut crop dummy and Zone 7 which has a positive correlation of 0.40. However, the signs and significance of all para meter estimates do not change whether or not one variable is omitted, therefore we leave both in the model.


27 households that initially had larger farms or more agricultural plots had better economic 590 outcomes as measured by agricultural income change Of all the crop dummy variables, 591 only the parameter coefficient for hous eholds that engage in cash cropping in 2002 592 ( DCC i 02 ) is significant, and the relationship is positive. This indicates that a household 593 having some form of cash crop income in 2002 is associated with more positive 594 agricultural income change 595 Turning to o ur rainfall zone dummy variables we find that the parameter 596 coefficients on Zones 1 (DZ1 i ), 8 (DZ8 i ) and 9 (DZ9 i ) are insignificant. This suggests 597 that households in these rainfall zones after controlling for the agricultural and 598 demographic characterist ics included in the regression model, do no better in terms of 599 agricultural income change than those located in Zone 4, the area most impacted by 600 Delfina. Both Zone 8 and 9 as described earlier in the paper, experienced severe drought 601 during the study per iod particularly in Season 1 Our findings for Zones 8 and 9 also 602 comport with those of Mather et al ( 2008 ) who find that the number of drought days 603 during the agricultural growing season is associated with lower incomes for Mozambican 604 farmers and Cungu ara and Kelly (2009) who find that lower monthly rainfall totals 605 during the growing seasons have detrimental effects on agricultural income in rural 606 Mozambique. Importantly our results show that the impacts on crop income from 607 prolonged drought cannot be differentiated from those of cyclones producing wide spread 608 heavy rainfall Future studies need to consider farmer responses to cyclones as well as 609 drought to gain a better understanding of weather related effects on rural agriculturalists. 610 The parameter co efficient for Zone 1 was also insignificant T his area did not 611 experience prolonged dry periods and it is in the agricultural ly productive area of 612


28 northern Mozambique The descriptive statistics on median income (Table 1), show that 613 households experienced a 4% decrease in earnings from all crop sources, which is less 614 than in other zones. T his suggests that households in this region would fare better than 615 those impacted by Delfina. It also points to limitations in using descriptive statistics to 616 determine ec onomic performance across rainfall zones and illustrates the benefits of 617 using a regression based approach that allows us to for control for other variables in order 618 to isolate the effects of rainfall patterns. As discussed previously, several villages in Zone 619 1 received extreme rainfall from Atang in November 2002 and this could account for the 620 insignificant parameter coefficient, indicating that households in this area, when 621 controlling for other predictors in the model, did no better in terms of agricul tural income 622 than households in the Delfina region T o further investigate the impact of Atang on 623 economic performance in this region, w e correlated rainfall totals from Atang and 624 agricultural income change. W e find a significant, negative association betw een rainfall 625 totals during the two day event and change in household per capita agricultural income 626 r (203) = .14, p = .05 This 627 suggests that high rainfall from Atang had income decreasing effects for h ouseholds in 628 Zone 1. Given that Atang occurred in Season 1, and the subsequent seasons were 629 characterized by relatively normal rainfall patterns, our finding indicates that even such 630 localized extreme events may have enduring effects on rural agricultural households 631 T he parameter coefficients for Zone 2, 3, 5, 6, and 7 dummy variables (DZ2 i 632 DZ 3 i DZ 5 i DZ 6 i DZ 7 i ) w ere positive and significant, indicating that, after controlling 633 for the agricultural and demographic characteristics included in the regressi on model, 634 households in these rainfall zones experienced more positive agricultural income change 635


29 than households located in Zone 4 As previously discussed Zones 2, 3, and 5 did not 636 experience any extreme wet or dry events although they did exhibit some month to 637 month variability (Fig 3 ) Thus, the positive and significant coefficient on these 638 parameter estimates supports our hypothesis that regions with more stable weather would 639 be characterized by better economic performance in terms of crop income a fter 640 controlling for other predictors in the model, as compared to the omitted area that 641 experienced a tropical cyclone (Zone 4). 642 The findings for Zone s 6 and7 merit closer attention given that the rainfall 643 patterns they represent each show signs of extre mes. Zone 7 experienced drought 644 conditions in Seasons 1 and 2, yet they ended with normal or close to normal rainfall and 645 may have allowed some households to replant crops that failed earlier in the season This 646 comports with FAO reports ( 2003, 2004, 2005 ) that good rainfall at the end of the 647 growing season may help farmers offset crop losses from/recover from earlier bad 648 weather. In addition, many households in this zone are located near the coast and the 649 TRMM product tends to underestimate rainfall in coa stal locations (Chen et al. 2013), 650 thus some households could have received more rainfall that our analysis showed. T hese 651 explanations could account for the fact that households in this region did better relative to 652 those in Zone 4 while Zones 8 and 9 did not. Our findings also indicate that the economic 653 effects of low rainfall in coastal regions merits further study as better climate data 654 become available for Mozambique. 655 The positive and significant parameter coefficient for Zone 6 (Japhet) indicates 656 that households in this region did better relative to households affected by Delfina 657 suggest ing that rainfall from Japhet had positive effects on household agricultural 658


30 income. This finding comports with those of Matyas and Silva ( 2013 ) Kadoma (2005) 659 (both of which use TRMM data) and the FAO ( 2003 ) which all concluded that Japhet had 660 beneficial effects on rural agricultural households because it brought late season rai n to 661 areas that were previously experiencing drought. Moreover, Seasons 2 and 3 had no 662 extreme events and experienced better weather relative to other climate zones farther 663 south These factors may account for why positive effects from Japhet may still be 664 evident over two years later 665 T o further test whether Japhet was indeed a drought busting cyclone as suggested 666 by the significance of our zone dummy and the findings of previous empirical work we 667 correlate rainfall totals from days of the storm event and agricultural income change for 668 all 861 households that received more than one third of their March 2003 rainfall from 669 Japhet This enables us to investigate if Japhet rainfall benefited other nearby households 670 with absolute totals less than 150 mm We find that Japhet rainfall totals and change in all 671 crop income are positively correlated, r ( 859 ) = .24, p = (<.01). This provides further 672 support that Japhet was a drought busting cyclone for households in our study and 673 supports our interpretation of the Zone 6 dummy parameter coefficient However we find 674 that agricultural income became more unevenly distributed across households between 675 2002 and 2005, with a coefficient of variation of 23.4 and 30, respectively. This suggests 676 that some households in the region were better positioned to take advantage of the 677 benefits brought by Japhet 9 Although beyond the scope of this study, examining the 678 characteristics that account for the differential benefits of drought busting cyclones 679 merits future research. 680 681 9 We would like to thank an anonymous reviewer for calling attention to this point.


31 4. Policy Impl ications for Rural Development 682 policies promoting 683 greater production of agricultural exports by small and medium scale farmers (GOM 684 2001 2006 ). Examples of such policies include the provision of agricultural extension 685 services for farmers producing internationally exportable non staple crops (GOM 2001). 686 687 widely criticized in the literature and associated with increasing rural poverty (Hanlon 688 and Smart 2008; Cunguara and Hanlon 2010 ). Our findings indicate that farmers more 689 reliant on staple crops fare worse economically, after controlling for initial wealth levels, 690 distinctive rainfall patterns, and other agri cultural and demographic characteristics. 691 Moreover, we find evidence that households which earned some income from cash crops 692 at the beginning of the study period fared better than households that did not These 693 findings lend some support to the idea that rural households can improve their economic 694 position via the production of non staple crops. However, median household per capita 695 agricultural income declined in most parts of the country and is extremely low Thus 696 increasing the share of income from non s taple crop production does not guarantee that 697 households can adequately raise their incomes to a degree that improves standards of 698 living. 699 Our weather related findings suggest that households in cyclone impacted areas 700 generally have similar negative econo mic outcomes as those living in drought areas 701 which indicates that the majority of rural farmers operate in an extremely risky 702 environment. Some researchers argue that rural farmers are increasingly facing limits to 703 successful adaptation in the context of climate change ( Eakin and Luers 2006; Eriksen 704


32 and Silva 2009; Adger et al. 2009 ; Silva el at. 2010 ). We 705 find that median household incomes from all crops, and staple crops in particular, have 706 decreased in most areas and that re liance on staple crops has an income diminishing 707 effect. This suggests that in a country already characterized by endemic poverty, rural 708 households in this analysis may be confronting such limits. Given predictions that 709 extreme events are likely to increas e, the situation confronting farmers in countries like 710 Mozambique makes climate mitigation and adaptation research increasingly important 711 for poverty reduction policies. 712 The Mozambican case illustrates the difficulty in assessing whether development 713 policy should focus on strengthening the small farm agricultural sector or helping people 714 move into non farm economic activities. Tschirley and Benfica (2001) advocate a dopting 715 a mixed approach However, i t is difficult to argue against prioritizing policies aim ed at 716 enhancing the resilience and profitability of the agricultural sector. T he vast majority of 717 rural households in Mozambique still engage in crop production Policies that 718 successfully increase agricultural productivity through advances such as improve d 719 efficiency, enhanced insurance mechanisms, and better access to markets would have a 720 large economic impact because they are relevant to the majority of the rural population. 721 However, as climate predictions suggest that conditions for rainfed agricultural ists will 722 continue to worsen, policies that stop short of irrigation infrastructure and state sponsored 723 crop insurance will likely lose their effectiveness over time. 724 W e find households exposed to extreme wet events may experience comparable 725 declines in ag ricultural income as those in drought effect regions which calls attention to 726 the need for more research on the relationship between rainfall from tropical cyclones 727


33 and agricultural outcomes for farmers. Moreover, a lthough drought receives considerably 728 mo re attention in the rural development literature, strategies focused on raising 729 agricultural incomes also requ ire policies that mitigate risks associated with extreme wet 730 events. In the Mozambican context, the entire country is vulnerable to rainfall from 731 tropical cyclones. In addition, some of the most drought prone regions are in the south of 732 the country, an area characterized by higher levels of economic development and stronger 733 linkages to labor and commodity markets ( INGC UEM FEWS NET MIND 2003; Silva 734 2007 ). However, areas in the northern and central regions of the country may, in fact, be 735 the most in need of policies that foster an enabling environment for non farm start up 736 activities, improve abilities to trade, and increase mobility. More c ase studie s of how 737 households may take advantage of extreme events could help in the development of 738 policies to help farmers capitalize on the aftermath of disasters when food supplies 739 become limited and demand is high. Taken together our findings suggest that 740 devel opment strategies relying primarily on increasing agricultural income face serious 741 limitations in countries such as Mozambique, where endemic poverty and very low levels 742 of infrastructure make it hard for farmers to access or compete in international 743 commo dity markets. 744 745 5. Concluding Remarks 746 This study linked rainfall patterns to the economic well being of rural agriculturalists in 747 Mozambique, who rely on rainfed agriculture for food and income. We developed this 748 link by dividing the study region into rainfa ll zones based the percentage of normal 749 rainfall received during the growing season and from two tropical cyclones at each of 536 750


34 village s The rainfall analysis revealed that four of the nine rainfall zones experienced 751 extreme wet or dry conditions, while high monthly and interannual variability in rainfall 752 occurred throughout the country. W e then developed a regression model utilizing 753 measures of agricultural and demographic characteristics of households and controls for 754 distinctive rainfall patterns to i nvestigate the change in all crop income over the three 755 year study period. The agricultural variables attained a higher level of statistical 756 significance than the socio economic variables that are not associated with the weather, 757 and we find that rainfall patterns do have an important influence on economic well being 758 in Mozambique. 759 W e find that higher initial dependence on staple food crops within the study 760 period is generally associated with declining per capita agricultural incomes for rural 761 Mozambican h ouseholds with small and medium size farms. We found this relationship 762 regardless of whether or not households were in areas that had experienced extremely 763 high or low rainfall over the time frame of the study Our results also indicate that the 764 extreme w eather in Season 1 is strongly associated with declining agricultural income It 765 is important to note that t his is not only true for regions where all households are strongly 766 impacted by a tropical cyclone in Season 1 (Zone 4 Delfina ) but for those where there 767 are a relatively small number of cyclone affected households (Zone 1 Atang ) In 768 addition, regions characterized by dry events (Zones 8 and 9) fare no better than those 769 impacted by tropical cyclones. However, some climate factors that are unobservab le in 770 this study (e.g., coastal rainfall ) may improve agricultural outcomes as could be the case 771 with Zone 7. The evidence of l ingering negative effects of extreme weather in Zones 1, 4, 772 8, and 9 suggest that coping and adaptation strategies of rural agri culturalists may not be 773


35 sufficiently effective to halt the worsening economic position of farmers over time. The 774 lingering positive effects of Japhet suggest that some tropical cyclones are beneficial to 775 farmers particularly if these events are followed b y consecutive seasons of relatively 776 normal rainfall. 777 In terms of policy implications, this study suggests that a continued attention to 778 the rainfed staple food crop sector remains important to rural livelihoods in Mozambique. 779 Since reliance on staple crop income is associated with higher economic vulnerability, it 780 follows that better provi sion of support services and inputs, as well as enactment of 781 agricultu ral insurance schemes would enhance coping and adaptation to variable climate 782 conditions. Improving f 783 capacity of rural agriculturalists to mitigate the impacts of erratic or extreme weather. But 784 agricultural strategies alone are unlikely to halt the declining economic position of rural 785 agricult uralist s particularly given the climate predictions for the region. Thus, future 786 research should examine best practices for improving access to non agricultural types of 787 activities that may be a more effective in the long term in improving the well being of 788 rural households. 789 790


36 Acknowledgements 791 The authors acknowledge the United States Agency for International Development 792 (USAID)/Mozambique for their support of the TIA data collection and analysis in 793 Mozambique. We appreciate the comments from three anonymo us reviewers th at were 794 helpful in reorganizing the manuscript into its current form. [All other acknowledgements 795 removed for review.] 796


37 References 797 Adeyewa ZD, Nakamura K, 2003: Validation of TRMM radar rainfall data over major 798 climatic regions in Africa. J. Appl. Met ., 42 331 347. 799 Adger, M., S. Dessai, M. Goulden, M. Hulme, I. Lorenzoni, D.R. Nelson, L.O. Naess, J. 800 Wolf, and A. Wreford, 2009: Are there social limits to adaptation to climate 801 change? Clim. Change 93, 335 354. 802 Alemaw, B. F., and T. R. Chaoka, 2006: The 1950 1998 warm ENSO events and regional 803 implications to river flow variability in Southern Africa. Water SA 32, 459 463. 804 Arn d t, C. and M. Baou, 2000: Economy wide effects of climate variability and climate 805 prediction in Mozambique. Agric. Appl Econ. Assoc. 82, 750 754. 806 Ash, K. D., and C. J. Matyas, 2012: The influences of ENSO and the Subtropical Indian 807 Ocean Dipole on tropical cyclone trajectories in the South Indian Ocean. Int. J. 808 Climatol. 32, 46 51. 809 Central Intelligence Agency (CIA), 201 3: The World F actbook Entry on Mozambique. 810 Available at world 811 factbook/geos/mz.html 812 Chen, Y., E.E. Ebert, K.J.E. Walsh, and N.E. Davidson, 2013: Evaluation of TRMM 813 3B42 precipitation estimates of tropical cycl one rainfall using PACRAIN data J. 814 Geophs. Res.:Atmos. 118 2184 96. 815 Christie F., and J. Hanlon 2001 : Mozambique and the Great Flood of 2000 The 816 International African Institute, 176 pp. 817 Cole, S.R., 2001: Analysis of complex survey data using SAS Comp ut. Meth. Prog. Bio. 818 64, 65 9. 819


38 Cunguara, B., 2008: Pathways out of Poverty in Rural Mozambique. MS thesis, Michigan 820 State University. 821 Cunguara B. and B. Kelly, 2009: The impact of the PARPA II in promoting the 822 agricultural sector in rural Mozambique. Wo rking Paper. Ministry of Planning 823 and Development, Maputo Mozambique. 824 Cunguara, B., and J. Hanlon, 2010: Poverty is not being reduced in Mozambique. Crisis 825 States Research Centre working papers series 2, 74 Crisis States Research Centre, 826 London School of Economics and Political Science, London, UK. 827 Cunguara, B., and I. Darnhofer, 2011: Assessing the impact of improved agricultural 828 technologies on household income in rural Mozambique. Food Policy 36, 378 829 390. 830 Eakin, H. and A.L. Luers, 2006: Assessing the vulnerability of social environmental 831 systems. Annu. Rev. Environ. Resour ., 31, 364 394. 832 Easterling, W.E., P.K. Aggarwal, P. Batima, K.M. Brander, L. Erda, S.M. Howden, A. 833 Kirilenko, J. Morton, J. F. Soussana, J. Schmidhuber, and F.N. Tubiello, 2007: 834 Food, fibre and forest products. Climate Change 2007: Impacts, Adaptation and 835 Vulnerability M.L. Parry, O.F. Canziani, J.P. Palutikof, P.J van der Linden, and 836 C.E. Hanson, Eds., Contribution of Working Group II to the Fourth Assessment 837 Report of the Intergover nmental Panel on Climate Change. Cambridge University 838 Press, Cambridge, UK, 273 313. 839 Ellis, F. 2000: Rural Livelihoods and Diversity in Developing Countries Oxford: Oxford 840 University Press, 273 pp. 841


39 Ellis, F., and S. Biggs, 2001: Evolving themes in rural d evelopment 1950 2000s. Dev. 842 Policy Rev. 19, 437 448. 843 Eriksen S, and J.A. Silva, 2009: The vulnerability context of a savanna area in 844 Mozambique: household drought coping strategies and responses to economic 845 change. Environ. Sci. Policy 12, 35 52. 846 FAO (Fo od and Agriculture Organization of the United Nations), 2003: Special report 847 2003: FAO/WFP crop and food supply assessment mission to Mozambique. 848 Available at 849 FAO (Food and Agriculture Organization of the United Nations), 2004: Special report 850 2004: FAO/WFP crop and food supply assessment mission to Mozambique. 851 Available at 852 FAO (Food and Agriculture Organization of the United Nations), 2005: Special report 853 2005: FAO/WFP crop and food supply assessment mission to Mozambique. 854 Available at 855 Ghosh, S, P. Pahwa, and D.C. Rennie, 2008: Comparison of design based and model 856 based methods to estimate the variance us ing National Population Health Survey 857 data Model Assist. Stat. App ., 3, 33 42. 858 GOM (Government of Mozambique), 2001: Action Plan for the Reduction of Absolute 859 Poverty (2001 2005). Maputo, Mozambique. 860 GOM ( Government of Mozambique), 2006 : Action plan for the reduction of absolute 861 poverty (2006 2009). Maputo, Mozambique 862


40 Hahn, M.B., A.M. Riederer, and S.O. Foster, 2009: The Livelihood Vulnerability Index: 863 A pragmatic approach to assessing risks from climate variability and change A 864 case study in Mozambique Global Environ. Change 19, 74 88. 865 Hanlon, J., and T. Smart, 2008: Do bicycles equal development in Mozambique? 866 Woodbridge, Suffolk, UK: James Currey, 256 pp. 867 Hermele, K. 1992: Stick and carrot: Political alliances and nascent capitalism in 868 Mozambique. Authoritarianism, democracy and adjustment P. Gibbon, Y. 869 Bangura, and A. Ofstad, Ed., Scandinavian Institute of African Studies 169 86 870 IFPRI, 2002: Ending hunger in Africa: only the small farmer can do it. IFPRI 871 Washington DC. 872 IMF (International Moneta ry Fund), 2007: Republic of Mozambique: Poverty Reduction 873 Strategy Paper Joint Staff Advisory Note. Washington, DC, IMF. 874 INE (Instituto Nacional de Estatstica), 2008: Resultos preliminaries do censo 2007 [2007 875 Preliminary census results]. Maputo, Mozamb ique, INE. Available at 876 877 INGC UEM FEWS NET MIND (Instituto Nacional de Gesto de Calamidades 878 Universidade Eduardo Mondlane FEWS NET Mozambique Integrated 879 Information Network for Decision Making), 2003: Atlas For Disaster 880 Preparedness and Response in the Limpopo Basin. Maputo, Mozambique: INGC. 881 Available at 882 Isaacman, A., and B. Is aacman, 1983 : Mozambique: From Colonialism to R evolution, 883 1900 1982 Westview Press 235 pp 884


41 Isaacman, A., 1996 : Cotton is the Mother of Poverty Peasants, W ork, and Rural 885 Struggle in C olonial Mozambique, 1938 1961 Heinemann 272 pp 886 Joint Typhoon Warn ing Center (JTWC), 2003 : A nnual Tropical Cyclone R eport. Pearl 887 Harbor, Hawaii. 888 Jury, M. R., W. B. White, and C. J. C. Reason, 2004: Modelling the dominant climate 889 signals around southern Africa. Climate Dyn. 23, 717 726. 890 Kadomura, H., 2005: Climate anomal ies and extreme events in Africa in 2003. Afr. Study 891 Monogr. Suppl. Issue 30, 165 181. 892 Knight, J., S. Weirb, and T. Woldehannac, 2003: The role of education in facilitating 893 risk taking and innovation in agriculture. J. Dev Stud 39, 1 22. 894 Kruger, A. C., 2007: Trends in cloud cover from 1960 to 2005 over South Africa. Water 895 SA 33, 603 608. 896 Langousis, A., and D. Veneziano, 2009: Theoretical model of rainfall in tropical cyclones 897 for the assessment of long term risk. J. Geophys. Res. Atmos. 114, 11. 898 Leich Double Exposure: Global Environmental Change 899 in an Era of Globalization Oxford University Press, 192 pp. 900 MADER (Ministrio da Agricultura e Desenvolvimento Rural), 2002: Trabalho de 901 inqurito agrcola 2002 (National agricult ure survey 2002). MADER, Maputo, 902 Mozambique. 903 MADER (Ministrio da Agricultura e Desenvolvimento Rural), 2005: Trabalho de 904 inqurito agrcola 2005 (National agriculture survey 2005). MADER, Maputo, 905 Mozambique. 906


42 Manhique, A. J., C. J. C. Reason, L. Rydberg, a nd N. Fauchereau, 2011: ENSO and 907 Indian Ocean sea surface temperatures and their relationships with tropical 908 temperate troughs over Mozambique and the Southwest Indian Ocean. Int. J. 909 Climatol. 31, 1 13. 910 Mathanga, D.P., C.H. Campbell, J.V. Eng, A. Wolkon, R.N. Bronzan, G.J. Malenga1, D. 911 Ali and M. Desai 2010: Comparison of anaemia and parasitaemia as indicators of 912 malaria control in household and EPI health facility surveys in Malawi Malaria 913 J. 9( 107). 914 Mather D ., D. Cunguara, and D. Boughton, 2008: Hous ehold income and assets in rural 915 Mozambique, 2002 2005: can pro poor growth be sustained? Directorate of 916 Economics Research Working Paper No. 66. Ministry of Agriculture, Maputo, 917 Mozambique. Available at 918 Ma ther, D., 2012: Determinants of Crop Income in Rural Mozambique, 2002 5. 919 Directorate of Economics Research Report 71E. Ministry of Agriculture, Maputo, 920 Mozambique. 921 Matyas, C. J., and J. A. Silva, 2013: Extreme weather and economic well being in rural 922 Mozam bique. Nat. Hazards 66, 31 49. 923 Molua, E.L., 2002: Climate variability, vulnerability and effectiveness of farm level 924 adaptation options: the challenges and implications for food security in 925 Southwestern Cameroon. Environ. Dev. Econ. 7, 529 545. 926 Parry, M. L., C. Rosenzweig, A. Iglesias, M. Livermore, and G. Fischer, 2004: Effects of 927 climate change on global food production under SRES emissions and socio 928 economic scenarios. Global Environ. Change 14, 53 67. 929


43 Patt, A.G., and D. Schrter, 2008: Perceptions of climate risk in Mozambique: 930 Implications for the success of adaptation strategies. Global Environ. Change 18, 931 458 467. 932 Pitcher, M.A., 2002 : Transforming Mozambique: The politics of privatization, 1975 933 2000 Cambridge University Press 320 pp 934 Reason, C. J C., and R. F. Phaladi, 2005: Evolution of the 2002 2004 drought over 935 northern South Africa and potential forcing mechanisms. S. Afr. J. Sci. 101, 544 936 552. 937 Reason, C. J. C., S. Hachigonta, and R. F. Phaladi, 2005: Interannual variability in rainy 938 season characteristics over the Limpopo region of southern Africa. Int. J. 939 Climatol. 25, 1835 1853. 940 Rosenzweig, C., and M.L. Parry, 1994: Potential impact of climate change on world food 941 supply. Nature 367, 133 138. 942 Rouault, M., and Y. Richard, 2005: Intensity and spatial extent of droughts in southern 943 Africa. Geophys. Res. Lett. 32, L15702, DOI: 10.1029/2005GL022436. 944 Sheldon, K.E., 2002: Pounders of grain: A history of women, work, and politics in 945 Mozambique Heinemann, Portsmouth, NH, 311pp. 946 Shongwe, M. E., G J. van Oldenborgh, B. van den Hurk, B. de Boer, C. A. S. Coelho, 947 and M. K. van Aalst, 2009: Projected changes in mean and extreme precipitation 948 in Africa under global warming. Part I: southern Africa. J. Climate 22, 3819 949 3837. 950


44 Sietz, D., M. Boschutz, an d R.J.T Klein, 2011: Mainstreaming climate adaptation into 951 development assistance: Rationale, institutional barriers and opportunities in 952 Mozambique. Environ. Sci. Policy 14, 493 502. 953 Silva, J.A., 2007: Trade and income inequality in a less developed coun try: The case of 954 Mozambique. Econ. Geogr. 83, 111 136. 955 Silva, J.A., 2008: A multilevel analysis of agricultural trade and socioeconomic 956 inequality in rural Mozambique. Prof. Geogr ., 60, 174 189. 957 Silva, J.A., S. Eriksen, and Z.A. Ombe, 2010: Double exposure 958 Limpopo River Basin. Geogr. J. 176, 6 24. 959 Thomas D.S.G., C. Twyman, H. Osbahr, and B. Hewitson, 2007: Adaptation to climate 960 change and variability: farmer responses to intra seasonal precipitation trends in 961 South Africa. Climate Change 8 3, 301 322. 962 Todd, M., and R. Washington, 1999: Circulation anomalies associated with tropical 963 temperate troughs in southern Africa and the south west Indian Ocean. Climate 964 Dyn. 15, 937 951. 965 Tschirley, D. and R. Benfica, 2001: Smallholder agriculture, wage labour and rural 966 poverty alleviation in land abundant areas of Africa: evidence from Mozambique. 967 J. Mod. Afr. Stud. 39, 333 358. 968 UNDP (United Nations Development Programme), 2011: Human Development Report 969 2011 Sustainability and Equity: A Better Future f or All. New York: Palgrave 970 Macmillan. Available at 971 972


45 Usman, M. T., and C. J. C. Reason, 2004: Dry spell frequencies and their variability over 973 southern Africa. Climate Res. 26, 199 211. 974 Walker T., D.L. Tschirley, J. Low, M. Pequenino Tanque, D. Boughton, E. 975 Payongayong, E., and M.T. Weber, 2004: Determinants of rural income, poverty, 976 and perceived well being in Mozambique in 2001 2002 Mi nistry of Agriculture 977 and Rural Development, Economics Directorate Maputo, Mozambique:. 978 Washington, R., and A. Preston, 2006: Extreme wet years over southern Africa: Role of 979 Indian Ocean sea surface temperatures. J. Geophs. Res.:Atmos. 111, DOI: 980 10.1029/ 2005jd006724. 981 Wen M. J., and L. Yeh, 2001 : Using SAS/GENMOD procedure to fit GEE regression 982 m odels Technical Report Number 42, National Cheng Kung University Taiwan. 983 World Bank, 2006: Mozambique Agricultural Development Strategy: Stimulating 984 Smallholder Agricultural Growth. Report No. 32416 MZ. World Bank 985 Agriculture, Environment, and Social Development Unit,Washington, D.C. 986 World Bank 2013. World DataBank Available at 987 988 Wuyts, M., 1996 : Foreign aid, structur al adjustment, and public management: The 989 Mozambican experience. Dev. Change, 27 717 749. 990 991 From conflict to reconstruction in Africa T. Addison, Ed., Oxford University 992 P ress, London, 141 154. 993 Zorn, C.J.W., 2001: Generalized Estimating Equation Models for Correlated Data: A 994 Review with Applications Amer. J. Polit. Sci ., 45, 470 90 995


46 Tables 996 Table 1: Median Annual Household Per Capita Incomes (in 2002 USD) from Staple Food Crops and All Crop Agriculture 1 997 Rainfall Zones 1 2 3 4 5 6 7 8 9 (n=205) (n=435) (n=455) (n=711) (n=882) (n=196) (n=347) (n=355) (n=283) Median household per capita income from staple crops: 2 In 2002 (USD) 3 38.41 35.55 65.91 52.35 32.95 2 9.46 46.97 34.22 59.12 In 2005 (USD) 4 35.52 30.94 43.29 23.52 25.93 36.15 39.15 17.89 23.94 per capita income 5 3% 23% 23% 53% 23% 8% 12% 38% 58% Median household per capita income from all crop agriculture: 6 In 2002 (USD) 3 45 .98 47.62 73.70 60.17 37.78 33.17 63.88 36.81 60.71 In 2005 (USD) 4 37.04 46.01 54.59 29.97 31.60 43.88 56.80 24.30 26.26 per capita income 5 4% 18% 23% 48% 17% 28% 5% 26% 54% 998 Notes : 1 Calculations include all 3,859 households in the analysis that reported agricultural income for either 2002 or 2005. 999 2 Total annual staple crop income households is calculated by summing all earnings from staple food crops minus the cost of seeds and fertilizers. 1000 Staple food crops consist of maize, rice, sorghum, millet, peanuts, butter beans, cowpeas, Bambara groundnuts, pigeon beans, Irish potatoes, 1001 cassava, and sweet potato. The total staple crop income is then divided by the number of household members to construct per capita values. 1002 3 Mozambican New Meticais (MTN) in 2002 converted into 2002 USD, adjusted for purchasing power parity (PPP), using the World Bank PPP 1003 conversion factor for private household consumption. 1004 4 Mozambican New Meticais (MTN) in 2005 first converted to 1005 are then converted into 2002 USD, adjusted for purchasing power parity (PPP), using the World Bank PPP conversion factor for private 1006 household consumption. 1007


47 5 We rep ort median percentage change in household per capita annual incomes. 1008 6 Total crop income is calculated by adding the sum of crop sales plus imputed income from food retained and consumed by the ho usehold. Thus 1009 crop income includes the retained a nd sold value of staple crops, and sales from cash crops (i.e., cotton, tobacco, tea, sisal, soybeans, paprika, 1010 sunflower and sesame seeds, and ginger), vegetables, cashew, and coconut. Costs of seeds, fertizers, and pesticides are subtracted from gross 1011 cr op income. Total crop income is then divided by the number of household members to construct per capita values. 1012 1013


48 Table 2: Percentage of Households with Income from Each Crop Source in 2002 and 2005. 1014 Rainfall Zones 1 (n=205) 2 (n=435) 3 (n=455) 4 (n= 711) 5 (n=882) 6 (n=196) 7 (n=347) 8 (n=355) 9 (n=283) Staple crops in 2002 in 2005 Difference 99% 96% 2% 99% 97% 1% 99% 98% 1% 100% 98% 1% 98% 95% 3% 99% 91% 8% 99% 82% 17% 97% 77% 20% 97% 84% 12% Cash crops in 2002 in 20 05 Difference 10% 5% 5% 42% 35% 7% 26% 24% 2% 18% 16% 2% 21% 17% 4% 27% 10% 17% 2% 2% 0% 1% 3% 2% 2% 7% 5% Vegetables in 2002 in 2005 Difference 14% 7% 7% 14% 5% 9% 28% 8% 20% 13% 8% 5% 21% 8% 14% 16% 7% 9% 10% 6 % 4% 8% 9% 2% 16% 12% 4% Fruit in 2002 in 2005 Difference 16% 17% 1% 16% 17% 1% 34% 22% 12% 17% 20% 3% 19% 12% 7% 23% 20% 3% 19% 17% 3% 10% 13% 3% 16% 6% 10% Cashew in 2002 in 2005 Difference 24% 27% 2% 35% 30% 5% 1 % 0 % 0 % 37% 44% 7% 7% 7% 1% 42% 54% 11% 54% 59% 5% 41% 52% 12% 15% 25% 10%


49 Coconut in 2002 in 2005 Difference 19% 23% 5% 5% 6% 1% 0% 0% 0% 11% 13% 1% 7% 8% 1% 8% 11% 3% 51% 56% 5% 8% 9% 1% 1% 0% 0% Notes : 1 Calculations inclu de all 3,859 households in the analysis that reported agricultural income for either 2002 or 2005 1015 1016


50 Table 3 : Mean and Percentage Values for Demographic Characteristics of Households in 2002 and 2005 1, 2 1017 1018 Rainfall Zone 1 (n=205) 2 (n=425) 3 (n=455) 4 (n=711 ) 5 (n=882) 6 (n=196) 7 (n=347) 8 (n=355) 9 (n=283) Year 02 05 02 05 02 05 02 05 02 05 02 05 02 05 02 05 02 05 Household Size 4.23 4.78 4.59 4.97 4.98 5.39 4.49 4.87 5.39 5.86 6.04 6.51 5.45 5.85 5.76 6.15 5.55 5.85 0.19 0.19 0.11 0 .12 0.13 0.16 0.08 0.09 0.12 0.15 0.31 0.30 0.22 0.27 0.25 0.32 0.24 0.27 Household 43.3 46.4 39.6 42.6 41.1 44.9 39.2 42.4 40.9 43.3 45.9 49.0 49.4 49.9 49.6 52.4 47.1 49.9 1.52 1.34 0.71 0.66 0.92 0.80 0.56 0.56 0.62 0.59 1.42 1 .40 1.05 1.14 1.11 1.09 1.25 1.14 Dependency Ratio 3 0.72 0.73 1.08 1.10 1.04 1.09 0.99 1.09 1.03 1.02 0.98 1.11 0.77 0.86 0.64 0.73 0.77 0.73 0.07 0.06 0.04 0.04 0.05 0.06 0.03 0.04 0.04 0.03 0.08 0.06 0.05 0.06 0.05 0.05 0.05 0 .05 Maximum Education 4 2.37 3.31 3.09 3.75 2.99 3.81 3.12 3.77 3.70 4.56 3.45 4.09 4.07 5.26 3.86 4.65 4.30 5.26 0.23 0.25 0.16 0.17 0.18 0.19 0.13 0.14 0.14 0.17 0.24 0.28 0.25 0.28 0.22 0.25 0.28 0.28 Per capita cultivated land ( ha ) 0.75 0.66 0.63 0.71 0.71 0.75 0.70 0.66 0.68 0.62 0.65 0.74 0.71 0.67 0.80 0.77 0.70 0.67 0.07 0.03 0.03 0.04 0.03 0.03 0.05 0.02 0.04 0.03 0.04 0.05 0.05 0.05 0.06 0.05 0.08 0.05 % Female headed 21% 23% 27% 26% 29% 28% 18% 20% 23% 28% 17% 25% 29% 40% 32% 37% 35% 40% % Chronic Illness 5 4% 19% 3% 27% 5% 24% 6% 23% 6% 17% 6% 15% 8% 27% 10% 23% 5% 14% % Fertilizer or Pesticides 0% 0% 18 % 19% 14% 22% 8% 4% 3% 3% 0% 0% 2% 2% 5% 4% 1% 4%


51 Notes: 1 Source surveys ( MADER 2002, 2005) 2 Standard errors are presented below the means in italics 3 The dependency ratio in this analysi s is calculated as the number of children under the age of 15 divided by the number of household members over the age of 15. We do not include the elderly (above age 64) as dependents in our measure. In Mozambique, due in part to the HIV/AIDS epidemic, the very elderly often head some households where there are no working age adults. 4 The number of years of schooling of the most educated member of the household. 5 C hronic illness is defined as a household with one or m ore members (of any age) reported to be suffering from serious illness at the time of the survey or have suffered a serious illness for at least 3 of the preceding 12 months. The survey does not specify the type of illness. 1020


52 52 Table 4: Construction of D ependent and Independent Variables for Regression Analysis Variable Definition Dependent: Change in the annual household per capita total crop income between 2002 2005 (lnY i ) The change in total annual crop income is calculated by summing all earnings from staple food crops 1 (sales plus value of produce retained for home consumption), and sales of cash crops 2 vegetable, fruit, cashew, and coconut in 2002 and 2005. Costs of seeds, pesticides, and fertilizer are subtracted from the total. The 2005 incom es are converted to 2002 values cs for Mozambique. The total household income is then divided by the number of household members to construct per capita values for each year. We take the natural log of our income figures t o normalize the distribution. The change variable is the difference between the natural log of per capita household annual agricultural crop income in 2005 and 2002. 3, 4 Independents: Household p er capita total crop income (lnPCI i 02 ) Total annual crop income for households is calculated by summing all earnings from staple food crops 1 (sales plus value of produce retained for home consumption), and sales of cash crops 2 vegetables, fruit, cashew, and coconut in 2002. Costs of seeds, pesticides, and fert ilizer are subtracted from the total. The total household income is then divided by the number of household members to construct per capita values 3 Share of total crop income from staple food crops (SFC i 02 ) crop income derived from staple crops (sales plus value of produced retained for home consumption) in 2002. 1, 2 Household p er capita land area under cultivation (lnPCLand i 02 ) The total annual land area (measured in hectares) under cultivation by a household, divided by the number of household members to construct per capita values in 2002. We take the natural log of our land area figures to normalize the distribution. 3 1021


53 53 1022 Crop income dummy variables: Cash crops (DCC i 02 ) Vegetables (DV i 02 ) Frui ts (DF i 02 ) Cashew (DCA i 02 ) Coconut (DCO i 02 ) The dummy takes a value of 1 if a household reported any income from that crop type in 2002, and 0 otherwise. We constructed dummy variables for cash crops, vegetables, fruits, cashew, and coconut as th ese are the six agricultural categories specified in the TIA data. 3 Age of household head (HAGE i 02 ) 3 Chronically ill household members dummy (DILL i 02 ) The dummy takes a value of 1 if the household has one or more members (of any age) suffered from serious illness either at the time of the 2002 survey or for at least 3 of the preceding 12 months, and 0 otherwise. The survey does not specify the type of illness. 3 Dependency ratio (HDR i 02 ) The number of childr en under the age of 15 divided by the number of household members over 15 in 2002. The elderly (above age 64) are not considered dependents in our measure. In Mozambique, due in part to the HIV/AIDS epidemic, the very elderly often head some households whe re there are no working age adults. 3 Female headed household dummy (DFEM i 02 ) Maximum educational attainment (EDU i 02 ) The dummy takes a value of 1 if the household is headed by a female in 2002, and 0 otherwise. 3 The highest number of years of educat ion of the most educated household member in 2002. 3 Rainfall zone dummy variables i, i ) The dummy takes a value of 1 if a household is located in the rainfall zone, and 0 otherwise. We construct dummy variables for all nine rain fall zones. The geographic coordinates of the village location of households is included in the TIA 2002 survey. 3, 5


54 54 Notes: 1 Staple food crops consist of maize, rice, sorghum, millet, peanuts, butter beans, cowpeas, bambara groundnuts, pigeon be ans, Irish potatoes, cassava, and sweet potato. 2 Cash crops consist of cotton, tobacco, tea, sunflowers, sesame, sisal, soybeans, paprika, and ginger. 3 Source: Authors' calculations using data from 2002 TIA (MADER 20 02). All values are for the 12 months preceding the administration of the survey. The survey was administered before the 2002 2003 growing season. 4 Source : using data from 2005 TIA (MADER 2005). All values are for the 12 months preceding the administration of the survey. The survey was administered after the 2004 2005 growing season. 5 Source : and geographic coor dinates of villages from the TIA 2002 survey (MADER 2002) 1023


55 55 Table 5: Regression Results 1 1024 Dependent: Change in annual household per capita total crop income between 2002 2005, logged (lnY i ) Estimate Z Score Intercept 4.4698 *** 19.58 Annual h ous ehold per capita total crop income, logged (lnPCI i 02 ) 0.6920 *** 19.60 Share of total crop income from staple food crops (SFC i 02 ) 0.6429 *** 3.69 Household per capita land area under cultivation, logged (lnPCLand i 02 ) 0.3803 *** 5.83 Cash crops (DCC i 02 ) 0.1350 ** 1.76 Vegetables (DV i 02 ) 0.0124 *** 0.18 Fruits (DF i 02 ) 0.0162 *** 0.24 Cashew (DCA i 02 ) 0.0874 *** 1.33 Coconut (DCO i 02 ) 0.1360 *** 1.43 Age of household head (HAGE i 02 ) 0.0018 *** 1.01 Chronically ill household member(s) (DILL i 02 ) 0.138 4 *** 1.33 Dependency ratio (HDR i 02 ) 0.0046 *** 0.13 Female headed household (DFEM i 02 ) 0.0565 *** 0.95 Maximum educational attainment (EDU i 02 ) 0.0004 *** 0.04 Zone 1 (DZ1 i ) 0.1217 *** 0.70 Zone 2 (DZ2 i ) 0.2263 ** 2.10 Zone 3 (DZ3 i ) 0.6144 *** 5.61 Zone 5 (DZ5 i ) 0.2071 ** 2.00 Zone 6 (DZ6 i ) 0.3390 ** 1.98 Zone 7 (DZ7 i ) 0.5654*** 3.31 Zone 8 (DZ8 i ) 0.2350 *** 1.13 Zone 9 (DZ9 i ) 0.3358 *** 1.36 significant at p < .10*, p < .05**; p < .01*** (2 tailed) 1025 1 The regression is estimated as a gener al linear model using a generalized estimating equation with 3,859 1026 observations. 1027


56 56 Figure Caption List 1028 Figure 1 : Map of Mozambique with v illage locations in nine rainf all zones 1029 1030 Figure 2: Rainfall data depicting a) average rainfall received during the main growing 1031 season months of November March, b) proportion of normal rainfall for Season 1, c) 1032 proportion of normal rainfall for Season 2, and c) proportion of normal rainfall for 1033 Season 3 1034 1035 Figure 3: Boxplots summarizing the variation in percentage of normal monthly rainfall 1036 November March for each rainfall zone in a) Season 1, b) Season 2, c) Season 3 1037 1038 Figure 4: Composition of total agricultural income by source in 2002 and 2005 across 1039 nine rainfall zones 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050


57 57 1051 Figures 1052 1053 Figure 1 : Map of Mozambique wi th v illage locations in nine rainf all zones 1054 1055 1056


58 58 1057 1058 1059 Figure 2: Rainfall data depicting a) average rainfall received during the main 1060 growing season months of November March, b) proportion of normal rainfall for 1061 Season 1, c) proportion of normal rainfall for Season 2, and c) proportion of normal 1062 rainfall for Season 3 1063


59 59 1064 1065 Figure 3: Boxplots summarizing the variation in percentage of normal monthly rainfall November March for each rainfall 1066 zone in a) Season 1, b) Season 2, c) Season 3 1067


60 60 1068 1069 1070 1071 Figure 4: Compositi on of total agricultural income by source in 2002 and 2005 across nine rainfall zones 1072