Spatial analysis of poverty in Ghana - World Bank



Spatial analysis of poverty: Combining geospatial data and survey data to study regional inequality in GhanaTomomi Tanaka (World Bank)Jia Jun Lee (World Bank)AbstractThis study combines district level poverty rates, population census data, income data, and geospatial data to investigate how human capital, structural change, infrastructure, and environmental degradation impacted poverty and employment in Ghana. We find that poverty reduction was primarily achieved through increased share of working age population, employment rates and income in the service sector, shift of labor from agriculture to the service sector, expansion of access to electricity, and increased rainfall. Further, the paper investigates the factors that have affected changes in agricultural income and shift of labor from agriculture to industry and services. Soil erosion had a large impact on the changes in agricultural income. Improved access to electricity, road, and market was crucial for creating jobs in non-agricultural sectors. In areas where droughts are severe, more people became engaged in agriculture and less in industry and services. It may be because severe droughts prevent people from investing in non-agricultural sectors. The results of this study suggest that for Ghana to reduce poverty and create jobs in non-agricultural sectors in lagging areas, it needs to invest in infrastructure, and take actions to mitigate damages from soil degradation and droughts.Introduction Ghana has achieved significant poverty reduction since the 1990s and accomplished the first Millennium Development Goal (MDG) of reducing the poverty rate by more than half. Between 1991 and 2012, the national poverty rate declined from 52.7 percent to 24.2 percent. The national poverty rate further declined to 23.4 percent by 2016. Starting from higher than the mean for low- and middle-income countries (LMICs), Ghana’s international poverty headcount today is lower than the current LMIC average. Even though the poverty rate fell at the national level, spatial inequality intensified with some regions lagging. Regional poverty rates had fallen dramatically in the Ashanti, Eastern, Greater Accra, Brong Ahafo, Central, and Western regions (Figure 1), while poverty rates remain above 50 percent in the Northern, Upper West, and Upper East regions. In addition, the pace of poverty reduction has slowed down in the Volta region.Poverty rates vary not only across regions but also across districts within regions. As Figure 2 shows, poverty rates were generally high in most districts in the three northern regions (Northern, Upper West, and Upper East regions) in 2000. However, the spatial distribution of poverty has changed dramatically between 2000 and 2010. The eastern part of the three Northern regions achieved significant poverty reduction, while several districts in the western side of the country witnessed rising poverty. Clusters of districts with high poverty rates also developed in the inland parts of the country and in the Volta region. A key question of this study is how some districts were able to reduce poverty, while others were not. We investigate how human capital development, structural change, infrastructure, and environmental degradation affected poverty, income and employment in Ghana. Human capital development is critical for poverty reduction as educated and healthy labor force is central to wealth creation and productivity growth ADDIN EN.CITE <EndNote><Cite><Author>Schultz</Author><Year>1961</Year><RecNum>51</RecNum><DisplayText>(Schultz 1961)</DisplayText><record><rec-number>51</rec-number><foreign-keys><key app="EN" db-id="2wfd5rtfos2e0qe2998vx90isetx99ew9ppp" timestamp="1545127531">51</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Schultz, Theodore W</author></authors></contributors><titles><title>Investment in human capital</title><secondary-title>The American economic review</secondary-title></titles><periodical><full-title>The American Economic Review</full-title></periodical><pages>1-17</pages><volume>51</volume><number>1</number><dates><year>1961</year></dates><isbn>0002-8282</isbn><urls></urls></record></Cite></EndNote>(Schultz 1961). A higher share of working age population in the economy helps the country achieve economic development and poverty reduction, since there are more people available for productive activities. Investment on education is a plausible way to escape poverty, as education is often a prime means of social mobility. Reduction of fertility helps families to invest more on each of their children. For these reasons we select a share of working age population, educational attainments among working age population, and fertility as key variables for the anlysis. Even when there is a higher share of working age population, economic development is difficult if there are not enough jobs in productive sectors. ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>McMillan</Author><Year>2014</Year><RecNum>88</RecNum><DisplayText>McMillan and Harttgen (2014)</DisplayText><record><rec-number>88</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1530242309">88</key></foreign-keys><ref-type name="Report">27</ref-type><contributors><authors><author>McMillan, Margaret S</author><author>Harttgen, Kenneth</author></authors></contributors><titles><title>What is driving the&apos;African Growth Miracle&apos;?</title></titles><dates><year>2014</year></dates><publisher>National Bureau of Economic Research</publisher><urls></urls></record></Cite></EndNote>McMillan and Harttgen (2014) and ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Diao</Author><Year>2017</Year><RecNum>55</RecNum><DisplayText>Diao, Harttgen, and McMillan (2017)</DisplayText><record><rec-number>55</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1530238166">55</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Diao, Xinshen</author><author>Harttgen, Kenneth</author><author>McMillan, Margaret</author></authors></contributors><titles><title>The Changing Structure of Africa’s Economies</title><secondary-title>The World Bank Economic Review</secondary-title></titles><periodical><full-title>The World Bank Economic Review</full-title></periodical><pages>412-433</pages><volume>31</volume><number>2</number><dates><year>2017</year></dates><isbn>0258-6770</isbn><urls></urls></record></Cite></EndNote>Diao, Harttgen, and McMillan (2017) show much of Africa’s recent growth and poverty reduction has been associated with structural change; a substantive decline in the share of the labor force engaged in agriculture. In Ghana, the share of employment in agriculture fell from 50 percent to 42 percent between 2000 and 2010, while the share of employment in services rose from 34 to 42 percent during that period. Meanwhile, the employment share in industry slightly declined. Ghana’s pattern of structural change is consistent with the general observation in Sub-Saharan Africa, i.e. premature deindustrialization ADDIN EN.CITE <EndNote><Cite><Author>Rodrik</Author><Year>2016</Year><RecNum>60</RecNum><DisplayText>(Rodrik 2016)</DisplayText><record><rec-number>60</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1530238836">60</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Rodrik, Dani</author></authors></contributors><titles><title>Premature deindustrialization</title><secondary-title>Journal of Economic Growth</secondary-title></titles><periodical><full-title>Journal of Economic Growth</full-title></periodical><pages>1-33</pages><volume>21</volume><number>1</number><dates><year>2016</year></dates><isbn>1381-4338</isbn><urls></urls></record></Cite></EndNote>(Rodrik 2016). In Africa, workers mainly relocated from agriculture to trade services, instead of manufacturing ADDIN EN.CITE <EndNote><Cite><Author>De Vries</Author><Year>2015</Year><RecNum>67</RecNum><DisplayText>(De Vries, Timmer, and De Vries 2015)</DisplayText><record><rec-number>67</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1530239341">67</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>De Vries, Gaaitzen</author><author>Timmer, Marcel</author><author>De Vries, Klaas</author></authors></contributors><titles><title>Structural transformation in Africa: Static gains, dynamic losses</title><secondary-title>The Journal of Development Studies</secondary-title></titles><periodical><full-title>The Journal of Development Studies</full-title></periodical><pages>674-688</pages><volume>51</volume><number>6</number><dates><year>2015</year></dates><isbn>0022-0388</isbn><urls></urls></record></Cite></EndNote>(De Vries, Timmer, and De Vries 2015). ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>De Vries</Author><Year>2015</Year><RecNum>67</RecNum><DisplayText>De Vries, Timmer, and De Vries (2015)</DisplayText><record><rec-number>67</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1530239341">67</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>De Vries, Gaaitzen</author><author>Timmer, Marcel</author><author>De Vries, Klaas</author></authors></contributors><titles><title>Structural transformation in Africa: Static gains, dynamic losses</title><secondary-title>The Journal of Development Studies</secondary-title></titles><periodical><full-title>The Journal of Development Studies</full-title></periodical><pages>674-688</pages><volume>51</volume><number>6</number><dates><year>2015</year></dates><isbn>0022-0388</isbn><urls></urls></record></Cite></EndNote>De Vries, Timmer, and De Vries (2015) and ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Duarte</Author><Year>2010</Year><RecNum>56</RecNum><DisplayText>Duarte and Restuccia (2010)</DisplayText><record><rec-number>56</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1530238465">56</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Duarte, Margarida</author><author>Restuccia, Diego</author></authors></contributors><titles><title>The role of the structural transformation in aggregate productivity</title><secondary-title>The Quarterly Journal of Economics</secondary-title></titles><periodical><full-title>The Quarterly Journal of Economics</full-title></periodical><pages>129-173</pages><volume>125</volume><number>1</number><dates><year>2010</year></dates><isbn>1531-4650</isbn><urls></urls></record></Cite></EndNote>Duarte and Restuccia (2010) point out that even though productivity levels in trade services are often higher than productivity in agriculture in Africa, its productivity growth is sluggish and stagnant. ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Geiger</Author><Year>2018</Year><RecNum>183</RecNum><DisplayText>Geiger, Trenczek, and Wacker (2018)</DisplayText><record><rec-number>183</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1538575689">183</key></foreign-keys><ref-type name="Unpublished Work">34</ref-type><contributors><authors><author>Geiger, Michael</author><author>Trenczek, Jan </author><author>Wacker, Konstantin M. </author></authors></contributors><titles><title>Understanding Economic Growth in Ghana in Comparative Perspective </title></titles><dates><year>2018</year></dates><publisher>World Bank</publisher><urls></urls></record></Cite></EndNote>Geiger, Trenczek, and Wacker (2018) confirm that Ghana is following the same trend; productivity in services is stagnant and has been declining in Ghana since 2005. The question arises whether structural change, a shift of labor from agriculture to services, contributed to poverty reduction in Ghana. ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Gollin</Author><Year>2002</Year><RecNum>85</RecNum><DisplayText>Gollin, Parente, and Rogerson (2002)</DisplayText><record><rec-number>85</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1530241834">85</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Gollin, Douglas</author><author>Parente, Stephen</author><author>Rogerson, Richard</author></authors></contributors><titles><title>The role of agriculture in development</title><secondary-title>American Economic Review</secondary-title></titles><periodical><full-title>American Economic Review</full-title></periodical><pages>160-164</pages><volume>92</volume><number>2</number><dates><year>2002</year></dates><isbn>0002-8282</isbn><urls></urls></record></Cite></EndNote>Gollin, Parente, and Rogerson (2002) and ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Christiaensen</Author><Year>2011</Year><RecNum>64</RecNum><DisplayText>Christiaensen, Demery, and Kuhl (2011)</DisplayText><record><rec-number>64</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1530239100">64</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Christiaensen, Luc</author><author>Demery, Lionel</author><author>Kuhl, Jesper</author></authors></contributors><titles><title>The (evolving) role of agriculture in poverty reduction—An empirical perspective</title><secondary-title>Journal of development economics</secondary-title></titles><periodical><full-title>Journal of development economics</full-title></periodical><pages>239-254</pages><volume>96</volume><number>2</number><dates><year>2011</year></dates><isbn>0304-3878</isbn><urls></urls></record></Cite></EndNote>Christiaensen, Demery, and Kuhl (2011) demonstrate how agriculture can be effective in reducing poverty, especially among the poorest. ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Breisinger</Author><Year>2008</Year><RecNum>184</RecNum><DisplayText>Breisinger et al. (2008)</DisplayText><record><rec-number>184</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1538578584">184</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Breisinger, Clemens</author><author>Diao, Xinshen</author><author>Thurlow, James</author><author>Al-Hassan, Ramatu M.</author></authors></contributors><titles><title>Agriculture for Development in Ghana: New Opportunities and Challenges</title><secondary-title>International Food Policy Research Institute Paper</secondary-title></titles><periodical><full-title>International Food Policy Research Institute Paper</full-title></periodical><pages>1-48</pages><volume>00784</volume><keywords><keyword>CAADP, agriculture, poverty, Africa, Ghana, CGE</keyword></keywords><dates><year>2008</year></dates><urls><related-urls><url> et al. (2008) shows the poverty rate among cocoa farmers declined from 60 percent to 24 percent in Ghana between 1991 and 2005. Beginning in the late 1990s, cocoa production rapidly grew due to favorable prices, and Ghana has become the world’s second-largest cocoa producer. As a result, the poverty rate among cocoa farmers significantly declined. This paper validate that increases in agricultural income contributed to increased consumption not only in cocoa production areas, but also in the whole country. People in poor areas are often disadvantaged by lack of access to infrastructure such as roads and electricity. Access to infrastructure is necessary for developing household enterprises, raising productivity and increasing incomes. In South Africa, household electrification encouraged the establishments of microenterprises and raised employment ADDIN EN.CITE <EndNote><Cite><Author>Dinkelman</Author><Year>2011</Year><RecNum>105</RecNum><DisplayText>(Dinkelman 2011)</DisplayText><record><rec-number>105</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1535298370">105</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Dinkelman, Taryn</author></authors></contributors><titles><title>The effects of rural electrification on employment: New evidence from South Africa</title><secondary-title>American Economic Review</secondary-title></titles><periodical><full-title>American Economic Review</full-title></periodical><pages>3078-3108</pages><volume>101</volume><number>7</number><dates><year>2011</year></dates><isbn>0002-8282</isbn><urls></urls></record></Cite></EndNote>(Dinkelman 2011). Improved roads stimulate industrial development ADDIN EN.CITE <EndNote><Cite><Author>Fernald</Author><Year>1999</Year><RecNum>107</RecNum><DisplayText>(Fernald 1999)</DisplayText><record><rec-number>107</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1535298787">107</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Fernald, John G</author></authors></contributors><titles><title>Roads to prosperity? Assessing the link between public capital and productivity</title><secondary-title>American economic review</secondary-title></titles><periodical><full-title>American Economic Review</full-title></periodical><pages>619-638</pages><volume>89</volume><number>3</number><dates><year>1999</year></dates><isbn>0002-8282</isbn><urls></urls></record></Cite></EndNote>(Fernald 1999), and benefits the poor ADDIN EN.CITE <EndNote><Cite><Author>Jacoby</Author><Year>2000</Year><RecNum>109</RecNum><DisplayText>(Jacoby 2000)</DisplayText><record><rec-number>109</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1535298931">109</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Jacoby, Hanan G</author></authors></contributors><titles><title>Access to markets and the benefits of rural roads</title><secondary-title>The Economic Journal</secondary-title></titles><periodical><full-title>The Economic Journal</full-title></periodical><pages>713-737</pages><volume>110</volume><number>465</number><dates><year>2000</year></dates><isbn>0013-0133</isbn><urls></urls></record></Cite></EndNote>(Jacoby 2000). Growth of road and electricity-generating capacity accounted for approximately half the growth of the productivity residual of India’s manufacturing sector ADDIN EN.CITE <EndNote><Cite><Author>Hulten</Author><Year>2006</Year><RecNum>117</RecNum><DisplayText>(Hulten, Bennathan, and Srinivasan 2006)</DisplayText><record><rec-number>117</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1535300472">117</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Hulten, Charles R</author><author>Bennathan, Esra</author><author>Srinivasan, Sylaja</author></authors></contributors><titles><title>Infrastructure, externalities, and economic development: a study of the Indian manufacturing industry</title><secondary-title>The World Bank Economic Review</secondary-title></titles><periodical><full-title>The World Bank Economic Review</full-title></periodical><pages>291-308</pages><volume>20</volume><number>2</number><dates><year>2006</year></dates><isbn>1564-698X</isbn><urls></urls></record></Cite></EndNote>(Hulten, Bennathan, and Srinivasan 2006). In Ghana, there is a substantial difference in access to electricity, market and roads between poor and rich areas. It is conceivable that lack of access to infrastructure inhibits the development of non-agricultural sectors, and misallocate labor across sectors ADDIN EN.CITE <EndNote><Cite><Author>Gollin</Author><Year>2013</Year><RecNum>50</RecNum><DisplayText>(Gollin, Lagakos, and Waugh 2013)</DisplayText><record><rec-number>50</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1530237705">50</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Gollin, Douglas</author><author>Lagakos, David</author><author>Waugh, Michael E</author></authors></contributors><titles><title>The agricultural productivity gap</title><secondary-title>The Quarterly Journal of Economics</secondary-title></titles><periodical><full-title>The Quarterly Journal of Economics</full-title></periodical><pages>939-993</pages><volume>129</volume><number>2</number><dates><year>2013</year></dates><isbn>0033-5533</isbn><urls></urls></record></Cite></EndNote>(Gollin, Lagakos, and Waugh 2013). The spatial inequality may not only reflect disparities in infrastructure but also ecological differences. Even though agriculture remains the dominant sector in the three Northern regions, the climate is not suitable for cocoa and other cash crops. Farmers in these regions are mainly engaged in rain-fed, traditional subsistence agriculture, as they have limited access to irrigation. In addition, the northern regions are frequently affected by floods and droughts, accompanied by high temperatures and intense heat (Figure 4). The northern floods of 2007, for example, affected 317,000 people, destroyed 1,000 km of roads, 210 schools, and 45 health clinics, and damaged or contaminated 630 water facilities. Furthermore, it was?followed immediately by drought. Crop losses are higher in districts with no irrigation, and when floods and droughts occur one after another ADDIN EN.CITE <EndNote><Cite><Author>Yiran</Author><Year>2016</Year><RecNum>175</RecNum><DisplayText>(Yiran and Stringer 2016)</DisplayText><record><rec-number>175</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1537921910">175</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Yiran, Gerald AB</author><author>Stringer, Lindsay C</author></authors></contributors><titles><title>Spatio-temporal analyses of impacts of multiple climatic hazards in a savannah ecosystem of Ghana</title><secondary-title>Climate Risk Management</secondary-title></titles><periodical><full-title>Climate Risk Management</full-title></periodical><pages>11-26</pages><volume>14</volume><dates><year>2016</year></dates><isbn>2212-0963</isbn><urls></urls></record></Cite></EndNote>(Yiran and Stringer 2016). Poverty in the northern regions may have been also influenced by land degradation, as land degradation is concentrated in the poor northern regions (Figure 4). land degradation leads to lower agricultural production, which in turn leads to food insecurity, poverty and vulnerability ADDIN EN.CITE <EndNote><Cite><Author>Lal</Author><Year>2004</Year><RecNum>127</RecNum><DisplayText>(Lal 2004, West et al. 2014)</DisplayText><record><rec-number>127</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1537396504">127</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Lal, Rattan</author></authors></contributors><titles><title>Soil carbon sequestration impacts on global climate change and food security</title><secondary-title>science</secondary-title></titles><periodical><full-title>science</full-title></periodical><pages>1623-1627</pages><volume>304</volume><number>5677</number><dates><year>2004</year></dates><isbn>0036-8075</isbn><urls></urls></record></Cite><Cite><Author>West</Author><Year>2014</Year><RecNum>128</RecNum><record><rec-number>128</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1537396682">128</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>West, Paul C</author><author>Gerber, James S</author><author>Engstrom, Peder M</author><author>Mueller, Nathaniel D</author><author>Brauman, Kate A</author><author>Carlson, Kimberly M</author><author>Cassidy, Emily S</author><author>Johnston, Matt</author><author>MacDonald, Graham K</author><author>Ray, Deepak K</author></authors></contributors><titles><title>Leverage points for improving global food security and the environment</title><secondary-title>Science</secondary-title></titles><periodical><full-title>science</full-title></periodical><pages>325-328</pages><volume>345</volume><number>6194</number><dates><year>2014</year></dates><isbn>0036-8075</isbn><urls></urls></record></Cite></EndNote>(Lal 2004, West et al. 2014).This study combines district level poverty rates, population census data, income data, and geospatial data to investigate how human capital, structural transformation, access to infrastructure, and environmental factors affected poverty, income and employment. We find poverty reduction was achieved through increased employment rates, especially women, increased income in the service sector, shifts of labor from agriculture to services, expansion of access to electricity, and increased rainfall. In addition, education, access to electricity, shift of labor from agriculture to services are important determining factors of the level of poverty. The paper further investigates what determines agricultural income and structural change. Better access to electricity is the most critical factor facilitating the shift of labor from agriculture to industry and services. Soil erosion had a large impact on the changes in agricultural income. Severe drought inhibits people from moving from agriculture to non-agricultural sectors. It may be because a large economic loss from drought prevents people from accumulating capital for investment. The results of this study suggest that for Ghana to reduce poverty and create jobs in non-agricultural sectors in lagging regions and districts, it needs to invest on infrastructure, and mitigate economic losses from soil degradation and drought.This paper relates to a large literature on income and welfare mobility ADDIN EN.CITE <EndNote><Cite><Author>Fields</Author><Year>2003</Year><RecNum>54</RecNum><DisplayText>(Fields et al. 2003, J?ntti and Jenkins 2015)</DisplayText><record><rec-number>54</rec-number><foreign-keys><key app="EN" db-id="2wfd5rtfos2e0qe2998vx90isetx99ew9ppp" timestamp="1545134258">54</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Fields, Gary</author><author>Cichello, Paul</author><author>Freije, Samuel</author><author>Menéndez, Marta</author><author>Newhouse, David</author></authors></contributors><titles><title>Household income dynamics: a four-country story</title><secondary-title>The journal of development studies</secondary-title></titles><periodical><full-title>The journal of development studies</full-title></periodical><pages>30-54</pages><volume>40</volume><number>2</number><dates><year>2003</year></dates><isbn>0022-0388</isbn><urls></urls></record></Cite><Cite><Author>J?ntti</Author><Year>2015</Year><RecNum>53</RecNum><record><rec-number>53</rec-number><foreign-keys><key app="EN" db-id="2wfd5rtfos2e0qe2998vx90isetx99ew9ppp" timestamp="1545134227">53</key></foreign-keys><ref-type name="Book Section">5</ref-type><contributors><authors><author>J?ntti, Markus</author><author>Jenkins, Stephen P</author></authors></contributors><titles><title>Income mobility</title><secondary-title>Handbook of income distribution</secondary-title></titles><pages>807-935</pages><volume>2</volume><dates><year>2015</year></dates><publisher>Elsevier</publisher><isbn>1574-0056</isbn><urls></urls></record></Cite></EndNote>(Fields et al. 2003, J?ntti and Jenkins 2015), and spatial dynamics PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5IaWdnaW5zPC9BdXRob3I+PFllYXI+MjAxMDwvWWVhcj48

UmVjTnVtPjU1PC9SZWNOdW0+PERpc3BsYXlUZXh0PihIaWdnaW5zLCBCaXJkLCBhbmQgSGFycmlz

IDIwMTAsIEphbGFuIGFuZCBSYXZhbGxpb24gMjAwMiwgRGUgVnJleWVyLCBIZXJyZXJhLCBhbmQg

TWVzcGzDqS1Tb21wcyAyMDA5KTwvRGlzcGxheVRleHQ+PHJlY29yZD48cmVjLW51bWJlcj41NTwv

cmVjLW51bWJlcj48Zm9yZWlnbi1rZXlzPjxrZXkgYXBwPSJFTiIgZGItaWQ9IjJ3ZmQ1cnRmb3My

ZTBxZTI5OTh2eDkwaXNldHg5OWV3OXBwcCIgdGltZXN0YW1wPSIxNTQ1MTM0OTgzIj41NTwva2V5

PjwvZm9yZWlnbi1rZXlzPjxyZWYtdHlwZSBuYW1lPSJKb3VybmFsIEFydGljbGUiPjE3PC9yZWYt

dHlwZT48Y29udHJpYnV0b3JzPjxhdXRob3JzPjxhdXRob3I+SGlnZ2lucywgS2F0ZTwvYXV0aG9y

PjxhdXRob3I+QmlyZCwgS2F0ZTwvYXV0aG9yPjxhdXRob3I+SGFycmlzLCBEYW5pZWw8L2F1dGhv

cj48L2F1dGhvcnM+PC9jb250cmlidXRvcnM+PHRpdGxlcz48dGl0bGU+UG9saWN5IHJlc3BvbnNl

cyB0byB0aGUgc3BhdGlhbCBkaW1lbnNpb25zIG9mIHBvdmVydHk8L3RpdGxlPjwvdGl0bGVzPjxk

YXRlcz48eWVhcj4yMDEwPC95ZWFyPjwvZGF0ZXM+PGlzYm4+MTkwNzI4ODEzOTwvaXNibj48dXJs

cz48L3VybHM+PC9yZWNvcmQ+PC9DaXRlPjxDaXRlPjxBdXRob3I+SmFsYW48L0F1dGhvcj48WWVh

cj4yMDAyPC9ZZWFyPjxSZWNOdW0+NTY8L1JlY051bT48cmVjb3JkPjxyZWMtbnVtYmVyPjU2PC9y

ZWMtbnVtYmVyPjxmb3JlaWduLWtleXM+PGtleSBhcHA9IkVOIiBkYi1pZD0iMndmZDVydGZvczJl

MHFlMjk5OHZ4OTBpc2V0eDk5ZXc5cHBwIiB0aW1lc3RhbXA9IjE1NDUxMzUwMzciPjU2PC9rZXk+

PC9mb3JlaWduLWtleXM+PHJlZi10eXBlIG5hbWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3JlZi10

eXBlPjxjb250cmlidXRvcnM+PGF1dGhvcnM+PGF1dGhvcj5KYWxhbiwgSnlvdHNuYTwvYXV0aG9y

PjxhdXRob3I+UmF2YWxsaW9uLCBNYXJ0aW48L2F1dGhvcj48L2F1dGhvcnM+PC9jb250cmlidXRv

cnM+PHRpdGxlcz48dGl0bGU+R2VvZ3JhcGhpYyBwb3ZlcnR5IHRyYXBzPyBBIG1pY3JvIG1vZGVs

IG9mIGNvbnN1bXB0aW9uIGdyb3d0aCBpbiBydXJhbCBDaGluYTwvdGl0bGU+PHNlY29uZGFyeS10

aXRsZT5Kb3VybmFsIG9mIGFwcGxpZWQgZWNvbm9tZXRyaWNzPC9zZWNvbmRhcnktdGl0bGU+PC90

aXRsZXM+PHBlcmlvZGljYWw+PGZ1bGwtdGl0bGU+Sm91cm5hbCBvZiBhcHBsaWVkIGVjb25vbWV0

cmljczwvZnVsbC10aXRsZT48L3BlcmlvZGljYWw+PHBhZ2VzPjMyOS0zNDY8L3BhZ2VzPjx2b2x1

bWU+MTc8L3ZvbHVtZT48bnVtYmVyPjQ8L251bWJlcj48ZGF0ZXM+PHllYXI+MjAwMjwveWVhcj48

L2RhdGVzPjxpc2JuPjA4ODMtNzI1MjwvaXNibj48dXJscz48L3VybHM+PC9yZWNvcmQ+PC9DaXRl

PjxDaXRlPjxBdXRob3I+RGUgVnJleWVyPC9BdXRob3I+PFllYXI+MjAwOTwvWWVhcj48UmVjTnVt

PjU3PC9SZWNOdW0+PHJlY29yZD48cmVjLW51bWJlcj41NzwvcmVjLW51bWJlcj48Zm9yZWlnbi1r

ZXlzPjxrZXkgYXBwPSJFTiIgZGItaWQ9IjJ3ZmQ1cnRmb3MyZTBxZTI5OTh2eDkwaXNldHg5OWV3

OXBwcCIgdGltZXN0YW1wPSIxNTQ1MTM1MDc5Ij41Nzwva2V5PjwvZm9yZWlnbi1rZXlzPjxyZWYt

dHlwZSBuYW1lPSJKb3VybmFsIEFydGljbGUiPjE3PC9yZWYtdHlwZT48Y29udHJpYnV0b3JzPjxh

dXRob3JzPjxhdXRob3I+RGUgVnJleWVyLCBQaGlsaXBwZTwvYXV0aG9yPjxhdXRob3I+SGVycmVy

YSwgSmF2aWVyPC9hdXRob3I+PGF1dGhvcj5NZXNwbMOpLVNvbXBzLCBTYW5kcmluZTwvYXV0aG9y

PjwvYXV0aG9ycz48L2NvbnRyaWJ1dG9ycz48dGl0bGVzPjx0aXRsZT5Db25zdW1wdGlvbiBncm93

dGggYW5kIHNwYXRpYWwgcG92ZXJ0eSB0cmFwczogYW4gYW5hbHlzaXMgb2YgdGhlIGVmZmVjdCBv

ZiBzb2NpYWwgc2VydmljZXMgYW5kIGNvbW11bml0eSBpbmZyYXN0cnVjdHVyZXMgb24gbGl2aW5n

IHN0YW5kYXJkcyBpbiBydXJhbCBQZXJ1PC90aXRsZT48c2Vjb25kYXJ5LXRpdGxlPlBvdmVydHks

IEluZXF1YWxpdHksIGFuZCBQb2xpY3kgaW4gTGF0aW4gQW1lcmljYTwvc2Vjb25kYXJ5LXRpdGxl

PjwvdGl0bGVzPjxwZXJpb2RpY2FsPjxmdWxsLXRpdGxlPlBvdmVydHksIEluZXF1YWxpdHksIGFu

ZCBQb2xpY3kgaW4gTGF0aW4gQW1lcmljYTwvZnVsbC10aXRsZT48L3BlcmlvZGljYWw+PHBhZ2Vz

PjEyOS0xNTU8L3BhZ2VzPjxkYXRlcz48eWVhcj4yMDA5PC95ZWFyPjwvZGF0ZXM+PHVybHM+PC91

cmxzPjwvcmVjb3JkPjwvQ2l0ZT48L0VuZE5vdGU+AG==

ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5IaWdnaW5zPC9BdXRob3I+PFllYXI+MjAxMDwvWWVhcj48

UmVjTnVtPjU1PC9SZWNOdW0+PERpc3BsYXlUZXh0PihIaWdnaW5zLCBCaXJkLCBhbmQgSGFycmlz

IDIwMTAsIEphbGFuIGFuZCBSYXZhbGxpb24gMjAwMiwgRGUgVnJleWVyLCBIZXJyZXJhLCBhbmQg

TWVzcGzDqS1Tb21wcyAyMDA5KTwvRGlzcGxheVRleHQ+PHJlY29yZD48cmVjLW51bWJlcj41NTwv

cmVjLW51bWJlcj48Zm9yZWlnbi1rZXlzPjxrZXkgYXBwPSJFTiIgZGItaWQ9IjJ3ZmQ1cnRmb3My

ZTBxZTI5OTh2eDkwaXNldHg5OWV3OXBwcCIgdGltZXN0YW1wPSIxNTQ1MTM0OTgzIj41NTwva2V5

PjwvZm9yZWlnbi1rZXlzPjxyZWYtdHlwZSBuYW1lPSJKb3VybmFsIEFydGljbGUiPjE3PC9yZWYt

dHlwZT48Y29udHJpYnV0b3JzPjxhdXRob3JzPjxhdXRob3I+SGlnZ2lucywgS2F0ZTwvYXV0aG9y

PjxhdXRob3I+QmlyZCwgS2F0ZTwvYXV0aG9yPjxhdXRob3I+SGFycmlzLCBEYW5pZWw8L2F1dGhv

cj48L2F1dGhvcnM+PC9jb250cmlidXRvcnM+PHRpdGxlcz48dGl0bGU+UG9saWN5IHJlc3BvbnNl

cyB0byB0aGUgc3BhdGlhbCBkaW1lbnNpb25zIG9mIHBvdmVydHk8L3RpdGxlPjwvdGl0bGVzPjxk

YXRlcz48eWVhcj4yMDEwPC95ZWFyPjwvZGF0ZXM+PGlzYm4+MTkwNzI4ODEzOTwvaXNibj48dXJs

cz48L3VybHM+PC9yZWNvcmQ+PC9DaXRlPjxDaXRlPjxBdXRob3I+SmFsYW48L0F1dGhvcj48WWVh

cj4yMDAyPC9ZZWFyPjxSZWNOdW0+NTY8L1JlY051bT48cmVjb3JkPjxyZWMtbnVtYmVyPjU2PC9y

ZWMtbnVtYmVyPjxmb3JlaWduLWtleXM+PGtleSBhcHA9IkVOIiBkYi1pZD0iMndmZDVydGZvczJl

MHFlMjk5OHZ4OTBpc2V0eDk5ZXc5cHBwIiB0aW1lc3RhbXA9IjE1NDUxMzUwMzciPjU2PC9rZXk+

PC9mb3JlaWduLWtleXM+PHJlZi10eXBlIG5hbWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3JlZi10

eXBlPjxjb250cmlidXRvcnM+PGF1dGhvcnM+PGF1dGhvcj5KYWxhbiwgSnlvdHNuYTwvYXV0aG9y

PjxhdXRob3I+UmF2YWxsaW9uLCBNYXJ0aW48L2F1dGhvcj48L2F1dGhvcnM+PC9jb250cmlidXRv

cnM+PHRpdGxlcz48dGl0bGU+R2VvZ3JhcGhpYyBwb3ZlcnR5IHRyYXBzPyBBIG1pY3JvIG1vZGVs

IG9mIGNvbnN1bXB0aW9uIGdyb3d0aCBpbiBydXJhbCBDaGluYTwvdGl0bGU+PHNlY29uZGFyeS10

aXRsZT5Kb3VybmFsIG9mIGFwcGxpZWQgZWNvbm9tZXRyaWNzPC9zZWNvbmRhcnktdGl0bGU+PC90

aXRsZXM+PHBlcmlvZGljYWw+PGZ1bGwtdGl0bGU+Sm91cm5hbCBvZiBhcHBsaWVkIGVjb25vbWV0

cmljczwvZnVsbC10aXRsZT48L3BlcmlvZGljYWw+PHBhZ2VzPjMyOS0zNDY8L3BhZ2VzPjx2b2x1

bWU+MTc8L3ZvbHVtZT48bnVtYmVyPjQ8L251bWJlcj48ZGF0ZXM+PHllYXI+MjAwMjwveWVhcj48

L2RhdGVzPjxpc2JuPjA4ODMtNzI1MjwvaXNibj48dXJscz48L3VybHM+PC9yZWNvcmQ+PC9DaXRl

PjxDaXRlPjxBdXRob3I+RGUgVnJleWVyPC9BdXRob3I+PFllYXI+MjAwOTwvWWVhcj48UmVjTnVt

PjU3PC9SZWNOdW0+PHJlY29yZD48cmVjLW51bWJlcj41NzwvcmVjLW51bWJlcj48Zm9yZWlnbi1r

ZXlzPjxrZXkgYXBwPSJFTiIgZGItaWQ9IjJ3ZmQ1cnRmb3MyZTBxZTI5OTh2eDkwaXNldHg5OWV3

OXBwcCIgdGltZXN0YW1wPSIxNTQ1MTM1MDc5Ij41Nzwva2V5PjwvZm9yZWlnbi1rZXlzPjxyZWYt

dHlwZSBuYW1lPSJKb3VybmFsIEFydGljbGUiPjE3PC9yZWYtdHlwZT48Y29udHJpYnV0b3JzPjxh

dXRob3JzPjxhdXRob3I+RGUgVnJleWVyLCBQaGlsaXBwZTwvYXV0aG9yPjxhdXRob3I+SGVycmVy

YSwgSmF2aWVyPC9hdXRob3I+PGF1dGhvcj5NZXNwbMOpLVNvbXBzLCBTYW5kcmluZTwvYXV0aG9y

PjwvYXV0aG9ycz48L2NvbnRyaWJ1dG9ycz48dGl0bGVzPjx0aXRsZT5Db25zdW1wdGlvbiBncm93

dGggYW5kIHNwYXRpYWwgcG92ZXJ0eSB0cmFwczogYW4gYW5hbHlzaXMgb2YgdGhlIGVmZmVjdCBv

ZiBzb2NpYWwgc2VydmljZXMgYW5kIGNvbW11bml0eSBpbmZyYXN0cnVjdHVyZXMgb24gbGl2aW5n

IHN0YW5kYXJkcyBpbiBydXJhbCBQZXJ1PC90aXRsZT48c2Vjb25kYXJ5LXRpdGxlPlBvdmVydHks

IEluZXF1YWxpdHksIGFuZCBQb2xpY3kgaW4gTGF0aW4gQW1lcmljYTwvc2Vjb25kYXJ5LXRpdGxl

PjwvdGl0bGVzPjxwZXJpb2RpY2FsPjxmdWxsLXRpdGxlPlBvdmVydHksIEluZXF1YWxpdHksIGFu

ZCBQb2xpY3kgaW4gTGF0aW4gQW1lcmljYTwvZnVsbC10aXRsZT48L3BlcmlvZGljYWw+PHBhZ2Vz

PjEyOS0xNTU8L3BhZ2VzPjxkYXRlcz48eWVhcj4yMDA5PC95ZWFyPjwvZGF0ZXM+PHVybHM+PC91

cmxzPjwvcmVjb3JkPjwvQ2l0ZT48L0VuZE5vdGU+AG==

ADDIN EN.CITE.DATA (Higgins, Bird, and Harris 2010, Jalan and Ravallion 2002, De Vreyer, Herrera, and Mesplé-Somps 2009). We make a significant contribution to the existing literature by combining household survey data with geospatial information. Geospatial information allows us to examine the impacts of environmental degradation on poverty and structural transformation. This paper shows climate change and environmental degradation are disproportionally affecting the poor areas. DataThis study combines poverty rates, population census data, mean income by sector estimated from household survey data (Ghana Living Standards Survey [GLSS]), and the Hidden Dimensions Dataset (HDD) at the district level to identify determining factors of the level and change of poverty rates. District poverty rateWe use poverty rates estimated at two time periods, 2000 and 2010. ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Coulombe</Author><Year>2005</Year><RecNum>1</RecNum><DisplayText>Coulombe (2005)</DisplayText><record><rec-number>1</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1529714835">1</key></foreign-keys><ref-type name="Press Release">63</ref-type><contributors><authors><author>Coulombe, Harold</author></authors></contributors><titles><title>Ghana Census-Based Poverty Map: District and Sub-District Level Results</title></titles><dates><year>2005</year></dates><publisher>Ghana Statistical Service,&#xD;Department for International Development UK,</publisher><urls><related-urls><url> (2005) applies the small-area estimation technique developed by ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Elbers</Author><Year>2003</Year><RecNum>3</RecNum><DisplayText>Elbers, Lanjouw, and Lanjouw (2003)</DisplayText><record><rec-number>3</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1529957288">3</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Elbers, Chris</author><author>Lanjouw, Jean O</author><author>Lanjouw, Peter</author></authors></contributors><titles><title>Micro–level estimation of poverty and inequality</title><secondary-title>Econometrica</secondary-title></titles><periodical><full-title>Econometrica</full-title></periodical><pages>355-364</pages><volume>71</volume><number>1</number><dates><year>2003</year></dates><isbn>1468-0262</isbn><urls></urls></record></Cite></EndNote>Elbers, Lanjouw, and Lanjouw (2003), and estimates the poverty rates of 110 districts, using the 2000 Population and Housing Census and GLSS 4. The data collection of GLSS 4 was carried out from April 1998 to March 1999. We call them ‘2000 poverty rates’ throughout the paper, as the 2000 Population and Housing Census was used for the estimation. The district with the highest poverty rate was East Mamprusi in the Northern region, with a poverty rate of 86.1 percent. The district with the lowest poverty rate was Accra Metropolitan in Greater Accra, with a poverty rate of 5.2 percent.Using the 2010 Population and Housing Census and GLSS 6 conducted from October 2012 to October 2013, ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Ghana Statistical Service</Author><Year>2015</Year><RecNum>2</RecNum><DisplayText>Ghana Statistical Service (2015)</DisplayText><record><rec-number>2</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1529714848">2</key></foreign-keys><ref-type name="Report">27</ref-type><contributors><authors><author>Ghana Statistical Service,</author></authors></contributors><titles><title>Ghana Poverty Mapping Report</title></titles><dates><year>2015</year></dates><publisher>Ghana Statistical Service,</publisher><urls><related-urls><url> Statistical Service (2015) estimates poverty rates of 216 districts, using the same small-area estimation technique ADDIN EN.CITE <EndNote><Cite><Author>Elbers</Author><Year>2003</Year><RecNum>3</RecNum><DisplayText>(Elbers, Lanjouw, and Lanjouw 2003)</DisplayText><record><rec-number>3</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1529957288">3</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Elbers, Chris</author><author>Lanjouw, Jean O</author><author>Lanjouw, Peter</author></authors></contributors><titles><title>Micro–level estimation of poverty and inequality</title><secondary-title>Econometrica</secondary-title></titles><periodical><full-title>Econometrica</full-title></periodical><pages>355-364</pages><volume>71</volume><number>1</number><dates><year>2003</year></dates><isbn>1468-0262</isbn><urls></urls></record></Cite></EndNote>(Elbers, Lanjouw, and Lanjouw 2003) as ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Coulombe</Author><Year>2005</Year><RecNum>1</RecNum><DisplayText>Coulombe (2005)</DisplayText><record><rec-number>1</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1529714835">1</key></foreign-keys><ref-type name="Press Release">63</ref-type><contributors><authors><author>Coulombe, Harold</author></authors></contributors><titles><title>Ghana Census-Based Poverty Map: District and Sub-District Level Results</title></titles><dates><year>2005</year></dates><publisher>Ghana Statistical Service,&#xD;Department for International Development UK,</publisher><urls><related-urls><url> (2005). We call them ‘2010 poverty rates’ throughout the paper, as the 2010 Population and Housing Census was used for the estimation. The district with the lowest poverty rate was La Dade Kotopon Municipal in Greater Accra, with a poverty rate of 1.3 percent. The district with the highest poverty rate was Wa West in the Upper West region, with a poverty rate of 92.4 percent. The 2010 poverty rates are the latest district level data on poverty at moment. This project will be extended and new district poverty rates will be estimated using the survey data from GLSS 7, which was collected from October 2016 to October 2017. Between 2000 and 2010, the number of districts increased from 110 to 216. It was done by dividing the original districts into smaller districts. To analyze the changes of district poverty rates between the two-time periods, we need to create a panel data of district poverty rates. The paper estimated the poverty rates of the 110 old district units for 2010, by merging the data of new 216 districts into 110 old districts using district population as weights. Population census The 2000 and 2010 Population and Housing Census gathered information on household members in various aspects including educational attainment, migration, demographic characteristics, employment, and fertility. Mean income by sectorMean income per worker is estimated for three major sectors, i.e. agriculture, industry and service, using the Rural Income Generating Activities Labor (RIGA-L) database, constructed by the Food and Agriculture Organization of the United Nations (FAO). RIGA-L is an individual employment dataset created from household survey data. This study uses the RIGA-L created from GLSS 4 (1998/99) and GLSS 6 (2012/13). The RIGA income data is constructed using GLSS data with the idea of including all sources household income that are often neglected in the computation of income. Such sources include income from self-employment and income from own-farm production of crops. The paper calculates the mean income of workers who hold jobs in a specific sector (either as primary jobs or secondary jobs) in each of the 110 original districts. It is not possible to obtain a reliable mean income of a sector if there are not sufficient number of workers in a sector. There are fewer workers in the industry and service sectors than in agriculture, particularly in poor districts in earlier survey data (GLSS 4). Reliable data of mean income in the industry and service sectors was obtained only for 88 and 93 out of 110 districts in 1998/99, respectively. Because of the reliability concern, the paper uses the mean income data of industry and services only in the regressions presented in Tables 2 and 3, in which we investigate the impacts of changes in per capita income and share of employment on the changes on poverty rates and consumption. Hidden Dimensions Dataset (HHD)HDD is a global geospatial dataset created by the World Bank, linking environment and natural resource measures to poverty and other human development indicators. This paper uses the following variables from the dataset: normalized difference vegetation index (NDVI), net primary productivity (NPP), night time lights (NTL), soil degradation, less favored agricultural land (LFAL), road density, distance to market, flood frequency, drought severity, and air pollution. The dataset was created for the 110-district boundary for this study. Precipitation, NDVI, NPP, NTL, and air pollution data are available for multiple years including 2000 and 2010, but data on soil erosion, LFAL, road density, distance to market, flood frequency, and drought severity are not available for multiple years. Normalized Difference Vegetation index (NDVI)Normalized Difference Vegetation Index (NDVI) is a satellite imagery-derived measure of “greenness”, or the relative density and health of vegetation, of the earth’s surface. It specifies where and how much green vegetation is growing at a certain time. The equation combining the bands of visible red (VIS) and near-infrared (NIR) that produces the NDVI is given by:NDVI=NIR-VISNIR+VISThe data is produced and collected by the Moderate Resolution Imaging Spectroradiometer (MODIS) aboard NASA’s terra satellite. The resolution image used as an input is 10km2, and the values range from -1 and +1. Values greater than 0.1 generally indicate increasing degrees in the greenness and intensity of vegetation. Values between 0 and 0.1 are normally representative of rocks and bare soil, and values less than 0 often imply clouds, rain, and snow. The NDVI for each pixel is averaged across all grid cells within the boundary of districts to obtain the mean NDVI of each district. This paper uses the mean NDVI estimated at the 110-district level in 2000 and 2010. NDVI is used to predict crop prices PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5Ccm93bjwvQXV0aG9yPjxZZWFyPjIwMDg8L1llYXI+PFJl

Y051bT40MzwvUmVjTnVtPjxEaXNwbGF5VGV4dD4oQnJvd24sIFBpbnpvbiwgYW5kIFByaW5jZSAy

MDA4LCBIaWdnaW5zLCBIaW50ZXJtYW5uLCBhbmQgQnJvd24gMjAxNSk8L0Rpc3BsYXlUZXh0Pjxy

ZWNvcmQ+PHJlYy1udW1iZXI+NDM8L3JlYy1udW1iZXI+PGZvcmVpZ24ta2V5cz48a2V5IGFwcD0i

RU4iIGRiLWlkPSJ3eHB4YXd3emV3ZWZ6NmVzcng1eGRhMm90dmUyNXBwcDVlMGYiIHRpbWVzdGFt

cD0iMTUzMDIzNjQzMCI+NDM8L2tleT48L2ZvcmVpZ24ta2V5cz48cmVmLXR5cGUgbmFtZT0iSm91

cm5hbCBBcnRpY2xlIj4xNzwvcmVmLXR5cGU+PGNvbnRyaWJ1dG9ycz48YXV0aG9ycz48YXV0aG9y

PkJyb3duLCBNb2xseSBFLjwvYXV0aG9yPjxhdXRob3I+UGluem9uLCBKb3JnZSBFLjwvYXV0aG9y

PjxhdXRob3I+UHJpbmNlLCBTdGVwaGVuIEQuPC9hdXRob3I+PC9hdXRob3JzPjwvY29udHJpYnV0

b3JzPjxhdXRoLWFkZHJlc3M+U1NBSS1OQVNBIEdvZGRhcmQgU3BhY2UgRmxpZ2h0IENlbnRlciYj

eEQ7VSBNRDwvYXV0aC1hZGRyZXNzPjx0aXRsZXM+PHRpdGxlPlVzaW5nIFNhdGVsbGl0ZSBSZW1v

dGUgU2Vuc2luZyBEYXRhIGluIGEgU3BhdGlhbGx5IEV4cGxpY2l0IFByaWNlIE1vZGVsOiBWZWdl

dGF0aW9uIER5bmFtaWNzIGFuZCBNaWxsZXQgUHJpY2VzPC90aXRsZT48c2Vjb25kYXJ5LXRpdGxl

PkxhbmQgRWNvbm9taWNzPC9zZWNvbmRhcnktdGl0bGU+PC90aXRsZXM+PHBlcmlvZGljYWw+PGZ1

bGwtdGl0bGU+TGFuZCBFY29ub21pY3M8L2Z1bGwtdGl0bGU+PC9wZXJpb2RpY2FsPjxwYWdlcz4z

NDAtMzU3PC9wYWdlcz48dm9sdW1lPjg0PC92b2x1bWU+PG51bWJlcj4yPC9udW1iZXI+PGtleXdv

cmRzPjxrZXl3b3JkPkVjb25vbWljIERldmVsb3BtZW50OiBBZ3JpY3VsdHVyZTwva2V5d29yZD48

a2V5d29yZD5OYXR1cmFsIFJlc291cmNlczwva2V5d29yZD48a2V5d29yZD5FbmVyZ3k8L2tleXdv

cmQ+PGtleXdvcmQ+RW52aXJvbm1lbnQ8L2tleXdvcmQ+PGtleXdvcmQ+T3RoZXIgUHJpbWFyeSBQ

cm9kdWN0cyBPMTM8L2tleXdvcmQ+PGtleXdvcmQ+QWdyaWN1bHR1cmFsIFImYW1wO0Q8L2tleXdv

cmQ+PGtleXdvcmQ+QWdyaWN1bHR1cmFsIFRlY2hub2xvZ3k8L2tleXdvcmQ+PGtleXdvcmQ+Qmlv

ZnVlbHM8L2tleXdvcmQ+PGtleXdvcmQ+QWdyaWN1bHR1cmFsIEV4dGVuc2lvbiBTZXJ2aWNlcyBR

MTY8L2tleXdvcmQ+PC9rZXl3b3Jkcz48ZGF0ZXM+PHllYXI+MjAwODwveWVhcj48L2RhdGVzPjxp

c2JuPjAwMjM3NjM5PC9pc2JuPjxhY2Nlc3Npb24tbnVtPjA5NzY3OTYgQWx0ZXJuYXRlIEFjY2Vz

c2lvbiBOdW1iZXI6IEVQMzIxOTcwNTg8L2FjY2Vzc2lvbi1udW0+PHVybHM+PHJlbGF0ZWQtdXJs

cz48dXJsPmh0dHA6Ly9zZWFyY2guZWJzY29ob3N0LmNvbS9sb2dpbi5hc3B4P2RpcmVjdD10cnVl

JmFtcDtkYj1lb2gmYW1wO0FOPTA5NzY3OTYmYW1wO3NpdGU9ZWhvc3QtbGl2ZTwvdXJsPjx1cmw+

aHR0cDovL2xlLnV3cHJlc3Mub3JnL2NvbnRlbnQvYnkveWVhcjwvdXJsPjwvcmVsYXRlZC11cmxz

PjwvdXJscz48ZWxlY3Ryb25pYy1yZXNvdXJjZS1udW0+aHR0cDovL2xlLnV3cHJlc3Mub3JnL2Nv

bnRlbnQvYnkveWVhcjwvZWxlY3Ryb25pYy1yZXNvdXJjZS1udW0+PHJlbW90ZS1kYXRhYmFzZS1u

YW1lPmVvaDwvcmVtb3RlLWRhdGFiYXNlLW5hbWU+PHJlbW90ZS1kYXRhYmFzZS1wcm92aWRlcj5F

QlNDT2hvc3Q8L3JlbW90ZS1kYXRhYmFzZS1wcm92aWRlcj48L3JlY29yZD48L0NpdGU+PENpdGU+

PEF1dGhvcj5IaWdnaW5zPC9BdXRob3I+PFllYXI+MjAxNTwvWWVhcj48UmVjTnVtPjE5PC9SZWNO

dW0+PHJlY29yZD48cmVjLW51bWJlcj4xOTwvcmVjLW51bWJlcj48Zm9yZWlnbi1rZXlzPjxrZXkg

YXBwPSJFTiIgZGItaWQ9Ind4cHhhd3d6ZXdlZno2ZXNyeDV4ZGEyb3R2ZTI1cHBwNWUwZiIgdGlt

ZXN0YW1wPSIxNTMwMjMzODIyIj4xOTwva2V5PjwvZm9yZWlnbi1rZXlzPjxyZWYtdHlwZSBuYW1l

PSJKb3VybmFsIEFydGljbGUiPjE3PC9yZWYtdHlwZT48Y29udHJpYnV0b3JzPjxhdXRob3JzPjxh

dXRob3I+SGlnZ2lucywgTmF0aGFuaWVsPC9hdXRob3I+PGF1dGhvcj5IaW50ZXJtYW5uLCBCZWF0

PC9hdXRob3I+PGF1dGhvcj5Ccm93biwgTW9sbHkgRS48L2F1dGhvcj48L2F1dGhvcnM+PC9jb250

cmlidXRvcnM+PGF1dGgtYWRkcmVzcz5FY29ub21pYyBSZXNlYXJjaCBTZXJ2aWNlLCBVUyBEZXBh

cnRtZW50IG9mIEFncmljdWx0dXJlJiN4RDtVIEJhc2VsJiN4RDtOQVNBIEdvZGRhcmQgU3BhY2Ug

RmxpZ2h0IENlbnRlcjwvYXV0aC1hZGRyZXNzPjx0aXRsZXM+PHRpdGxlPkEgTW9kZWwgb2YgV2Vz

dCBBZnJpY2FuIE1pbGxldCBQcmljZXMgaW4gUnVyYWwgTWFya2V0czwvdGl0bGU+PHNlY29uZGFy

eS10aXRsZT5Gb29kIFBvbGljeTwvc2Vjb25kYXJ5LXRpdGxlPjwvdGl0bGVzPjxwZXJpb2RpY2Fs

PjxmdWxsLXRpdGxlPkZvb2QgUG9saWN5PC9mdWxsLXRpdGxlPjwvcGVyaW9kaWNhbD48cGFnZXM+

MzMtNDM8L3BhZ2VzPjx2b2x1bWU+NTI8L3ZvbHVtZT48a2V5d29yZHM+PGtleXdvcmQ+RWNvbm9t

aWMgRGV2ZWxvcG1lbnQ6IEFncmljdWx0dXJlPC9rZXl3b3JkPjxrZXl3b3JkPk5hdHVyYWwgUmVz

b3VyY2VzPC9rZXl3b3JkPjxrZXl3b3JkPkVuZXJneTwva2V5d29yZD48a2V5d29yZD5FbnZpcm9u

bWVudDwva2V5d29yZD48a2V5d29yZD5PdGhlciBQcmltYXJ5IFByb2R1Y3RzIE8xMzwva2V5d29y

ZD48a2V5d29yZD5BZ3JpY3VsdHVyZTogQWdncmVnYXRlIFN1cHBseSBhbmQgRGVtYW5kIEFuYWx5

c2lzPC9rZXl3b3JkPjxrZXl3b3JkPlByaWNlcyBRMTE8L2tleXdvcmQ+PC9rZXl3b3Jkcz48ZGF0

ZXM+PHllYXI+MjAxNTwveWVhcj48L2RhdGVzPjxpc2JuPjAzMDY5MTkyPC9pc2JuPjxhY2Nlc3Np

b24tbnVtPjE0OTAwNjg8L2FjY2Vzc2lvbi1udW0+PHVybHM+PHJlbGF0ZWQtdXJscz48dXJsPmh0

dHA6Ly9zZWFyY2guZWJzY29ob3N0LmNvbS9sb2dpbi5hc3B4P2RpcmVjdD10cnVlJmFtcDtkYj1l

b2gmYW1wO0FOPTE0OTAwNjgmYW1wO3NpdGU9ZWhvc3QtbGl2ZTwvdXJsPjx1cmw+aHR0cDovL2R4

LmRvaS5vcmcvMTAuMTAxNi9qLmZvb2Rwb2wuMjAxNC4wOS4wMTE8L3VybD48dXJsPmh0dHA6Ly93

d3cuc2NpZW5jZWRpcmVjdC5jb20vc2NpZW5jZS9qb3VybmFsLzAzMDY5MTkyPC91cmw+PC9yZWxh

dGVkLXVybHM+PC91cmxzPjxlbGVjdHJvbmljLXJlc291cmNlLW51bT5odHRwOi8vd3d3LnNjaWVu

Y2VkaXJlY3QuY29tL3NjaWVuY2Uvam91cm5hbC8wMzA2OTE5MjwvZWxlY3Ryb25pYy1yZXNvdXJj

ZS1udW0+PHJlbW90ZS1kYXRhYmFzZS1uYW1lPmVvaDwvcmVtb3RlLWRhdGFiYXNlLW5hbWU+PHJl

bW90ZS1kYXRhYmFzZS1wcm92aWRlcj5FQlNDT2hvc3Q8L3JlbW90ZS1kYXRhYmFzZS1wcm92aWRl

cj48L3JlY29yZD48L0NpdGU+PC9FbmROb3RlPgB=

ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5Ccm93bjwvQXV0aG9yPjxZZWFyPjIwMDg8L1llYXI+PFJl

Y051bT40MzwvUmVjTnVtPjxEaXNwbGF5VGV4dD4oQnJvd24sIFBpbnpvbiwgYW5kIFByaW5jZSAy

MDA4LCBIaWdnaW5zLCBIaW50ZXJtYW5uLCBhbmQgQnJvd24gMjAxNSk8L0Rpc3BsYXlUZXh0Pjxy

ZWNvcmQ+PHJlYy1udW1iZXI+NDM8L3JlYy1udW1iZXI+PGZvcmVpZ24ta2V5cz48a2V5IGFwcD0i

RU4iIGRiLWlkPSJ3eHB4YXd3emV3ZWZ6NmVzcng1eGRhMm90dmUyNXBwcDVlMGYiIHRpbWVzdGFt

cD0iMTUzMDIzNjQzMCI+NDM8L2tleT48L2ZvcmVpZ24ta2V5cz48cmVmLXR5cGUgbmFtZT0iSm91

cm5hbCBBcnRpY2xlIj4xNzwvcmVmLXR5cGU+PGNvbnRyaWJ1dG9ycz48YXV0aG9ycz48YXV0aG9y

PkJyb3duLCBNb2xseSBFLjwvYXV0aG9yPjxhdXRob3I+UGluem9uLCBKb3JnZSBFLjwvYXV0aG9y

PjxhdXRob3I+UHJpbmNlLCBTdGVwaGVuIEQuPC9hdXRob3I+PC9hdXRob3JzPjwvY29udHJpYnV0

b3JzPjxhdXRoLWFkZHJlc3M+U1NBSS1OQVNBIEdvZGRhcmQgU3BhY2UgRmxpZ2h0IENlbnRlciYj

eEQ7VSBNRDwvYXV0aC1hZGRyZXNzPjx0aXRsZXM+PHRpdGxlPlVzaW5nIFNhdGVsbGl0ZSBSZW1v

dGUgU2Vuc2luZyBEYXRhIGluIGEgU3BhdGlhbGx5IEV4cGxpY2l0IFByaWNlIE1vZGVsOiBWZWdl

dGF0aW9uIER5bmFtaWNzIGFuZCBNaWxsZXQgUHJpY2VzPC90aXRsZT48c2Vjb25kYXJ5LXRpdGxl

PkxhbmQgRWNvbm9taWNzPC9zZWNvbmRhcnktdGl0bGU+PC90aXRsZXM+PHBlcmlvZGljYWw+PGZ1

bGwtdGl0bGU+TGFuZCBFY29ub21pY3M8L2Z1bGwtdGl0bGU+PC9wZXJpb2RpY2FsPjxwYWdlcz4z

NDAtMzU3PC9wYWdlcz48dm9sdW1lPjg0PC92b2x1bWU+PG51bWJlcj4yPC9udW1iZXI+PGtleXdv

cmRzPjxrZXl3b3JkPkVjb25vbWljIERldmVsb3BtZW50OiBBZ3JpY3VsdHVyZTwva2V5d29yZD48

a2V5d29yZD5OYXR1cmFsIFJlc291cmNlczwva2V5d29yZD48a2V5d29yZD5FbmVyZ3k8L2tleXdv

cmQ+PGtleXdvcmQ+RW52aXJvbm1lbnQ8L2tleXdvcmQ+PGtleXdvcmQ+T3RoZXIgUHJpbWFyeSBQ

cm9kdWN0cyBPMTM8L2tleXdvcmQ+PGtleXdvcmQ+QWdyaWN1bHR1cmFsIFImYW1wO0Q8L2tleXdv

cmQ+PGtleXdvcmQ+QWdyaWN1bHR1cmFsIFRlY2hub2xvZ3k8L2tleXdvcmQ+PGtleXdvcmQ+Qmlv

ZnVlbHM8L2tleXdvcmQ+PGtleXdvcmQ+QWdyaWN1bHR1cmFsIEV4dGVuc2lvbiBTZXJ2aWNlcyBR

MTY8L2tleXdvcmQ+PC9rZXl3b3Jkcz48ZGF0ZXM+PHllYXI+MjAwODwveWVhcj48L2RhdGVzPjxp

c2JuPjAwMjM3NjM5PC9pc2JuPjxhY2Nlc3Npb24tbnVtPjA5NzY3OTYgQWx0ZXJuYXRlIEFjY2Vz

c2lvbiBOdW1iZXI6IEVQMzIxOTcwNTg8L2FjY2Vzc2lvbi1udW0+PHVybHM+PHJlbGF0ZWQtdXJs

cz48dXJsPmh0dHA6Ly9zZWFyY2guZWJzY29ob3N0LmNvbS9sb2dpbi5hc3B4P2RpcmVjdD10cnVl

JmFtcDtkYj1lb2gmYW1wO0FOPTA5NzY3OTYmYW1wO3NpdGU9ZWhvc3QtbGl2ZTwvdXJsPjx1cmw+

aHR0cDovL2xlLnV3cHJlc3Mub3JnL2NvbnRlbnQvYnkveWVhcjwvdXJsPjwvcmVsYXRlZC11cmxz

PjwvdXJscz48ZWxlY3Ryb25pYy1yZXNvdXJjZS1udW0+aHR0cDovL2xlLnV3cHJlc3Mub3JnL2Nv

bnRlbnQvYnkveWVhcjwvZWxlY3Ryb25pYy1yZXNvdXJjZS1udW0+PHJlbW90ZS1kYXRhYmFzZS1u

YW1lPmVvaDwvcmVtb3RlLWRhdGFiYXNlLW5hbWU+PHJlbW90ZS1kYXRhYmFzZS1wcm92aWRlcj5F

QlNDT2hvc3Q8L3JlbW90ZS1kYXRhYmFzZS1wcm92aWRlcj48L3JlY29yZD48L0NpdGU+PENpdGU+

PEF1dGhvcj5IaWdnaW5zPC9BdXRob3I+PFllYXI+MjAxNTwvWWVhcj48UmVjTnVtPjE5PC9SZWNO

dW0+PHJlY29yZD48cmVjLW51bWJlcj4xOTwvcmVjLW51bWJlcj48Zm9yZWlnbi1rZXlzPjxrZXkg

YXBwPSJFTiIgZGItaWQ9Ind4cHhhd3d6ZXdlZno2ZXNyeDV4ZGEyb3R2ZTI1cHBwNWUwZiIgdGlt

ZXN0YW1wPSIxNTMwMjMzODIyIj4xOTwva2V5PjwvZm9yZWlnbi1rZXlzPjxyZWYtdHlwZSBuYW1l

PSJKb3VybmFsIEFydGljbGUiPjE3PC9yZWYtdHlwZT48Y29udHJpYnV0b3JzPjxhdXRob3JzPjxh

dXRob3I+SGlnZ2lucywgTmF0aGFuaWVsPC9hdXRob3I+PGF1dGhvcj5IaW50ZXJtYW5uLCBCZWF0

PC9hdXRob3I+PGF1dGhvcj5Ccm93biwgTW9sbHkgRS48L2F1dGhvcj48L2F1dGhvcnM+PC9jb250

cmlidXRvcnM+PGF1dGgtYWRkcmVzcz5FY29ub21pYyBSZXNlYXJjaCBTZXJ2aWNlLCBVUyBEZXBh

cnRtZW50IG9mIEFncmljdWx0dXJlJiN4RDtVIEJhc2VsJiN4RDtOQVNBIEdvZGRhcmQgU3BhY2Ug

RmxpZ2h0IENlbnRlcjwvYXV0aC1hZGRyZXNzPjx0aXRsZXM+PHRpdGxlPkEgTW9kZWwgb2YgV2Vz

dCBBZnJpY2FuIE1pbGxldCBQcmljZXMgaW4gUnVyYWwgTWFya2V0czwvdGl0bGU+PHNlY29uZGFy

eS10aXRsZT5Gb29kIFBvbGljeTwvc2Vjb25kYXJ5LXRpdGxlPjwvdGl0bGVzPjxwZXJpb2RpY2Fs

PjxmdWxsLXRpdGxlPkZvb2QgUG9saWN5PC9mdWxsLXRpdGxlPjwvcGVyaW9kaWNhbD48cGFnZXM+

MzMtNDM8L3BhZ2VzPjx2b2x1bWU+NTI8L3ZvbHVtZT48a2V5d29yZHM+PGtleXdvcmQ+RWNvbm9t

aWMgRGV2ZWxvcG1lbnQ6IEFncmljdWx0dXJlPC9rZXl3b3JkPjxrZXl3b3JkPk5hdHVyYWwgUmVz

b3VyY2VzPC9rZXl3b3JkPjxrZXl3b3JkPkVuZXJneTwva2V5d29yZD48a2V5d29yZD5FbnZpcm9u

bWVudDwva2V5d29yZD48a2V5d29yZD5PdGhlciBQcmltYXJ5IFByb2R1Y3RzIE8xMzwva2V5d29y

ZD48a2V5d29yZD5BZ3JpY3VsdHVyZTogQWdncmVnYXRlIFN1cHBseSBhbmQgRGVtYW5kIEFuYWx5

c2lzPC9rZXl3b3JkPjxrZXl3b3JkPlByaWNlcyBRMTE8L2tleXdvcmQ+PC9rZXl3b3Jkcz48ZGF0

ZXM+PHllYXI+MjAxNTwveWVhcj48L2RhdGVzPjxpc2JuPjAzMDY5MTkyPC9pc2JuPjxhY2Nlc3Np

b24tbnVtPjE0OTAwNjg8L2FjY2Vzc2lvbi1udW0+PHVybHM+PHJlbGF0ZWQtdXJscz48dXJsPmh0

dHA6Ly9zZWFyY2guZWJzY29ob3N0LmNvbS9sb2dpbi5hc3B4P2RpcmVjdD10cnVlJmFtcDtkYj1l

b2gmYW1wO0FOPTE0OTAwNjgmYW1wO3NpdGU9ZWhvc3QtbGl2ZTwvdXJsPjx1cmw+aHR0cDovL2R4

LmRvaS5vcmcvMTAuMTAxNi9qLmZvb2Rwb2wuMjAxNC4wOS4wMTE8L3VybD48dXJsPmh0dHA6Ly93

d3cuc2NpZW5jZWRpcmVjdC5jb20vc2NpZW5jZS9qb3VybmFsLzAzMDY5MTkyPC91cmw+PC9yZWxh

dGVkLXVybHM+PC91cmxzPjxlbGVjdHJvbmljLXJlc291cmNlLW51bT5odHRwOi8vd3d3LnNjaWVu

Y2VkaXJlY3QuY29tL3NjaWVuY2Uvam91cm5hbC8wMzA2OTE5MjwvZWxlY3Ryb25pYy1yZXNvdXJj

ZS1udW0+PHJlbW90ZS1kYXRhYmFzZS1uYW1lPmVvaDwvcmVtb3RlLWRhdGFiYXNlLW5hbWU+PHJl

bW90ZS1kYXRhYmFzZS1wcm92aWRlcj5FQlNDT2hvc3Q8L3JlbW90ZS1kYXRhYmFzZS1wcm92aWRl

cj48L3JlY29yZD48L0NpdGU+PC9FbmROb3RlPgB=

ADDIN EN.CITE.DATA (Brown, Pinzon, and Prince 2008, Higgins, Hintermann, and Brown 2015) and as an index for drought insurance PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5CYWNjaGluaTwvQXV0aG9yPjxZZWFyPjIwMTU8L1llYXI+

PFJlY051bT4yMDwvUmVjTnVtPjxEaXNwbGF5VGV4dD4oQmFjY2hpbmkgYW5kIE1pZ3VleiAyMDE1

LCBNYWthdWR6ZSBhbmQgTWlyYW5kYSAyMDEwKTwvRGlzcGxheVRleHQ+PHJlY29yZD48cmVjLW51

bWJlcj4yMDwvcmVjLW51bWJlcj48Zm9yZWlnbi1rZXlzPjxrZXkgYXBwPSJFTiIgZGItaWQ9Ind4

cHhhd3d6ZXdlZno2ZXNyeDV4ZGEyb3R2ZTI1cHBwNWUwZiIgdGltZXN0YW1wPSIxNTMwMjMzODUz

Ij4yMDwva2V5PjwvZm9yZWlnbi1rZXlzPjxyZWYtdHlwZSBuYW1lPSJKb3VybmFsIEFydGljbGUi

PjE3PC9yZWYtdHlwZT48Y29udHJpYnV0b3JzPjxhdXRob3JzPjxhdXRob3I+QmFjY2hpbmksIFJv

YmVydG8gRGFyaW88L2F1dGhvcj48YXV0aG9yPk1pZ3VleiwgRGFuaWVsIEZlcm5hbmRvPC9hdXRo

b3I+PC9hdXRob3JzPjwvY29udHJpYnV0b3JzPjxhdXRoLWFkZHJlc3M+VSBCdWVub3MgQWlyZXM8

L2F1dGgtYWRkcmVzcz48dGl0bGVzPjx0aXRsZT5BZ3JpY3VsdHVyYWwgUmlzayBNYW5hZ2VtZW50

IFVzaW5nIE5EVkkgUGFzdHVyZSBJbmRleC1CYXNlZCBJbnN1cmFuY2UgZm9yIExpdmVzdG9jayBQ

cm9kdWNlcnMgaW4gU291dGggV2VzdCBCdWVub3MgQWlyZXMgUHJvdmluY2U8L3RpdGxlPjxzZWNv

bmRhcnktdGl0bGU+QWdyaWN1bHR1cmFsIEZpbmFuY2UgUmV2aWV3PC9zZWNvbmRhcnktdGl0bGU+

PC90aXRsZXM+PHBlcmlvZGljYWw+PGZ1bGwtdGl0bGU+QWdyaWN1bHR1cmFsIEZpbmFuY2UgUmV2

aWV3PC9mdWxsLXRpdGxlPjwvcGVyaW9kaWNhbD48cGFnZXM+NzctOTE8L3BhZ2VzPjx2b2x1bWU+

NzU8L3ZvbHVtZT48bnVtYmVyPjE8L251bWJlcj48a2V5d29yZHM+PGtleXdvcmQ+SW5zdXJhbmNl

PC9rZXl3b3JkPjxrZXl3b3JkPkluc3VyYW5jZSBDb21wYW5pZXM8L2tleXdvcmQ+PGtleXdvcmQ+

QWN0dWFyaWFsIFN0dWRpZXMgRzIyPC9rZXl3b3JkPjxrZXl3b3JkPkVjb25vbWljIERldmVsb3Bt

ZW50OiBBZ3JpY3VsdHVyZTwva2V5d29yZD48a2V5d29yZD5OYXR1cmFsIFJlc291cmNlczwva2V5

d29yZD48a2V5d29yZD5FbmVyZ3k8L2tleXdvcmQ+PGtleXdvcmQ+RW52aXJvbm1lbnQ8L2tleXdv

cmQ+PGtleXdvcmQ+T3RoZXIgUHJpbWFyeSBQcm9kdWN0cyBPMTM8L2tleXdvcmQ+PGtleXdvcmQ+

QWdyaWN1bHR1cmU6IEFnZ3JlZ2F0ZSBTdXBwbHkgYW5kIERlbWFuZCBBbmFseXNpczwva2V5d29y

ZD48a2V5d29yZD5QcmljZXMgUTExPC9rZXl3b3JkPjxrZXl3b3JkPk1pY3JvIEFuYWx5c2lzIG9m

IEZhcm0gRmlybXMsIEZhcm0gSG91c2Vob2xkcywgYW5kIEZhcm0gSW5wdXQgTWFya2V0cyBRMTI8

L2tleXdvcmQ+PGtleXdvcmQ+QWdyaWN1bHR1cmFsIEZpbmFuY2UgUTE0PC9rZXl3b3JkPjxrZXl3

b3JkPkxhbmQgT3duZXJzaGlwIGFuZCBUZW51cmU8L2tleXdvcmQ+PGtleXdvcmQ+TGFuZCBSZWZv

cm08L2tleXdvcmQ+PGtleXdvcmQ+TGFuZCBVc2U8L2tleXdvcmQ+PGtleXdvcmQ+SXJyaWdhdGlv

bjwva2V5d29yZD48a2V5d29yZD5BZ3JpY3VsdHVyZSBhbmQgRW52aXJvbm1lbnQgUTE1PC9rZXl3

b3JkPjxrZXl3b3JkPkNsaW1hdGU8L2tleXdvcmQ+PGtleXdvcmQ+TmF0dXJhbCBEaXNhc3RlcnMg

YW5kIFRoZWlyIE1hbmFnZW1lbnQ8L2tleXdvcmQ+PGtleXdvcmQ+R2xvYmFsIFdhcm1pbmcgUTU0

PC9rZXl3b3JkPjwva2V5d29yZHM+PGRhdGVzPjx5ZWFyPjIwMTU8L3llYXI+PC9kYXRlcz48aXNi

bj4wMDAyMTQ2NjwvaXNibj48YWNjZXNzaW9uLW51bT4xNTE4NDg0PC9hY2Nlc3Npb24tbnVtPjx1

cmxzPjxyZWxhdGVkLXVybHM+PHVybD5odHRwOi8vc2VhcmNoLmVic2NvaG9zdC5jb20vbG9naW4u

YXNweD9kaXJlY3Q9dHJ1ZSZhbXA7ZGI9ZW9oJmFtcDtBTj0xNTE4NDg0JmFtcDtzaXRlPWVob3N0

LWxpdmU8L3VybD48dXJsPmh0dHA6Ly93d3cuZW1lcmFsZGluc2lnaHQuY29tL2pvdXJuYWxzLmh0

bT9pc3NuPTAwMDItMTQ2NjwvdXJsPjwvcmVsYXRlZC11cmxzPjwvdXJscz48cmVtb3RlLWRhdGFi

YXNlLW5hbWU+ZW9oPC9yZW1vdGUtZGF0YWJhc2UtbmFtZT48cmVtb3RlLWRhdGFiYXNlLXByb3Zp

ZGVyPkVCU0NPaG9zdDwvcmVtb3RlLWRhdGFiYXNlLXByb3ZpZGVyPjwvcmVjb3JkPjwvQ2l0ZT48

Q2l0ZT48QXV0aG9yPk1ha2F1ZHplPC9BdXRob3I+PFllYXI+MjAxMDwvWWVhcj48UmVjTnVtPjIy

PC9SZWNOdW0+PHJlY29yZD48cmVjLW51bWJlcj4yMjwvcmVjLW51bWJlcj48Zm9yZWlnbi1rZXlz

PjxrZXkgYXBwPSJFTiIgZGItaWQ9Ind4cHhhd3d6ZXdlZno2ZXNyeDV4ZGEyb3R2ZTI1cHBwNWUw

ZiIgdGltZXN0YW1wPSIxNTMwMjMzOTE3Ij4yMjwva2V5PjwvZm9yZWlnbi1rZXlzPjxyZWYtdHlw

ZSBuYW1lPSJKb3VybmFsIEFydGljbGUiPjE3PC9yZWYtdHlwZT48Y29udHJpYnV0b3JzPjxhdXRo

b3JzPjxhdXRob3I+TWFrYXVkemUsIEVwaGlhcyBNLjwvYXV0aG9yPjxhdXRob3I+TWlyYW5kYSwg

TWFyaW8gSi48L2F1dGhvcj48L2F1dGhvcnM+PC9jb250cmlidXRvcnM+PGF1dGgtYWRkcmVzcz5V

IFdlc3Rlcm4gQ2FwZSYjeEQ7T0ggU3RhdGUgVTwvYXV0aC1hZGRyZXNzPjx0aXRsZXM+PHRpdGxl

PkNhdGFzdHJvcGhpYyBEcm91Z2h0IEluc3VyYW5jZSBCYXNlZCBvbiB0aGUgUmVtb3RlbHkgU2Vu

c2VkIE5vcm1hbGlzZWQgRGlmZmVyZW5jZSBWZWdldGF0aW9uIEluZGV4IGZvciBTbWFsbGhvbGRl

ciBGYXJtZXJzIGluIFppbWJhYndlPC90aXRsZT48c2Vjb25kYXJ5LXRpdGxlPkFncmVrb248L3Nl

Y29uZGFyeS10aXRsZT48L3RpdGxlcz48cGVyaW9kaWNhbD48ZnVsbC10aXRsZT5BZ3Jla29uPC9m

dWxsLXRpdGxlPjwvcGVyaW9kaWNhbD48cGFnZXM+NDE4LTQzMjwvcGFnZXM+PHZvbHVtZT40OTwv

dm9sdW1lPjxudW1iZXI+NDwvbnVtYmVyPjxrZXl3b3Jkcz48a2V5d29yZD5JbnN1cmFuY2U8L2tl

eXdvcmQ+PGtleXdvcmQ+SW5zdXJhbmNlIENvbXBhbmllczwva2V5d29yZD48a2V5d29yZD5BY3R1

YXJpYWwgU3R1ZGllcyBHMjI8L2tleXdvcmQ+PGtleXdvcmQ+RWNvbm9taWMgRGV2ZWxvcG1lbnQ6

IEFncmljdWx0dXJlPC9rZXl3b3JkPjxrZXl3b3JkPk5hdHVyYWwgUmVzb3VyY2VzPC9rZXl3b3Jk

PjxrZXl3b3JkPkVuZXJneTwva2V5d29yZD48a2V5d29yZD5FbnZpcm9ubWVudDwva2V5d29yZD48

a2V5d29yZD5PdGhlciBQcmltYXJ5IFByb2R1Y3RzIE8xMzwva2V5d29yZD48a2V5d29yZD5FY29u

b21pYyBEZXZlbG9wbWVudDogRmluYW5jaWFsIE1hcmtldHM8L2tleXdvcmQ+PGtleXdvcmQ+U2F2

aW5nIGFuZCBDYXBpdGFsIEludmVzdG1lbnQ8L2tleXdvcmQ+PGtleXdvcmQ+Q29ycG9yYXRlIEZp

bmFuY2UgYW5kIEdvdmVybmFuY2UgTzE2PC9rZXl3b3JkPjxrZXl3b3JkPk1pY3JvIEFuYWx5c2lz

IG9mIEZhcm0gRmlybXMsIEZhcm0gSG91c2Vob2xkcywgYW5kIEZhcm0gSW5wdXQgTWFya2V0cyBR

MTI8L2tleXdvcmQ+PGtleXdvcmQ+QWdyaWN1bHR1cmFsIEZpbmFuY2UgUTE0PC9rZXl3b3JkPjxr

ZXl3b3JkPkNsaW1hdGU8L2tleXdvcmQ+PGtleXdvcmQ+TmF0dXJhbCBEaXNhc3RlcnMgYW5kIFRo

ZWlyIE1hbmFnZW1lbnQ8L2tleXdvcmQ+PGtleXdvcmQ+R2xvYmFsIFdhcm1pbmcgUTU0PC9rZXl3

b3JkPjwva2V5d29yZHM+PGRhdGVzPjx5ZWFyPjIwMTA8L3llYXI+PC9kYXRlcz48aXNibj4wMzAz

MTg1MzwvaXNibj48YWNjZXNzaW9uLW51bT4xNjIwNTc4PC9hY2Nlc3Npb24tbnVtPjx1cmxzPjxy

ZWxhdGVkLXVybHM+PHVybD5odHRwOi8vc2VhcmNoLmVic2NvaG9zdC5jb20vbG9naW4uYXNweD9k

aXJlY3Q9dHJ1ZSZhbXA7ZGI9ZW9oJmFtcDtBTj0xNjIwNTc4JmFtcDtzaXRlPWVob3N0LWxpdmU8

L3VybD48dXJsPmh0dHA6Ly93d3cudGFuZGZvbmxpbmUuY29tL2xvaS9yYWdyMjAjLlZHejQ0WWZ3

VG9JPC91cmw+PC9yZWxhdGVkLXVybHM+PC91cmxzPjxlbGVjdHJvbmljLXJlc291cmNlLW51bT5o

dHRwOi8vd3d3LnRhbmRmb25saW5lLmNvbS9sb2kvcmFncjIwIy5WR3o0NFlmd1RvSTwvZWxlY3Ry

b25pYy1yZXNvdXJjZS1udW0+PHJlbW90ZS1kYXRhYmFzZS1uYW1lPmVvaDwvcmVtb3RlLWRhdGFi

YXNlLW5hbWU+PHJlbW90ZS1kYXRhYmFzZS1wcm92aWRlcj5FQlNDT2hvc3Q8L3JlbW90ZS1kYXRh

YmFzZS1wcm92aWRlcj48L3JlY29yZD48L0NpdGU+PC9FbmROb3RlPgB=

ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5CYWNjaGluaTwvQXV0aG9yPjxZZWFyPjIwMTU8L1llYXI+

PFJlY051bT4yMDwvUmVjTnVtPjxEaXNwbGF5VGV4dD4oQmFjY2hpbmkgYW5kIE1pZ3VleiAyMDE1

LCBNYWthdWR6ZSBhbmQgTWlyYW5kYSAyMDEwKTwvRGlzcGxheVRleHQ+PHJlY29yZD48cmVjLW51

bWJlcj4yMDwvcmVjLW51bWJlcj48Zm9yZWlnbi1rZXlzPjxrZXkgYXBwPSJFTiIgZGItaWQ9Ind4

cHhhd3d6ZXdlZno2ZXNyeDV4ZGEyb3R2ZTI1cHBwNWUwZiIgdGltZXN0YW1wPSIxNTMwMjMzODUz

Ij4yMDwva2V5PjwvZm9yZWlnbi1rZXlzPjxyZWYtdHlwZSBuYW1lPSJKb3VybmFsIEFydGljbGUi

PjE3PC9yZWYtdHlwZT48Y29udHJpYnV0b3JzPjxhdXRob3JzPjxhdXRob3I+QmFjY2hpbmksIFJv

YmVydG8gRGFyaW88L2F1dGhvcj48YXV0aG9yPk1pZ3VleiwgRGFuaWVsIEZlcm5hbmRvPC9hdXRo

b3I+PC9hdXRob3JzPjwvY29udHJpYnV0b3JzPjxhdXRoLWFkZHJlc3M+VSBCdWVub3MgQWlyZXM8

L2F1dGgtYWRkcmVzcz48dGl0bGVzPjx0aXRsZT5BZ3JpY3VsdHVyYWwgUmlzayBNYW5hZ2VtZW50

IFVzaW5nIE5EVkkgUGFzdHVyZSBJbmRleC1CYXNlZCBJbnN1cmFuY2UgZm9yIExpdmVzdG9jayBQ

cm9kdWNlcnMgaW4gU291dGggV2VzdCBCdWVub3MgQWlyZXMgUHJvdmluY2U8L3RpdGxlPjxzZWNv

bmRhcnktdGl0bGU+QWdyaWN1bHR1cmFsIEZpbmFuY2UgUmV2aWV3PC9zZWNvbmRhcnktdGl0bGU+

PC90aXRsZXM+PHBlcmlvZGljYWw+PGZ1bGwtdGl0bGU+QWdyaWN1bHR1cmFsIEZpbmFuY2UgUmV2

aWV3PC9mdWxsLXRpdGxlPjwvcGVyaW9kaWNhbD48cGFnZXM+NzctOTE8L3BhZ2VzPjx2b2x1bWU+

NzU8L3ZvbHVtZT48bnVtYmVyPjE8L251bWJlcj48a2V5d29yZHM+PGtleXdvcmQ+SW5zdXJhbmNl

PC9rZXl3b3JkPjxrZXl3b3JkPkluc3VyYW5jZSBDb21wYW5pZXM8L2tleXdvcmQ+PGtleXdvcmQ+

QWN0dWFyaWFsIFN0dWRpZXMgRzIyPC9rZXl3b3JkPjxrZXl3b3JkPkVjb25vbWljIERldmVsb3Bt

ZW50OiBBZ3JpY3VsdHVyZTwva2V5d29yZD48a2V5d29yZD5OYXR1cmFsIFJlc291cmNlczwva2V5

d29yZD48a2V5d29yZD5FbmVyZ3k8L2tleXdvcmQ+PGtleXdvcmQ+RW52aXJvbm1lbnQ8L2tleXdv

cmQ+PGtleXdvcmQ+T3RoZXIgUHJpbWFyeSBQcm9kdWN0cyBPMTM8L2tleXdvcmQ+PGtleXdvcmQ+

QWdyaWN1bHR1cmU6IEFnZ3JlZ2F0ZSBTdXBwbHkgYW5kIERlbWFuZCBBbmFseXNpczwva2V5d29y

ZD48a2V5d29yZD5QcmljZXMgUTExPC9rZXl3b3JkPjxrZXl3b3JkPk1pY3JvIEFuYWx5c2lzIG9m

IEZhcm0gRmlybXMsIEZhcm0gSG91c2Vob2xkcywgYW5kIEZhcm0gSW5wdXQgTWFya2V0cyBRMTI8

L2tleXdvcmQ+PGtleXdvcmQ+QWdyaWN1bHR1cmFsIEZpbmFuY2UgUTE0PC9rZXl3b3JkPjxrZXl3

b3JkPkxhbmQgT3duZXJzaGlwIGFuZCBUZW51cmU8L2tleXdvcmQ+PGtleXdvcmQ+TGFuZCBSZWZv

cm08L2tleXdvcmQ+PGtleXdvcmQ+TGFuZCBVc2U8L2tleXdvcmQ+PGtleXdvcmQ+SXJyaWdhdGlv

bjwva2V5d29yZD48a2V5d29yZD5BZ3JpY3VsdHVyZSBhbmQgRW52aXJvbm1lbnQgUTE1PC9rZXl3

b3JkPjxrZXl3b3JkPkNsaW1hdGU8L2tleXdvcmQ+PGtleXdvcmQ+TmF0dXJhbCBEaXNhc3RlcnMg

YW5kIFRoZWlyIE1hbmFnZW1lbnQ8L2tleXdvcmQ+PGtleXdvcmQ+R2xvYmFsIFdhcm1pbmcgUTU0

PC9rZXl3b3JkPjwva2V5d29yZHM+PGRhdGVzPjx5ZWFyPjIwMTU8L3llYXI+PC9kYXRlcz48aXNi

bj4wMDAyMTQ2NjwvaXNibj48YWNjZXNzaW9uLW51bT4xNTE4NDg0PC9hY2Nlc3Npb24tbnVtPjx1

cmxzPjxyZWxhdGVkLXVybHM+PHVybD5odHRwOi8vc2VhcmNoLmVic2NvaG9zdC5jb20vbG9naW4u

YXNweD9kaXJlY3Q9dHJ1ZSZhbXA7ZGI9ZW9oJmFtcDtBTj0xNTE4NDg0JmFtcDtzaXRlPWVob3N0

LWxpdmU8L3VybD48dXJsPmh0dHA6Ly93d3cuZW1lcmFsZGluc2lnaHQuY29tL2pvdXJuYWxzLmh0

bT9pc3NuPTAwMDItMTQ2NjwvdXJsPjwvcmVsYXRlZC11cmxzPjwvdXJscz48cmVtb3RlLWRhdGFi

YXNlLW5hbWU+ZW9oPC9yZW1vdGUtZGF0YWJhc2UtbmFtZT48cmVtb3RlLWRhdGFiYXNlLXByb3Zp

ZGVyPkVCU0NPaG9zdDwvcmVtb3RlLWRhdGFiYXNlLXByb3ZpZGVyPjwvcmVjb3JkPjwvQ2l0ZT48

Q2l0ZT48QXV0aG9yPk1ha2F1ZHplPC9BdXRob3I+PFllYXI+MjAxMDwvWWVhcj48UmVjTnVtPjIy

PC9SZWNOdW0+PHJlY29yZD48cmVjLW51bWJlcj4yMjwvcmVjLW51bWJlcj48Zm9yZWlnbi1rZXlz

PjxrZXkgYXBwPSJFTiIgZGItaWQ9Ind4cHhhd3d6ZXdlZno2ZXNyeDV4ZGEyb3R2ZTI1cHBwNWUw

ZiIgdGltZXN0YW1wPSIxNTMwMjMzOTE3Ij4yMjwva2V5PjwvZm9yZWlnbi1rZXlzPjxyZWYtdHlw

ZSBuYW1lPSJKb3VybmFsIEFydGljbGUiPjE3PC9yZWYtdHlwZT48Y29udHJpYnV0b3JzPjxhdXRo

b3JzPjxhdXRob3I+TWFrYXVkemUsIEVwaGlhcyBNLjwvYXV0aG9yPjxhdXRob3I+TWlyYW5kYSwg

TWFyaW8gSi48L2F1dGhvcj48L2F1dGhvcnM+PC9jb250cmlidXRvcnM+PGF1dGgtYWRkcmVzcz5V

IFdlc3Rlcm4gQ2FwZSYjeEQ7T0ggU3RhdGUgVTwvYXV0aC1hZGRyZXNzPjx0aXRsZXM+PHRpdGxl

PkNhdGFzdHJvcGhpYyBEcm91Z2h0IEluc3VyYW5jZSBCYXNlZCBvbiB0aGUgUmVtb3RlbHkgU2Vu

c2VkIE5vcm1hbGlzZWQgRGlmZmVyZW5jZSBWZWdldGF0aW9uIEluZGV4IGZvciBTbWFsbGhvbGRl

ciBGYXJtZXJzIGluIFppbWJhYndlPC90aXRsZT48c2Vjb25kYXJ5LXRpdGxlPkFncmVrb248L3Nl

Y29uZGFyeS10aXRsZT48L3RpdGxlcz48cGVyaW9kaWNhbD48ZnVsbC10aXRsZT5BZ3Jla29uPC9m

dWxsLXRpdGxlPjwvcGVyaW9kaWNhbD48cGFnZXM+NDE4LTQzMjwvcGFnZXM+PHZvbHVtZT40OTwv

dm9sdW1lPjxudW1iZXI+NDwvbnVtYmVyPjxrZXl3b3Jkcz48a2V5d29yZD5JbnN1cmFuY2U8L2tl

eXdvcmQ+PGtleXdvcmQ+SW5zdXJhbmNlIENvbXBhbmllczwva2V5d29yZD48a2V5d29yZD5BY3R1

YXJpYWwgU3R1ZGllcyBHMjI8L2tleXdvcmQ+PGtleXdvcmQ+RWNvbm9taWMgRGV2ZWxvcG1lbnQ6

IEFncmljdWx0dXJlPC9rZXl3b3JkPjxrZXl3b3JkPk5hdHVyYWwgUmVzb3VyY2VzPC9rZXl3b3Jk

PjxrZXl3b3JkPkVuZXJneTwva2V5d29yZD48a2V5d29yZD5FbnZpcm9ubWVudDwva2V5d29yZD48

a2V5d29yZD5PdGhlciBQcmltYXJ5IFByb2R1Y3RzIE8xMzwva2V5d29yZD48a2V5d29yZD5FY29u

b21pYyBEZXZlbG9wbWVudDogRmluYW5jaWFsIE1hcmtldHM8L2tleXdvcmQ+PGtleXdvcmQ+U2F2

aW5nIGFuZCBDYXBpdGFsIEludmVzdG1lbnQ8L2tleXdvcmQ+PGtleXdvcmQ+Q29ycG9yYXRlIEZp

bmFuY2UgYW5kIEdvdmVybmFuY2UgTzE2PC9rZXl3b3JkPjxrZXl3b3JkPk1pY3JvIEFuYWx5c2lz

IG9mIEZhcm0gRmlybXMsIEZhcm0gSG91c2Vob2xkcywgYW5kIEZhcm0gSW5wdXQgTWFya2V0cyBR

MTI8L2tleXdvcmQ+PGtleXdvcmQ+QWdyaWN1bHR1cmFsIEZpbmFuY2UgUTE0PC9rZXl3b3JkPjxr

ZXl3b3JkPkNsaW1hdGU8L2tleXdvcmQ+PGtleXdvcmQ+TmF0dXJhbCBEaXNhc3RlcnMgYW5kIFRo

ZWlyIE1hbmFnZW1lbnQ8L2tleXdvcmQ+PGtleXdvcmQ+R2xvYmFsIFdhcm1pbmcgUTU0PC9rZXl3

b3JkPjwva2V5d29yZHM+PGRhdGVzPjx5ZWFyPjIwMTA8L3llYXI+PC9kYXRlcz48aXNibj4wMzAz

MTg1MzwvaXNibj48YWNjZXNzaW9uLW51bT4xNjIwNTc4PC9hY2Nlc3Npb24tbnVtPjx1cmxzPjxy

ZWxhdGVkLXVybHM+PHVybD5odHRwOi8vc2VhcmNoLmVic2NvaG9zdC5jb20vbG9naW4uYXNweD9k

aXJlY3Q9dHJ1ZSZhbXA7ZGI9ZW9oJmFtcDtBTj0xNjIwNTc4JmFtcDtzaXRlPWVob3N0LWxpdmU8

L3VybD48dXJsPmh0dHA6Ly93d3cudGFuZGZvbmxpbmUuY29tL2xvaS9yYWdyMjAjLlZHejQ0WWZ3

VG9JPC91cmw+PC9yZWxhdGVkLXVybHM+PC91cmxzPjxlbGVjdHJvbmljLXJlc291cmNlLW51bT5o

dHRwOi8vd3d3LnRhbmRmb25saW5lLmNvbS9sb2kvcmFncjIwIy5WR3o0NFlmd1RvSTwvZWxlY3Ry

b25pYy1yZXNvdXJjZS1udW0+PHJlbW90ZS1kYXRhYmFzZS1uYW1lPmVvaDwvcmVtb3RlLWRhdGFi

YXNlLW5hbWU+PHJlbW90ZS1kYXRhYmFzZS1wcm92aWRlcj5FQlNDT2hvc3Q8L3JlbW90ZS1kYXRh

YmFzZS1wcm92aWRlcj48L3JlY29yZD48L0NpdGU+PC9FbmROb3RlPgB=

ADDIN EN.CITE.DATA (Bacchini and Miguez 2015, Makaudze and Miranda 2010). NDVI is correlated with the probability of child malnourishment PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5CYXVlcjwvQXV0aG9yPjxZZWFyPjIwMTc8L1llYXI+PFJl

Y051bT4xNjwvUmVjTnVtPjxEaXNwbGF5VGV4dD4oQmF1ZXIgYW5kIE1idXJ1IDIwMTcpPC9EaXNw

bGF5VGV4dD48cmVjb3JkPjxyZWMtbnVtYmVyPjE2PC9yZWMtbnVtYmVyPjxmb3JlaWduLWtleXM+

PGtleSBhcHA9IkVOIiBkYi1pZD0id3hweGF3d3pld2VmejZlc3J4NXhkYTJvdHZlMjVwcHA1ZTBm

IiB0aW1lc3RhbXA9IjE1MzAyMzM2OTQiPjE2PC9rZXk+PC9mb3JlaWduLWtleXM+PHJlZi10eXBl

IG5hbWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3JlZi10eXBlPjxjb250cmlidXRvcnM+PGF1dGhv

cnM+PGF1dGhvcj5CYXVlciwgSmFuIE0uPC9hdXRob3I+PGF1dGhvcj5NYnVydSwgU2FtdWVsPC9h

dXRob3I+PC9hdXRob3JzPjwvY29udHJpYnV0b3JzPjxhdXRoLWFkZHJlc3M+Q29wZW5oYWdlbiBC

dXNpbmVzcyBTY2hvb2wgYW5kIFUgSG9oZW5oZWltJiN4RDtVIEhvaGVuaGVpbTwvYXV0aC1hZGRy

ZXNzPjx0aXRsZXM+PHRpdGxlPkVmZmVjdHMgb2YgRHJvdWdodCBvbiBDaGlsZCBIZWFsdGggaW4g

TWFyc2FiaXQgRGlzdHJpY3QsIE5vcnRoZXJuIEtlbnlhPC90aXRsZT48c2Vjb25kYXJ5LXRpdGxl

PkVjb25vbWljcyBhbmQgSHVtYW4gQmlvbG9neTwvc2Vjb25kYXJ5LXRpdGxlPjwvdGl0bGVzPjxw

ZXJpb2RpY2FsPjxmdWxsLXRpdGxlPkVjb25vbWljcyBhbmQgSHVtYW4gQmlvbG9neTwvZnVsbC10

aXRsZT48L3BlcmlvZGljYWw+PHBhZ2VzPjc0LTc5PC9wYWdlcz48dm9sdW1lPjI0PC92b2x1bWU+

PGtleXdvcmRzPjxrZXl3b3JkPkhlYWx0aCBCZWhhdmlvciBJMTI8L2tleXdvcmQ+PGtleXdvcmQ+

RmVydGlsaXR5PC9rZXl3b3JkPjxrZXl3b3JkPkZhbWlseSBQbGFubmluZzwva2V5d29yZD48a2V5

d29yZD5DaGlsZCBDYXJlPC9rZXl3b3JkPjxrZXl3b3JkPkNoaWxkcmVuPC9rZXl3b3JkPjxrZXl3

b3JkPllvdXRoIEoxMzwva2V5d29yZD48a2V5d29yZD5FY29ub21pYyBEZXZlbG9wbWVudDogQWdy

aWN1bHR1cmU8L2tleXdvcmQ+PGtleXdvcmQ+TmF0dXJhbCBSZXNvdXJjZXM8L2tleXdvcmQ+PGtl

eXdvcmQ+RW5lcmd5PC9rZXl3b3JkPjxrZXl3b3JkPkVudmlyb25tZW50PC9rZXl3b3JkPjxrZXl3

b3JkPk90aGVyIFByaW1hcnkgUHJvZHVjdHMgTzEzPC9rZXl3b3JkPjxrZXl3b3JkPkVjb25vbWlj

IERldmVsb3BtZW50OiBIdW1hbiBSZXNvdXJjZXM8L2tleXdvcmQ+PGtleXdvcmQ+SHVtYW4gRGV2

ZWxvcG1lbnQ8L2tleXdvcmQ+PGtleXdvcmQ+SW5jb21lIERpc3RyaWJ1dGlvbjwva2V5d29yZD48

a2V5d29yZD5NaWdyYXRpb24gTzE1PC9rZXl3b3JkPjxrZXl3b3JkPkVjb25vbWljIERldmVsb3Bt

ZW50OiBVcmJhbiwgUnVyYWwsIFJlZ2lvbmFsLCBhbmQgVHJhbnNwb3J0YXRpb24gQW5hbHlzaXM8

L2tleXdvcmQ+PGtleXdvcmQ+SG91c2luZzwva2V5d29yZD48a2V5d29yZD5JbmZyYXN0cnVjdHVy

ZSBPMTg8L2tleXdvcmQ+PGtleXdvcmQ+QWdyaWN1bHR1cmFsIFBvbGljeTwva2V5d29yZD48a2V5

d29yZD5Gb29kIFBvbGljeSBRMTg8L2tleXdvcmQ+PC9rZXl3b3Jkcz48ZGF0ZXM+PHllYXI+MjAx

NzwveWVhcj48L2RhdGVzPjxpc2JuPjE1NzA2NzdYPC9pc2JuPjxhY2Nlc3Npb24tbnVtPjE2NDY3

NDI8L2FjY2Vzc2lvbi1udW0+PHVybHM+PHJlbGF0ZWQtdXJscz48dXJsPmh0dHA6Ly9zZWFyY2gu

ZWJzY29ob3N0LmNvbS9sb2dpbi5hc3B4P2RpcmVjdD10cnVlJmFtcDtkYj1lb2gmYW1wO0FOPTE2

NDY3NDImYW1wO3NpdGU9ZWhvc3QtbGl2ZTwvdXJsPjx1cmw+aHR0cDovL2R4LmRvaS5vcmcvMTAu

MTAxNi9qLmVoYi4yMDE2LjEwLjAxMDwvdXJsPjx1cmw+aHR0cDovL3d3dy5zY2llbmNlZGlyZWN0

LmNvbS9zY2llbmNlL2pvdXJuYWwvMTU3MDY3N1g8L3VybD48L3JlbGF0ZWQtdXJscz48L3VybHM+

PGVsZWN0cm9uaWMtcmVzb3VyY2UtbnVtPmh0dHA6Ly93d3cuc2NpZW5jZWRpcmVjdC5jb20vc2Np

ZW5jZS9qb3VybmFsLzE1NzA2NzdYPC9lbGVjdHJvbmljLXJlc291cmNlLW51bT48cmVtb3RlLWRh

dGFiYXNlLW5hbWU+ZW9oPC9yZW1vdGUtZGF0YWJhc2UtbmFtZT48cmVtb3RlLWRhdGFiYXNlLXBy

b3ZpZGVyPkVCU0NPaG9zdDwvcmVtb3RlLWRhdGFiYXNlLXByb3ZpZGVyPjwvcmVjb3JkPjwvQ2l0

ZT48L0VuZE5vdGU+

ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5CYXVlcjwvQXV0aG9yPjxZZWFyPjIwMTc8L1llYXI+PFJl

Y051bT4xNjwvUmVjTnVtPjxEaXNwbGF5VGV4dD4oQmF1ZXIgYW5kIE1idXJ1IDIwMTcpPC9EaXNw

bGF5VGV4dD48cmVjb3JkPjxyZWMtbnVtYmVyPjE2PC9yZWMtbnVtYmVyPjxmb3JlaWduLWtleXM+

PGtleSBhcHA9IkVOIiBkYi1pZD0id3hweGF3d3pld2VmejZlc3J4NXhkYTJvdHZlMjVwcHA1ZTBm

IiB0aW1lc3RhbXA9IjE1MzAyMzM2OTQiPjE2PC9rZXk+PC9mb3JlaWduLWtleXM+PHJlZi10eXBl

IG5hbWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3JlZi10eXBlPjxjb250cmlidXRvcnM+PGF1dGhv

cnM+PGF1dGhvcj5CYXVlciwgSmFuIE0uPC9hdXRob3I+PGF1dGhvcj5NYnVydSwgU2FtdWVsPC9h

dXRob3I+PC9hdXRob3JzPjwvY29udHJpYnV0b3JzPjxhdXRoLWFkZHJlc3M+Q29wZW5oYWdlbiBC

dXNpbmVzcyBTY2hvb2wgYW5kIFUgSG9oZW5oZWltJiN4RDtVIEhvaGVuaGVpbTwvYXV0aC1hZGRy

ZXNzPjx0aXRsZXM+PHRpdGxlPkVmZmVjdHMgb2YgRHJvdWdodCBvbiBDaGlsZCBIZWFsdGggaW4g

TWFyc2FiaXQgRGlzdHJpY3QsIE5vcnRoZXJuIEtlbnlhPC90aXRsZT48c2Vjb25kYXJ5LXRpdGxl

PkVjb25vbWljcyBhbmQgSHVtYW4gQmlvbG9neTwvc2Vjb25kYXJ5LXRpdGxlPjwvdGl0bGVzPjxw

ZXJpb2RpY2FsPjxmdWxsLXRpdGxlPkVjb25vbWljcyBhbmQgSHVtYW4gQmlvbG9neTwvZnVsbC10

aXRsZT48L3BlcmlvZGljYWw+PHBhZ2VzPjc0LTc5PC9wYWdlcz48dm9sdW1lPjI0PC92b2x1bWU+

PGtleXdvcmRzPjxrZXl3b3JkPkhlYWx0aCBCZWhhdmlvciBJMTI8L2tleXdvcmQ+PGtleXdvcmQ+

RmVydGlsaXR5PC9rZXl3b3JkPjxrZXl3b3JkPkZhbWlseSBQbGFubmluZzwva2V5d29yZD48a2V5

d29yZD5DaGlsZCBDYXJlPC9rZXl3b3JkPjxrZXl3b3JkPkNoaWxkcmVuPC9rZXl3b3JkPjxrZXl3

b3JkPllvdXRoIEoxMzwva2V5d29yZD48a2V5d29yZD5FY29ub21pYyBEZXZlbG9wbWVudDogQWdy

aWN1bHR1cmU8L2tleXdvcmQ+PGtleXdvcmQ+TmF0dXJhbCBSZXNvdXJjZXM8L2tleXdvcmQ+PGtl

eXdvcmQ+RW5lcmd5PC9rZXl3b3JkPjxrZXl3b3JkPkVudmlyb25tZW50PC9rZXl3b3JkPjxrZXl3

b3JkPk90aGVyIFByaW1hcnkgUHJvZHVjdHMgTzEzPC9rZXl3b3JkPjxrZXl3b3JkPkVjb25vbWlj

IERldmVsb3BtZW50OiBIdW1hbiBSZXNvdXJjZXM8L2tleXdvcmQ+PGtleXdvcmQ+SHVtYW4gRGV2

ZWxvcG1lbnQ8L2tleXdvcmQ+PGtleXdvcmQ+SW5jb21lIERpc3RyaWJ1dGlvbjwva2V5d29yZD48

a2V5d29yZD5NaWdyYXRpb24gTzE1PC9rZXl3b3JkPjxrZXl3b3JkPkVjb25vbWljIERldmVsb3Bt

ZW50OiBVcmJhbiwgUnVyYWwsIFJlZ2lvbmFsLCBhbmQgVHJhbnNwb3J0YXRpb24gQW5hbHlzaXM8

L2tleXdvcmQ+PGtleXdvcmQ+SG91c2luZzwva2V5d29yZD48a2V5d29yZD5JbmZyYXN0cnVjdHVy

ZSBPMTg8L2tleXdvcmQ+PGtleXdvcmQ+QWdyaWN1bHR1cmFsIFBvbGljeTwva2V5d29yZD48a2V5

d29yZD5Gb29kIFBvbGljeSBRMTg8L2tleXdvcmQ+PC9rZXl3b3Jkcz48ZGF0ZXM+PHllYXI+MjAx

NzwveWVhcj48L2RhdGVzPjxpc2JuPjE1NzA2NzdYPC9pc2JuPjxhY2Nlc3Npb24tbnVtPjE2NDY3

NDI8L2FjY2Vzc2lvbi1udW0+PHVybHM+PHJlbGF0ZWQtdXJscz48dXJsPmh0dHA6Ly9zZWFyY2gu

ZWJzY29ob3N0LmNvbS9sb2dpbi5hc3B4P2RpcmVjdD10cnVlJmFtcDtkYj1lb2gmYW1wO0FOPTE2

NDY3NDImYW1wO3NpdGU9ZWhvc3QtbGl2ZTwvdXJsPjx1cmw+aHR0cDovL2R4LmRvaS5vcmcvMTAu

MTAxNi9qLmVoYi4yMDE2LjEwLjAxMDwvdXJsPjx1cmw+aHR0cDovL3d3dy5zY2llbmNlZGlyZWN0

LmNvbS9zY2llbmNlL2pvdXJuYWwvMTU3MDY3N1g8L3VybD48L3JlbGF0ZWQtdXJscz48L3VybHM+

PGVsZWN0cm9uaWMtcmVzb3VyY2UtbnVtPmh0dHA6Ly93d3cuc2NpZW5jZWRpcmVjdC5jb20vc2Np

ZW5jZS9qb3VybmFsLzE1NzA2NzdYPC9lbGVjdHJvbmljLXJlc291cmNlLW51bT48cmVtb3RlLWRh

dGFiYXNlLW5hbWU+ZW9oPC9yZW1vdGUtZGF0YWJhc2UtbmFtZT48cmVtb3RlLWRhdGFiYXNlLXBy

b3ZpZGVyPkVCU0NPaG9zdDwvcmVtb3RlLWRhdGFiYXNlLXByb3ZpZGVyPjwvcmVjb3JkPjwvQ2l0

ZT48L0VuZE5vdGU+

ADDIN EN.CITE.DATA (Bauer and Mburu 2017).Net Primary Productivity (NPP) NPP is the total amount of carbon dioxide taken in by plants. It is measured as the difference between how much carbon dioxide is taken in by plants and how much carbon dioxide is given out by plants. Carbon dioxide contributes to warming of the planet and it is therefore important to track plant productivity to understand where carbon dioxide comes from and where it goes. The change in NPP over time is often used as a measure of land degradation. The losses of NPP can be caused by human-induced dryland degradation ADDIN EN.CITE <EndNote><Cite><Author>Zika</Author><Year>2009</Year><RecNum>13</RecNum><DisplayText>(Zika and Erb 2009)</DisplayText><record><rec-number>13</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1530068526">13</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Zika, Michael</author><author>Erb, Karl-Heinz</author></authors></contributors><titles><title>The global loss of net primary production resulting from human-induced soil degradation in drylands</title><secondary-title>Ecological Economics</secondary-title></titles><periodical><full-title>Ecological Economics</full-title></periodical><pages>310-318</pages><volume>69</volume><number>2</number><dates><year>2009</year></dates><isbn>0921-8009</isbn><urls></urls></record></Cite></EndNote>(Zika and Erb 2009). Urban areas tend to have considerably higher NPP ADDIN EN.CITE <EndNote><Cite><Author>O&apos;Neill</Author><Year>2009</Year><RecNum>29</RecNum><DisplayText>(O&apos;Neill and Abson 2009)</DisplayText><record><rec-number>29</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1530234682">29</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>O&apos;Neill, Daniel W.</author><author>Abson, David J.</author></authors></contributors><auth-address>U Leeds</auth-address><titles><title>To Settle or Protect? A Global Analysis of Net Primary Production in Parks and Urban Areas</title><secondary-title>Ecological Economics</secondary-title></titles><periodical><full-title>Ecological Economics</full-title></periodical><pages>319-327</pages><volume>69</volume><number>2</number><keywords><keyword>Economic Development: Agriculture</keyword><keyword>Natural Resources</keyword><keyword>Energy</keyword><keyword>Environment</keyword><keyword>Other Primary Products O13</keyword><keyword>Renewable Resources and Conservation: Land Q24</keyword><keyword>Recreational Aspects of Natural Resources Q26</keyword><keyword>Ecological Economics: Ecosystem Services</keyword><keyword>Biodiversity Conservation</keyword><keyword>Bioeconomics</keyword><keyword>Industrial Ecology Q57</keyword></keywords><dates><year>2009</year></dates><isbn>09218009</isbn><accession-num>1090728</accession-num><urls><related-urls><url>;(O'Neill and Abson 2009), as humans often build settlements in areas of high biological productivity. In general, NPP is similar to NDVI in that they are both measures of vegetation density, however, NPP is usually a better measure of biomass productivity ADDIN EN.CITE <EndNote><Cite><Author>Xu</Author><Year>2012</Year><RecNum>178</RecNum><DisplayText>(Xu et al. 2012)</DisplayText><record><rec-number>178</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1538155845">178</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Xu, Chi</author><author>Li, Yutong</author><author>Hu, Jian</author><author>Yang, Xuejiao</author><author>Sheng, Sheng</author><author>Liu, Maosong</author></authors></contributors><titles><title>Evaluating the difference between the normalized difference vegetation index and net primary productivity as the indicators of vegetation vigor assessment at landscape scale</title><secondary-title>Environmental monitoring and assessment</secondary-title></titles><periodical><full-title>Environmental monitoring and assessment</full-title></periodical><pages>1275-1286</pages><volume>184</volume><number>3</number><dates><year>2012</year></dates><isbn>0167-6369</isbn><urls></urls></record></Cite></EndNote>(Xu et al. 2012).The unit of NPP is grams of carbon per m2 per day. The data is captured with Moderate Resolution Imaging Spectroradiometer (MODIS) aboard NASA’s Terra and Aqua satellites. This paper uses the mean NPP at the 110-district level in 2000 and 2010. The NPP for each pixel is averaged across all grid cells within each district to obtain the mean NPP. Night Time Lights (NTL)NTL is often used to study the distribution of economic activities PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5IZW5kZXJzb248L0F1dGhvcj48WWVhcj4yMDE3PC9ZZWFy

PjxSZWNOdW0+MTU1PC9SZWNOdW0+PERpc3BsYXlUZXh0PihIZW5kZXJzb24gZXQgYWwuIDIwMTcs

IE1pY2hhbG9wb3Vsb3MgYW5kIFBhcGFpb2Fubm91IDIwMTMsIENoZW4gYW5kIE5vcmRoYXVzIDIw

MTEpPC9EaXNwbGF5VGV4dD48cmVjb3JkPjxyZWMtbnVtYmVyPjE1NTwvcmVjLW51bWJlcj48Zm9y

ZWlnbi1rZXlzPjxrZXkgYXBwPSJFTiIgZGItaWQ9Ind4cHhhd3d6ZXdlZno2ZXNyeDV4ZGEyb3R2

ZTI1cHBwNWUwZiIgdGltZXN0YW1wPSIxNTM3NzY0NzE0Ij4xNTU8L2tleT48L2ZvcmVpZ24ta2V5

cz48cmVmLXR5cGUgbmFtZT0iSm91cm5hbCBBcnRpY2xlIj4xNzwvcmVmLXR5cGU+PGNvbnRyaWJ1

dG9ycz48YXV0aG9ycz48YXV0aG9yPkhlbmRlcnNvbiwgSiBWZXJub248L2F1dGhvcj48YXV0aG9y

PlNxdWlyZXMsIFRpbTwvYXV0aG9yPjxhdXRob3I+U3RvcmV5Z2FyZCwgQWRhbTwvYXV0aG9yPjxh

dXRob3I+V2VpbCwgRGF2aWQ8L2F1dGhvcj48L2F1dGhvcnM+PC9jb250cmlidXRvcnM+PHRpdGxl

cz48dGl0bGU+VGhlIGdsb2JhbCBkaXN0cmlidXRpb24gb2YgZWNvbm9taWMgYWN0aXZpdHk6IG5h

dHVyZSwgaGlzdG9yeSwgYW5kIHRoZSByb2xlIG9mIHRyYWRlPC90aXRsZT48c2Vjb25kYXJ5LXRp

dGxlPlRoZSBRdWFydGVybHkgSm91cm5hbCBvZiBFY29ub21pY3M8L3NlY29uZGFyeS10aXRsZT48

L3RpdGxlcz48cGVyaW9kaWNhbD48ZnVsbC10aXRsZT5UaGUgUXVhcnRlcmx5IEpvdXJuYWwgb2Yg

RWNvbm9taWNzPC9mdWxsLXRpdGxlPjwvcGVyaW9kaWNhbD48cGFnZXM+MzU3LTQwNjwvcGFnZXM+

PHZvbHVtZT4xMzM8L3ZvbHVtZT48bnVtYmVyPjE8L251bWJlcj48ZGF0ZXM+PHllYXI+MjAxNzwv

eWVhcj48L2RhdGVzPjxpc2JuPjAwMzMtNTUzMzwvaXNibj48dXJscz48L3VybHM+PC9yZWNvcmQ+

PC9DaXRlPjxDaXRlPjxBdXRob3I+TWljaGFsb3BvdWxvczwvQXV0aG9yPjxZZWFyPjIwMTM8L1ll

YXI+PFJlY051bT4xNjg8L1JlY051bT48cmVjb3JkPjxyZWMtbnVtYmVyPjE2ODwvcmVjLW51bWJl

cj48Zm9yZWlnbi1rZXlzPjxrZXkgYXBwPSJFTiIgZGItaWQ9Ind4cHhhd3d6ZXdlZno2ZXNyeDV4

ZGEyb3R2ZTI1cHBwNWUwZiIgdGltZXN0YW1wPSIxNTM3NzY1MjM1Ij4xNjg8L2tleT48L2ZvcmVp

Z24ta2V5cz48cmVmLXR5cGUgbmFtZT0iSm91cm5hbCBBcnRpY2xlIj4xNzwvcmVmLXR5cGU+PGNv

bnRyaWJ1dG9ycz48YXV0aG9ycz48YXV0aG9yPk1pY2hhbG9wb3Vsb3MsIFN0ZWxpb3M8L2F1dGhv

cj48YXV0aG9yPlBhcGFpb2Fubm91LCBFbGlhczwvYXV0aG9yPjwvYXV0aG9ycz48L2NvbnRyaWJ1

dG9ycz48dGl0bGVzPjx0aXRsZT5OYXRpb25hbCBpbnN0aXR1dGlvbnMgYW5kIHN1Ym5hdGlvbmFs

IGRldmVsb3BtZW50IGluIEFmcmljYTwvdGl0bGU+PHNlY29uZGFyeS10aXRsZT5UaGUgUXVhcnRl

cmx5IEpvdXJuYWwgb2YgRWNvbm9taWNzPC9zZWNvbmRhcnktdGl0bGU+PC90aXRsZXM+PHBlcmlv

ZGljYWw+PGZ1bGwtdGl0bGU+VGhlIFF1YXJ0ZXJseSBKb3VybmFsIG9mIEVjb25vbWljczwvZnVs

bC10aXRsZT48L3BlcmlvZGljYWw+PHBhZ2VzPjE1MS0yMTM8L3BhZ2VzPjx2b2x1bWU+MTI5PC92

b2x1bWU+PG51bWJlcj4xPC9udW1iZXI+PGRhdGVzPjx5ZWFyPjIwMTM8L3llYXI+PC9kYXRlcz48

aXNibj4wMDMzLTU1MzM8L2lzYm4+PHVybHM+PC91cmxzPjwvcmVjb3JkPjwvQ2l0ZT48Q2l0ZT48

QXV0aG9yPkNoZW48L0F1dGhvcj48WWVhcj4yMDExPC9ZZWFyPjxSZWNOdW0+MTczPC9SZWNOdW0+

PHJlY29yZD48cmVjLW51bWJlcj4xNzM8L3JlYy1udW1iZXI+PGZvcmVpZ24ta2V5cz48a2V5IGFw

cD0iRU4iIGRiLWlkPSJ3eHB4YXd3emV3ZWZ6NmVzcng1eGRhMm90dmUyNXBwcDVlMGYiIHRpbWVz

dGFtcD0iMTUzNzc2NTQzMiI+MTczPC9rZXk+PC9mb3JlaWduLWtleXM+PHJlZi10eXBlIG5hbWU9

IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3JlZi10eXBlPjxjb250cmlidXRvcnM+PGF1dGhvcnM+PGF1

dGhvcj5DaGVuLCBYaTwvYXV0aG9yPjxhdXRob3I+Tm9yZGhhdXMsIFdpbGxpYW0gRDwvYXV0aG9y

PjwvYXV0aG9ycz48L2NvbnRyaWJ1dG9ycz48dGl0bGVzPjx0aXRsZT5Vc2luZyBsdW1pbm9zaXR5

IGRhdGEgYXMgYSBwcm94eSBmb3IgZWNvbm9taWMgc3RhdGlzdGljczwvdGl0bGU+PHNlY29uZGFy

eS10aXRsZT5Qcm9jZWVkaW5ncyBvZiB0aGUgTmF0aW9uYWwgQWNhZGVteSBvZiBTY2llbmNlczwv

c2Vjb25kYXJ5LXRpdGxlPjwvdGl0bGVzPjxwZXJpb2RpY2FsPjxmdWxsLXRpdGxlPlByb2NlZWRp

bmdzIG9mIHRoZSBOYXRpb25hbCBBY2FkZW15IG9mIFNjaWVuY2VzPC9mdWxsLXRpdGxlPjwvcGVy

aW9kaWNhbD48cGFnZXM+ODU4OS04NTk0PC9wYWdlcz48dm9sdW1lPjEwODwvdm9sdW1lPjxudW1i

ZXI+MjE8L251bWJlcj48ZGF0ZXM+PHllYXI+MjAxMTwveWVhcj48L2RhdGVzPjxpc2JuPjAwMjct

ODQyNDwvaXNibj48dXJscz48L3VybHM+PC9yZWNvcmQ+PC9DaXRlPjwvRW5kTm90ZT5=

ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5IZW5kZXJzb248L0F1dGhvcj48WWVhcj4yMDE3PC9ZZWFy

PjxSZWNOdW0+MTU1PC9SZWNOdW0+PERpc3BsYXlUZXh0PihIZW5kZXJzb24gZXQgYWwuIDIwMTcs

IE1pY2hhbG9wb3Vsb3MgYW5kIFBhcGFpb2Fubm91IDIwMTMsIENoZW4gYW5kIE5vcmRoYXVzIDIw

MTEpPC9EaXNwbGF5VGV4dD48cmVjb3JkPjxyZWMtbnVtYmVyPjE1NTwvcmVjLW51bWJlcj48Zm9y

ZWlnbi1rZXlzPjxrZXkgYXBwPSJFTiIgZGItaWQ9Ind4cHhhd3d6ZXdlZno2ZXNyeDV4ZGEyb3R2

ZTI1cHBwNWUwZiIgdGltZXN0YW1wPSIxNTM3NzY0NzE0Ij4xNTU8L2tleT48L2ZvcmVpZ24ta2V5

cz48cmVmLXR5cGUgbmFtZT0iSm91cm5hbCBBcnRpY2xlIj4xNzwvcmVmLXR5cGU+PGNvbnRyaWJ1

dG9ycz48YXV0aG9ycz48YXV0aG9yPkhlbmRlcnNvbiwgSiBWZXJub248L2F1dGhvcj48YXV0aG9y

PlNxdWlyZXMsIFRpbTwvYXV0aG9yPjxhdXRob3I+U3RvcmV5Z2FyZCwgQWRhbTwvYXV0aG9yPjxh

dXRob3I+V2VpbCwgRGF2aWQ8L2F1dGhvcj48L2F1dGhvcnM+PC9jb250cmlidXRvcnM+PHRpdGxl

cz48dGl0bGU+VGhlIGdsb2JhbCBkaXN0cmlidXRpb24gb2YgZWNvbm9taWMgYWN0aXZpdHk6IG5h

dHVyZSwgaGlzdG9yeSwgYW5kIHRoZSByb2xlIG9mIHRyYWRlPC90aXRsZT48c2Vjb25kYXJ5LXRp

dGxlPlRoZSBRdWFydGVybHkgSm91cm5hbCBvZiBFY29ub21pY3M8L3NlY29uZGFyeS10aXRsZT48

L3RpdGxlcz48cGVyaW9kaWNhbD48ZnVsbC10aXRsZT5UaGUgUXVhcnRlcmx5IEpvdXJuYWwgb2Yg

RWNvbm9taWNzPC9mdWxsLXRpdGxlPjwvcGVyaW9kaWNhbD48cGFnZXM+MzU3LTQwNjwvcGFnZXM+

PHZvbHVtZT4xMzM8L3ZvbHVtZT48bnVtYmVyPjE8L251bWJlcj48ZGF0ZXM+PHllYXI+MjAxNzwv

eWVhcj48L2RhdGVzPjxpc2JuPjAwMzMtNTUzMzwvaXNibj48dXJscz48L3VybHM+PC9yZWNvcmQ+

PC9DaXRlPjxDaXRlPjxBdXRob3I+TWljaGFsb3BvdWxvczwvQXV0aG9yPjxZZWFyPjIwMTM8L1ll

YXI+PFJlY051bT4xNjg8L1JlY051bT48cmVjb3JkPjxyZWMtbnVtYmVyPjE2ODwvcmVjLW51bWJl

cj48Zm9yZWlnbi1rZXlzPjxrZXkgYXBwPSJFTiIgZGItaWQ9Ind4cHhhd3d6ZXdlZno2ZXNyeDV4

ZGEyb3R2ZTI1cHBwNWUwZiIgdGltZXN0YW1wPSIxNTM3NzY1MjM1Ij4xNjg8L2tleT48L2ZvcmVp

Z24ta2V5cz48cmVmLXR5cGUgbmFtZT0iSm91cm5hbCBBcnRpY2xlIj4xNzwvcmVmLXR5cGU+PGNv

bnRyaWJ1dG9ycz48YXV0aG9ycz48YXV0aG9yPk1pY2hhbG9wb3Vsb3MsIFN0ZWxpb3M8L2F1dGhv

cj48YXV0aG9yPlBhcGFpb2Fubm91LCBFbGlhczwvYXV0aG9yPjwvYXV0aG9ycz48L2NvbnRyaWJ1

dG9ycz48dGl0bGVzPjx0aXRsZT5OYXRpb25hbCBpbnN0aXR1dGlvbnMgYW5kIHN1Ym5hdGlvbmFs

IGRldmVsb3BtZW50IGluIEFmcmljYTwvdGl0bGU+PHNlY29uZGFyeS10aXRsZT5UaGUgUXVhcnRl

cmx5IEpvdXJuYWwgb2YgRWNvbm9taWNzPC9zZWNvbmRhcnktdGl0bGU+PC90aXRsZXM+PHBlcmlv

ZGljYWw+PGZ1bGwtdGl0bGU+VGhlIFF1YXJ0ZXJseSBKb3VybmFsIG9mIEVjb25vbWljczwvZnVs

bC10aXRsZT48L3BlcmlvZGljYWw+PHBhZ2VzPjE1MS0yMTM8L3BhZ2VzPjx2b2x1bWU+MTI5PC92

b2x1bWU+PG51bWJlcj4xPC9udW1iZXI+PGRhdGVzPjx5ZWFyPjIwMTM8L3llYXI+PC9kYXRlcz48

aXNibj4wMDMzLTU1MzM8L2lzYm4+PHVybHM+PC91cmxzPjwvcmVjb3JkPjwvQ2l0ZT48Q2l0ZT48

QXV0aG9yPkNoZW48L0F1dGhvcj48WWVhcj4yMDExPC9ZZWFyPjxSZWNOdW0+MTczPC9SZWNOdW0+

PHJlY29yZD48cmVjLW51bWJlcj4xNzM8L3JlYy1udW1iZXI+PGZvcmVpZ24ta2V5cz48a2V5IGFw

cD0iRU4iIGRiLWlkPSJ3eHB4YXd3emV3ZWZ6NmVzcng1eGRhMm90dmUyNXBwcDVlMGYiIHRpbWVz

dGFtcD0iMTUzNzc2NTQzMiI+MTczPC9rZXk+PC9mb3JlaWduLWtleXM+PHJlZi10eXBlIG5hbWU9

IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3JlZi10eXBlPjxjb250cmlidXRvcnM+PGF1dGhvcnM+PGF1

dGhvcj5DaGVuLCBYaTwvYXV0aG9yPjxhdXRob3I+Tm9yZGhhdXMsIFdpbGxpYW0gRDwvYXV0aG9y

PjwvYXV0aG9ycz48L2NvbnRyaWJ1dG9ycz48dGl0bGVzPjx0aXRsZT5Vc2luZyBsdW1pbm9zaXR5

IGRhdGEgYXMgYSBwcm94eSBmb3IgZWNvbm9taWMgc3RhdGlzdGljczwvdGl0bGU+PHNlY29uZGFy

eS10aXRsZT5Qcm9jZWVkaW5ncyBvZiB0aGUgTmF0aW9uYWwgQWNhZGVteSBvZiBTY2llbmNlczwv

c2Vjb25kYXJ5LXRpdGxlPjwvdGl0bGVzPjxwZXJpb2RpY2FsPjxmdWxsLXRpdGxlPlByb2NlZWRp

bmdzIG9mIHRoZSBOYXRpb25hbCBBY2FkZW15IG9mIFNjaWVuY2VzPC9mdWxsLXRpdGxlPjwvcGVy

aW9kaWNhbD48cGFnZXM+ODU4OS04NTk0PC9wYWdlcz48dm9sdW1lPjEwODwvdm9sdW1lPjxudW1i

ZXI+MjE8L251bWJlcj48ZGF0ZXM+PHllYXI+MjAxMTwveWVhcj48L2RhdGVzPjxpc2JuPjAwMjct

ODQyNDwvaXNibj48dXJscz48L3VybHM+PC9yZWNvcmQ+PC9DaXRlPjwvRW5kTm90ZT5=

ADDIN EN.CITE.DATA (Henderson et al. 2017, Michalopoulos and Papaioannou 2013, Chen and Nordhaus 2011), income growth ADDIN EN.CITE <EndNote><Cite><Author>Henderson</Author><Year>2012</Year><RecNum>159</RecNum><DisplayText>(Henderson, Storeygard, and Weil 2012)</DisplayText><record><rec-number>159</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1537764714">159</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Henderson, J Vernon</author><author>Storeygard, Adam</author><author>Weil, David N</author></authors></contributors><titles><title>Measuring economic growth from outer space</title><secondary-title>The American Economic Review</secondary-title></titles><periodical><full-title>The American Economic Review</full-title></periodical><pages>994-1028</pages><volume>102</volume><number>2</number><dates><year>2012</year></dates><isbn>0002-8282</isbn><urls></urls></record></Cite></EndNote>(Henderson, Storeygard, and Weil 2012), and poverty ADDIN EN.CITE <EndNote><Cite><Author>Jean</Author><Year>2016</Year><RecNum>165</RecNum><DisplayText>(Jean et al. 2016)</DisplayText><record><rec-number>165</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1537765198">165</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Jean, Neal</author><author>Burke, Marshall</author><author>Xie, Michael</author><author>Davis, W Matthew</author><author>Lobell, David B</author><author>Ermon, Stefano</author></authors></contributors><titles><title>Combining satellite imagery and machine learning to predict poverty</title><secondary-title>Science</secondary-title></titles><periodical><full-title>science</full-title></periodical><pages>790-794</pages><volume>353</volume><number>6301</number><dates><year>2016</year></dates><isbn>0036-8075</isbn><urls></urls></record></Cite></EndNote>(Jean et al. 2016).The data is collected using polar orbiting satellites, which have an Operation Linescan system that can detect low levels of visible-near infrared radiance at night. NTL data was compiled from the Defense Meteorological Satellite Program (DMSP) for the years 2000 and 2010 allowing for cross-comparability. NTL is averaged across all grid cells within each district to obtain the mean NTL. Soil ErosionSoil erosion can lead to land degradation in the form of nutrient loss, a decrease in the effective root depth, water imbalance in the root zone, productivity reduction ADDIN EN.CITE <EndNote><Cite><Author>Yang</Author><Year>2003</Year><RecNum>179</RecNum><DisplayText>(Yang et al. 2003)</DisplayText><record><rec-number>179</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1538156011">179</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Yang, Dawen</author><author>Kanae, Shinjiro</author><author>Oki, Taikan</author><author>Koike, Toshio</author><author>Musiake, Katumi</author></authors></contributors><titles><title>Global potential soil erosion with reference to land use and climate changes</title><secondary-title>Hydrological processes</secondary-title></titles><periodical><full-title>Hydrological processes</full-title></periodical><pages>2913-2928</pages><volume>17</volume><number>14</number><dates><year>2003</year></dates><isbn>0885-6087</isbn><urls></urls></record></Cite></EndNote>(Yang et al. 2003). Further, soil degradation can trigger food insecurity and contribute to vulnerability and poverty ADDIN EN.CITE <EndNote><Cite><Author>Lal</Author><Year>2004</Year><RecNum>127</RecNum><DisplayText>(Lal 2004, West et al. 2014)</DisplayText><record><rec-number>127</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1537396504">127</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Lal, Rattan</author></authors></contributors><titles><title>Soil carbon sequestration impacts on global climate change and food security</title><secondary-title>science</secondary-title></titles><periodical><full-title>science</full-title></periodical><pages>1623-1627</pages><volume>304</volume><number>5677</number><dates><year>2004</year></dates><isbn>0036-8075</isbn><urls></urls></record></Cite><Cite><Author>West</Author><Year>2014</Year><RecNum>128</RecNum><record><rec-number>128</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1537396682">128</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>West, Paul C</author><author>Gerber, James S</author><author>Engstrom, Peder M</author><author>Mueller, Nathaniel D</author><author>Brauman, Kate A</author><author>Carlson, Kimberly M</author><author>Cassidy, Emily S</author><author>Johnston, Matt</author><author>MacDonald, Graham K</author><author>Ray, Deepak K</author></authors></contributors><titles><title>Leverage points for improving global food security and the environment</title><secondary-title>Science</secondary-title></titles><periodical><full-title>science</full-title></periodical><pages>325-328</pages><volume>345</volume><number>6194</number><dates><year>2014</year></dates><isbn>0036-8075</isbn><urls></urls></record></Cite></EndNote>(Lal 2004, West et al. 2014). This paper uses the soil erosion dataset calculated using the adjusted Revised Universal Soil Loss Equation (RUSLE) model by ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Naipal</Author><Year>2015</Year><RecNum>8</RecNum><DisplayText>Naipal et al. (2015)</DisplayText><record><rec-number>8</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1529983195">8</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Naipal, Victoria</author><author>Reick, Christian H</author><author>Pongratz, Julia</author><author>Van Oost, Kristof</author></authors></contributors><titles><title>Improving the global applicability of the RUSLE model-adjustment of the topographical and rainfall erosivity factors</title><secondary-title>Geoscientific Model Development</secondary-title></titles><periodical><full-title>Geoscientific Model Development</full-title></periodical><pages>2893-2913</pages><volume>8</volume><dates><year>2015</year></dates><urls></urls></record></Cite></EndNote>Naipal et al. (2015). The RUSLE model predicts erosion rates and takes into account rainfall erosivity, slope steepness, slope length, soil erodibility, land cover, and support practice. The soil erosion rate for each pixel is averaged across all grid cells within each district to obtain the mean soil erosion at the 110-district level. Mean soil erosion is shown in Figure 3. Soil erosion is particularly serious in poor Upper West region. Less Favored Agricultural LandLess favored agricultural land (LFAL) is susceptible to low productivity and degradation, because its agricultural potential is constrained biophysically by terrain, poor soil quality, or limited rainfall ADDIN EN.CITE <EndNote><Cite><Author>Barbier</Author><Year>2014</Year><RecNum>14</RecNum><DisplayText>(Barbier and Hochard 2014)</DisplayText><record><rec-number>14</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1530070209">14</key></foreign-keys><ref-type name="Unpublished Work">34</ref-type><contributors><authors><author>Barbier, Edward B</author><author>Hochard, Jacob P</author></authors></contributors><titles><title>Land Degradation, Less Favored Lands and the Rural Poor: A Spatial and Economic Analysis</title><secondary-title>A Report for the Economics of Land Degradation Initiative. Department of Economics and Finance, University of Wyoming, available at: eld-initiative. org</secondary-title></titles><periodical><full-title>A Report for the Economics of Land Degradation Initiative. Department of Economics and Finance, University of Wyoming, available at: eld-initiative. org</full-title></periodical><dates><year>2014</year></dates><urls></urls></record></Cite></EndNote>(Barbier and Hochard 2014). This paper uses the LFAL variable defined by ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Pender</Author><Year>2000</Year><RecNum>10</RecNum><DisplayText>Pender and Hazell (2000)</DisplayText><record><rec-number>10</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1529983273">10</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Pender, John</author><author>Hazell, Peter</author></authors></contributors><titles><title>Promoting sustainable development in less-favored areas</title></titles><dates><year>2000</year></dates><urls></urls></record></Cite></EndNote>Pender and Hazell (2000), ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>World Bank</Author><Year>2008</Year><RecNum>11</RecNum><DisplayText>World Bank (2008)</DisplayText><record><rec-number>11</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1529983437">11</key></foreign-keys><ref-type name="Report">27</ref-type><contributors><authors><author>World Bank,</author></authors></contributors><titles><title>Word Development Report 2008: Agricultural Development</title></titles><dates><year>2008</year></dates><pub-location>Washington DC</pub-location><publisher>The World Bank</publisher><urls></urls></record></Cite></EndNote>World Bank (2008), and ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Barbier</Author><Year>2014</Year><RecNum>14</RecNum><DisplayText>Barbier and Hochard (2014)</DisplayText><record><rec-number>14</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1530070209">14</key></foreign-keys><ref-type name="Unpublished Work">34</ref-type><contributors><authors><author>Barbier, Edward B</author><author>Hochard, Jacob P</author></authors></contributors><titles><title>Land Degradation, Less Favored Lands and the Rural Poor: A Spatial and Economic Analysis</title><secondary-title>A Report for the Economics of Land Degradation Initiative. Department of Economics and Finance, University of Wyoming, available at: eld-initiative. org</secondary-title></titles><periodical><full-title>A Report for the Economics of Land Degradation Initiative. Department of Economics and Finance, University of Wyoming, available at: eld-initiative. org</full-title></periodical><dates><year>2014</year></dates><urls></urls></record></Cite></EndNote>Barbier and Hochard (2014). LFAL consists of irrigated land on terrain greater than 8 percent median slope or poor soil quality, rainfed land with a length of growing period of more than 120 days but either on terrain greater than 8 percent median slope or with poor soil quality, semi-arid land (land with length of growing period (LGP) of 60–119 days), and arid land (land with LGP of less than 60 days). The global extent of agriculture is defined as all areas with 10 percent or greater cropland, grazing land or irrigated area net of areas with a growing period of zero days. The data source is from various years. Road Density The Netherlands Environmental Assessment Agency maintains the Global Roads Inventory Project (GRIP). GRIP combines data collected from about 60 public sources including the United Nations, national spatial data sources, topographic agencies, non-governmental and international organizations, and volunteered geographic information. HDD defines road density as the mean length of road in kilometers per cell, re-scaled from GRIP’s original 10km resolution to the boundaries of 110 districts. Values are for the year 2013, and include total road density, as well as disaggregated densities of highways, primary roads, secondary roads, tertiary roads, and urban/residential roads. Road density by district is shown in Figure 4. Road density is high along the coast and in major cities. Distance to MarketThe European Commission’s Joint Research Centre created a global map of travel time to cities from the data in World Development Report ADDIN EN.CITE <EndNote><Cite><Author>World Bank</Author><Year>2009</Year><RecNum>15</RecNum><DisplayText>(World Bank 2009)</DisplayText><record><rec-number>15</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1530071966">15</key></foreign-keys><ref-type name="Report">27</ref-type><contributors><authors><author>World Bank,</author></authors></contributors><titles><title>World Development Report 2009</title></titles><dates><year>2009</year></dates><pub-location>Washington DC</pub-location><publisher>The World Bank</publisher><urls></urls></record></Cite></EndNote>(World Bank 2009). Accessibility is defined as the travel time to a city of 50,000 or more people using road or water-based travel. Factors that influence travel time include available transport networks, environmental factors like land cover and slope, and political boundaries and borders. A cost-distance algorithm is used to determine the ‘cost’ of travelling between raster cells, measured in time units. The resulting raster grid represents a ‘friction-surface’ of these costs. Mean distant to market by district is illustrated in Figure 4. It shows the distance to market is longer in poor rural areas. Flood Occurrence and Drought SeverityHDD uses the flood occurrence and drought severity data that is made available by the World Resources Institute’s Aquaduct platform. Flood occurrence is defined as the number of flood events recorded from 1985 to 2011. Flood events from news, governmental, instrumental, and remote sensing sources are aggregated by the Global Active Archive of Major Flood, which is then used to estimate the extent of flooding in the affected regions. The remote sensing data is taken from Brackenridge, the Dartmouth Flood Observatory and the University of Colorado (2011) and is in flood extent polygons. Drought severity measures the average length of droughts times the dryness of the droughts from 1901 to 2008. Drought is defined as a continuous period during which the soil moisture level is below the 20th percentile for a given location. Length of drought is the number of months, while dryness is the average number of percentage points soil moisture drops below the 20th percentile. The number of floods and drought severity are shown in Figure 3. The fire show how disproportionally floods and drought are affecting poor northern districts. Air pollutionGlobal fine particulate matter (PM2.5) concentrations were estimated using information from a combination of satellite-, simulation- and monitor-based sources. A Geographically Weighted Regression (GWR) is applied to global satellite-derived PM2.5 estimates. Aerosol optical depth from multiple satellite products was combined based on their relative uncertainties as determined using ground-based sun photometer (Aerosol Robotic Network (AERONET)) observations. The GWR predictors included simulated aerosol composition and land use information. The resultant PM2.5 estimates were highly consistent (R2 = 0.81) with out-of-sample cross-validated PM2.5 concentrations from monitors. The data is produced by and follow a similar methodology to the Global Burden of Disease report. This paper uses the mean PM2.5 air pollution levels at the 110-district level in 2000 and 2010 as an air pollution variable. Trends of poverty, economic structure, human capital, infrastructure and environmental degradation Trends of povertyRegional inequality widened over years. Figure 1 shows the changes in poverty rates at the national and regional levels. Between 1991 and 2016, the national poverty rate declined from 52.7 percent to 23.4 percent. Ashanti, Eastern, Greater Accra, Brong Ahafo, Central, and Western regions experienced significant poverty reduction, as the poverty rates dropped by more than 20 percent in these regions. However, the poverty rates dropped by less than 20 percent points in the Volta, Upper East, Northern, and Upper West regions. In 2016, the poverty rates remain above 50 percent in the Northern, Upper West, and Upper East regions. Poverty rates vary not only across regions but also across districts within regions. Figure 2 shows the district level poverty rates estimated by ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Coulombe</Author><Year>2005</Year><RecNum>1</RecNum><DisplayText>Coulombe (2005)</DisplayText><record><rec-number>1</rec-number><foreign-keys><key app="EN" db-id="wxpxawwzewefz6esrx5xda2otve25ppp5e0f" timestamp="1529714835">1</key></foreign-keys><ref-type name="Press Release">63</ref-type><contributors><authors><author>Coulombe, Harold</author></authors></contributors><titles><title>Ghana Census-Based Poverty Map: District and Sub-District Level Results</title></titles><dates><year>2005</year></dates><publisher>Ghana Statistical Service,&#xD;Department for International Development UK,</publisher><urls><related-urls><url> (2005) and Ghana Statistical Service (2015) for 2000 and 2010, respectively. In 2000, poverty was widely spread in the northern parts of the country. However, the spatial distribution of poverty has changed dramatically in 2010. The eastern part of the three Northern regions achieved significant poverty reduction, while several districts in the western side of the country witnessed increased poverty. Clusters of districts with high poverty rates also developed in the inland parts of the country and in Volta region.Table 2 shows that the average poverty rates at the 110-district level declined from 46.9 to 29.3 percent between 2000 and 2010. However, the poverty rate of the poorest district remains high (81 percent). Some of the poor districts in 2000 were able to drastically reduce poverty rates by 2010 while others were not. For example, Gushiegu-Karaga district in the Northern region reduced the poverty rate from 85.7 percent to 34.3 percent between 2000 and 2010, while the poverty rate dropped from 85.5 percent to only 70.3 percent in Nadowli district in Upper West region during the same period. Economic structure Table 1 shows the average employment rate at the 110-district level increased from 64.8 to 70.7 percent between 2000 and 2010. The employment rate increased for both men and women during the period. Ghana went through a significant structural transformation between 2000 and 2010. The average share of employment in agriculture declined from 63.3 percent to 53.1 percent, while the average share of employment in services increase from 23.5 percent to 33 percent. The share of labor force in industry increased from 13.3 percent to 13.9 percent. In addition, the average income in the industry and service sector rose dramatically. Figure 5 shows the correlations of variables with the district poverty rates in 2010. The shares of employment in service and agriculture is highly correlated with poverty. The share of migrants within districts is also highly correlated with poverty rates. The total employment rates, as well as the employment rates of both men and women, are moderately correlated with the district poverty rates. The mean income in non-agricultural sectors is not strongly correlated with the district poverty rates.Night time lights are not strongly correlated with poverty (Figure 5), however, they are highly correlated with agricultural income. As discussed in Section 2, night time lights often reflect the distribution of economic activities. Night time lights may be reflecting opportunities for farmers to get linked to market. Human capital Between 2000 and 2010, the mean fertility (average number of children per women) declined from 3.2 to 2.7 children. In addition, the share of working age adults who completed primary and secondary school increased from 50.6 to 67.7 percent, and 45.1 to 54.7 percent, respectively. Figure 5 demonstrates the share of working age adults who completed primary and secondary school completion is the most highly correlated variables with poverty among all variables considered in this study. The share of working age population as well as fertility are correlated with the district poverty rate, suggesting the human capital variables selected for this study are substantially important factors on poverty reduction. The human capital variables do not only correlate with poverty but also strongly relate to agricultural income and the share of labor by sector. Lower fertility and a higher share of working age population are associated with higher agricultural income, and a higher share of working age population and the education achievement are strongly correlated with a higher share of labor force in non-agricultural sector, and a lower share of labor in the agricultural sector (Figures 6 and 7). Access to infrastructureAccess to electricity significantly improved between 2000 and 2010 (Table 1). The mean percentage of households with access to electricity increased from 29.2 percent to 48,8 percent between 2000 and 2010. The correlation coefficients of access to electricity, roads, and markets and poverty rates are reported in Figure 5. Access to electricity is exceptionally highly correlated with poverty. Figures 6 and 7 suggest access to electricity is the key factor facilitating the shift of labor from agriculture to industry and services. Climate change and environmental degradationTable 1 shows NPP (biological productivity) declined dramatically between 2000 and 2010. Figure 5 indicates NPP is higher in rich districts than poor districts. This suggests rich districts are endowed with biological productivity, while poor districts are poorly endowed with biological productivity. Figure 5 suggests poor districts also suffer from insufficient rainfall, severe droughts, higher flood frequency and serious soil erosion. The mean poverty rate of the 10 districts with the most frequent floods is 41 percent, while the mean poverty rate of the 10 districts with the least flood frequency is 20.7 percent. Flood frequency is highest in Saboba, Chereponi, Zabzugu, and Tatale districts in the Northern region. In 2018 over 2,000 residents of Zabzugu district were evacuated due to floods. Zabzugu district often get cut off from the Northern region, as the main road gets flooded. In 2016, over 1,000 people were displaced from their homes in Chereponi district.The mean poverty rate of the 10 districts with the highest score of drought severity is 45.3 percent, while the mean poverty rate of the 10 districts with the least severe drought is 16.4 percent. Drought severity is highest in Bawku West and Bawku East districts in Upper The poverty rates of Bawku West and Bawku East districts are 47.1 percent and 68.1 percent, respectively.Many poor districts suffer from both floods and droughts. The old East Gonja district in Northern Region (which currently includes East Gonja and Kpandai districts under the 216-district system) had the highest poverty rates of 80.9 percent in 2012. It experienced 13.4 major floods between 1985 and 2011 on average, and its drought severity score is 24.1. These numbers are much higher than the average of 9.9 and 20.9 in the country, respectively. In districts with low agricultural income, drought severity and soil erosion are significantly higher (Figure 6). Districts with high shares of agricultural labor have less rainfall and biological productivity and suffer from more frequent floods and severe droughts as well as more intense soil erosion (Figure 6). Panel data analysisIn the previous section, we have demonstrated how selected variables are strongly correlated with poverty rates. The fundamental setback of such simple statistical analysis is causal inferences. In this section, we conduct panel data regressions to investigate causal effects of the changes in the key variables on the changes in poverty. Panel data analysis is useful in controlling for the effects of other variables that are not included as independent variables in the regression model. First, we estimate three alternative regression models and conduct several tests to find out which regression model is most appropriate for each estimation.If district specific effect does not exist, ordinary least squares (OLS) produces efficient and consistent estimates.Yit=α+βXit+εitui=0 (district specific effect=0)If there are individual differences in intercepts, then a fixed effect model is suitable. Yit=(α+ui)+βXit+vitA random effect model assumes individual specific errors. Yit=α+βXit+(ui+vit)Fixed effects are tested by the F-test, and random effects are tested by the Lagrange multiplier test ADDIN EN.CITE <EndNote><Cite><Author>Breusch</Author><Year>1980</Year><RecNum>52</RecNum><DisplayText>(Breusch and Pagan 1980)</DisplayText><record><rec-number>52</rec-number><foreign-keys><key app="EN" db-id="2wfd5rtfos2e0qe2998vx90isetx99ew9ppp" timestamp="1545131592">52</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Breusch, Trevor Stanley</author><author>Pagan, Adrian Rodney</author></authors></contributors><titles><title>The Lagrange multiplier test and its applications to model specification in econometrics</title><secondary-title>The Review of Economic Studies</secondary-title></titles><periodical><full-title>The Review of Economic Studies</full-title></periodical><pages>239-253</pages><volume>47</volume><number>1</number><dates><year>1980</year></dates><isbn>0034-6527</isbn><urls></urls></record></Cite></EndNote>(Breusch and Pagan 1980). If the null hypothesis is not rejected by the F-test, then the pooled OLD regression is favored to fixed effect regression. If the null hypothesis is not rejected by the Lagrange multiplier test, then the pooled OLD regression is favored to random effect regression. In addition, we conduct the Hausman specification test to compare a random effect model and a fixed effect model. The results of all tests, and a selected regression model for each estimation are summarized in Table A1 in the Appendix. In all cases, the Hausman specification test was not rejected. Thus, we use the random effect model (in Tables 2 and 3) unless it is rejected by the Lagrange multiplier test. When the random effect model is rejected by the Lagrange multiplier test but the fixed effect model was not rejected by the F-test, we use the fixed effect model (Table 4). Table 2 contains the results of random effect regression results. The left column contains the regression result for poverty rates, and the right column contains the regression result for consumption. The regression results demonstrate poverty rates significantly declined and consumption increased in districts where income per worker in the service sector increased. Between 2000 and 2010, increases in the service sector income by GHC 1 million led to 1.2 percent point reduction in district poverty rates. Consumption also increased in districts where agricultural income per worker increased. Increased shares of working age population within districts decreased poverty and increased consumption. It is probably because the share of dependent household members declined in districts where poverty reduced, and consumption increased. An increase in employment rates (the percentage of working population who are economically active) also lead to poverty reduction and consumption growth. An increased employment rate by 1 percent led to a reduction in poverty rates by 1 percent point. Increased shares of workers in the service sector contribute to both poverty reduction and consumption increases. One percent shift of labor from agriculture to the service sector led to 0.6 percentage point poverty reduction. We further investigate the difference in the impacts of labor participation by gender. In the regressions presented in Table 3, we disaggregate a share of working age population and employment rate into male and female. The regression results suggest an increase share of working population among male increases consumption and reduces poverty. On the other hand, increased employment rates of women weakly reduce poverty and increase consumption. In the regressions presented in Table 4, we examine how human capital development, infrastructure improvement, and environmental factors have affected poverty reduction. In the first regression, we used the completion rate of primary school among working age adults as in independent variable, and in the second regression, we used the completion rate of secondary school among working age adults as in independent variable. the Poverty reduction is associated with increased employment rates, improved access to electricity, and increased precipitation. One percent increase in the share of households with access to electricity led to 0.4 percent point drop in poverty rates. In addition, an increase in rainfall by 1 meter led to 0.2 percent reduction in poverty rates. The regression results indicate improved human capital indicators, such as fertility, educational achievement, and demographic change had no impacts on poverty reduction. It may be because ten years is not long enough to benefit from human capital development. Next, the paper investigates factors that have affected changes in agricultural income and structural change. Since we do not have panel data for soil erosion, distance to market, road density, LFAL, flood frequency and drought severity, we cannot run the same model of panel data analysis. We estimate the following regression model instead.?Yit=α+β1?X1it+β2X2i+εit where ?Yit is the annual change in income per worker in agriculture, annual change in the shares of agriculture, industry and services, respectively. ?X1it includes demographic and other variables that are panel data. X2i includes other variables that are not a part of the panel data, i.e., soil erosion, distance to market, road density, LFAL, flood frequency and drought severity.Table 5 shows soil erosion had a large impact on the changes in agricultural income. Agricultural income declined in districts with higher soil erosion. Improved access to electricity, roads, and markets were crucial for creating jobs in non-agricultural sectors. Increased shares of households with access to electricity led to a reduction in the share of workers in agriculture and an increase of the share of workers in industry. Between 2000 and 2010, a 1 percent improvement in the share of households with access to electricity increased the share of workers in industry and decreased the share of workers in agriculture both by 0.06 percent. Better access to markets and roads increased the share of employment in the service sector. In areas where droughts are severe, more people became engaged in agriculture and less in industry and services. It may be because severe droughts prevented people from accumulating capital to invest in non-agricultural sectors. Conclusion and Policy Recommendation This study combines district level poverty rates, population census data, income data, and geospatial data to investigate how human capital development, economic transformation, improved access to infrastructure and environmental degradation impacted poverty, agricultural income and allocations of labor in Ghana. We find poverty reduction was achieved through increased employment rates and income in the service sector, shifts of labor from agriculture to the service sector, expansion of access to electricity, and increased rainfall. In addition, educational achievement, access to electricity, shift of labor from agriculture to services are important determining factors of the level of poverty. The paper further investigates the factors that have affected changes in agricultural income and structural change. Soil erosion had a huge impact on the changes in agricultural income. Districts with high shares of agricultural labor have less rainfall and biological productivity and suffer from more frequent floods and severe droughts as well as more intense soil erosion. In addition, they are far from markets, have lower road density, and lower economic activities. Improved access to electricity, roads, and markets were crucial for creating jobs in non-agricultural sectors. In areas where droughts are severe, more people became engaged in agriculture and less in industry and services. It may be because severe droughts prevent people from investing in non-agricultural sectors.The results of this study have two key implications in terms of policy recommendations for Ghana to reduce poverty and create jobs in non-agricultural sectors in lagging areas. Firstly, Ghana needs to invest on infrastructure, which will improve access to markets, and value chains on agricultural commodities, which will enable structural change and the creation of jobs in agricultural and non-agricultural sectors. Such investments in infrastructure will reduce poverty whilst improving sustainable natural resource management.In addition, the Government of Ghana needs to invest on a transformative landscape, and sustainable land and water management initiatives in districts affected by soil erosion and land degradation in the northern regions to boost agricultural production and reduce rural poverty. Such initiatives will also build the resilience of communities and ecosystems to climate change and natural shocks such as floods and droughts. In particular, soil erosion can be mitigated by introducing sustainable farming system and regulation of reservoirs for proper water utilization. Improved water infrastructure can provide water supply for agriculture, as well as prevent flooding. The results of panel data analysis do not support the hypothesis that human capital development is a significant force behind poverty reduction. However, ten years may not be long enough to realize the benefit of human capital development. The analytical results suggest increase labor participation by women is effective in poverty reduction. Reducing fertility by family planning is a practicable way to encourage women to stay active in the labor force. References ADDIN EN.REFLIST Bacchini, Roberto Dario, and Daniel Fernando Miguez. 2015. "Agricultural Risk Management Using NDVI Pasture Index-Based Insurance for Livestock Producers in South West Buenos Aires Province." Agricultural Finance Review 75 (1):77-91.Barbier, Edward B, and Jacob P Hochard. 2014. "Title." A Report for the Economics of Land Degradation Initiative. Department of Economics and Finance, University of Wyoming, available at: eld-initiative. org.Bauer, Jan M., and Samuel Mburu. 2017. "Effects of Drought on Child Health in Marsabit District, Northern Kenya." Economics and Human Biology 24:74-79. doi: , Clemens, Xinshen Diao, James Thurlow, and Ramatu M. Al-Hassan. 2008. "Agriculture for Development in Ghana: New Opportunities and Challenges." International Food Policy Research Institute Paper 00784:1-48.Breusch, Trevor Stanley, and Adrian Rodney Pagan. 1980. "The Lagrange multiplier test and its applications to model specification in econometrics." The Review of Economic Studies 47 (1):239-253.Brown, Molly E., Jorge E. Pinzon, and Stephen D. Prince. 2008. "Using Satellite Remote Sensing Data in a Spatially Explicit Price Model: Vegetation Dynamics and Millet Prices." Land Economics 84 (2):340-357. doi: , Xi, and William D Nordhaus. 2011. "Using luminosity data as a proxy for economic statistics." Proceedings of the National Academy of Sciences 108 (21):8589-8594.Christiaensen, Luc, Lionel Demery, and Jesper Kuhl. 2011. "The (evolving) role of agriculture in poverty reduction—An empirical perspective." Journal of development economics 96 (2):239-254.Coulombe, Harold. 2005. Ghana Census-Based Poverty Map: District and Sub-District Level Results. Ghana Statistical Service,Department for International Development UK,.De Vreyer, Philippe, Javier Herrera, and Sandrine Mesplé-Somps. 2009. "Consumption growth and spatial poverty traps: an analysis of the effect of social services and community infrastructures on living standards in rural Peru." Poverty, Inequality, and Policy in Latin America:129-155.De Vries, Gaaitzen, Marcel Timmer, and Klaas De Vries. 2015. "Structural transformation in Africa: Static gains, dynamic losses." The Journal of Development Studies 51 (6):674-688.Diao, Xinshen, Kenneth Harttgen, and Margaret McMillan. 2017. "The Changing Structure of Africa’s Economies." The World Bank Economic Review 31 (2):412-433.Dinkelman, Taryn. 2011. "The effects of rural electrification on employment: New evidence from South Africa." American Economic Review 101 (7):3078-3108.Duarte, Margarida, and Diego Restuccia. 2010. "The role of the structural transformation in aggregate productivity." The Quarterly Journal of Economics 125 (1):129-173.Elbers, Chris, Jean O Lanjouw, and Peter Lanjouw. 2003. "Micro–level estimation of poverty and inequality." Econometrica 71 (1):355-364.Fernald, John G. 1999. "Roads to prosperity? Assessing the link between public capital and productivity." American economic review 89 (3):619-638.Fields, Gary, Paul Cichello, Samuel Freije, Marta Menéndez, and David Newhouse. 2003. "Household income dynamics: a four-country story." The journal of development studies 40 (2):30-54.Geiger, Michael, Jan Trenczek, and Konstantin M. Wacker. 2018.Ghana Statistical Service. 2015. Ghana Poverty Mapping Report. Ghana Statistical Service,.Gollin, Douglas, David Lagakos, and Michael E Waugh. 2013. "The agricultural productivity gap." The Quarterly Journal of Economics 129 (2):939-993.Gollin, Douglas, Stephen Parente, and Richard Rogerson. 2002. "The role of agriculture in development." American Economic Review 92 (2):160-164.Henderson, J Vernon, Tim Squires, Adam Storeygard, and David Weil. 2017. "The global distribution of economic activity: nature, history, and the role of trade." The Quarterly Journal of Economics 133 (1):357-406.Henderson, J Vernon, Adam Storeygard, and David N Weil. 2012. "Measuring economic growth from outer space." The American Economic Review 102 (2):994-1028.Higgins, Kate, Kate Bird, and Daniel Harris. 2010. "Policy responses to the spatial dimensions of poverty."Higgins, Nathaniel, Beat Hintermann, and Molly E. Brown. 2015. "A Model of West African Millet Prices in Rural Markets." Food Policy 52:33-43. doi: , Charles R, Esra Bennathan, and Sylaja Srinivasan. 2006. "Infrastructure, externalities, and economic development: a study of the Indian manufacturing industry." The World Bank Economic Review 20 (2):291-308.Jacoby, Hanan G. 2000. "Access to markets and the benefits of rural roads." The Economic Journal 110 (465):713-737.Jalan, Jyotsna, and Martin Ravallion. 2002. "Geographic poverty traps? A micro model of consumption growth in rural China." Journal of applied econometrics 17 (4):329-346.J?ntti, Markus, and Stephen P Jenkins. 2015. "Income mobility." In Handbook of income distribution, 807-935. Elsevier.Jean, Neal, Marshall Burke, Michael Xie, W Matthew Davis, David B Lobell, and Stefano Ermon. 2016. "Combining satellite imagery and machine learning to predict poverty." Science 353 (6301):790-794.Lal, Rattan. 2004. "Soil carbon sequestration impacts on global climate change and food security." science 304 (5677):1623-1627.Makaudze, Ephias M., and Mario J. Miranda. 2010. "Catastrophic Drought Insurance Based on the Remotely Sensed Normalised Difference Vegetation Index for Smallholder Farmers in Zimbabwe." Agrekon 49 (4):418-432. doi: , Margaret S, and Kenneth Harttgen. 2014. What is driving the'African Growth Miracle'? : National Bureau of Economic Research.Michalopoulos, Stelios, and Elias Papaioannou. 2013. "National institutions and subnational development in Africa." The Quarterly Journal of Economics 129 (1):151-213.Naipal, Victoria, Christian H Reick, Julia Pongratz, and Kristof Van Oost. 2015. "Improving the global applicability of the RUSLE model-adjustment of the topographical and rainfall erosivity factors." Geoscientific Model Development 8:2893-2913.O'Neill, Daniel W., and David J. Abson. 2009. "To Settle or Protect? A Global Analysis of Net Primary Production in Parks and Urban Areas." Ecological Economics 69 (2):319-327. doi: , John, and Peter Hazell. 2000. "Promoting sustainable development in less-favored areas."Rodrik, Dani. 2016. "Premature deindustrialization." Journal of Economic Growth 21 (1):1-33.Schultz, Theodore W. 1961. "Investment in human capital." The American economic review 51 (1):1-17.West, Paul C, James S Gerber, Peder M Engstrom, Nathaniel D Mueller, Kate A Brauman, Kimberly M Carlson, Emily S Cassidy, Matt Johnston, Graham K MacDonald, and Deepak K Ray. 2014. "Leverage points for improving global food security and the environment." Science 345 (6194):325-328.World Bank. 2008. Word Development Report 2008: Agricultural Development. Washington DC: The World Bank.World Bank. 2009. World Development Report 2009. Washington DC: The World Bank.Xu, Chi, Yutong Li, Jian Hu, Xuejiao Yang, Sheng Sheng, and Maosong Liu. 2012. "Evaluating the difference between the normalized difference vegetation index and net primary productivity as the indicators of vegetation vigor assessment at landscape scale." Environmental monitoring and assessment 184 (3):1275-1286.Yang, Dawen, Shinjiro Kanae, Taikan Oki, Toshio Koike, and Katumi Musiake. 2003. "Global potential soil erosion with reference to land use and climate changes." Hydrological processes 17 (14):2913-2928.Yiran, Gerald AB, and Lindsay C Stringer. 2016. "Spatio-temporal analyses of impacts of multiple climatic hazards in a savannah ecosystem of Ghana." Climate Risk Management 14:11-26.Zika, Michael, and Karl-Heinz Erb. 2009. "The global loss of net primary production resulting from human-induced soil degradation in drylands." Ecological Economics 69 (2):310-318.Table 1: Summary table?20002010?SampleMeanMinMaxSampleMeanMinMaxPoverty rate1100.4690.0520.8611100.2930.0250.809Economic structure Employment rate (%)1100.6480.4300.8521100.7070.5150.848of men1100.6610.4550.8741100.7160.4920.865of women1100.6340.4100.8331100.6980.5370.829-Sector of employment (%)- Agriculture1100.6330.0580.8901100.5310.0230.912 Industry1100.1330.0390.3681100.1390.0210.305 Service1100.2350.0700.6961100.3300.0670.786-Income (million)- Agriculture1010.0030.0010.0401100.0020.0010.023 Industry880.0020.0000.0081070.0290.0000.142 Service930.0030.0000.0421100.0250.0030.160NTL 1102.6560.00058.5661103.1020.00057.019Human capitalWorking age pop (%)1100.5210.4340.6471100.5520.4680.679Fertility 1103.2351.9124.0441102.7311.6303.464Completed primary school (%)1100.5060.1030.8381100.6770.1920.928Completed secondary school (%)1100.4510.0940.7701100.5470.1160.833Infrastructure Access to electricity (%)1100.2970.0390.8961100.4880.1120.910Road density 11012.9597.21534.590Distance to market1101.8130.0855.571Environmental factorsPrecipitation 1101.0020.8021.2551101.2920.9322.013NDVI 1100.2930.0930.5481100.3340.0800.548NPP 1100.6910.0041.1301100.3960.1140.687Air pollution11030.45822.91135.37711032.86619.35140.392Soil erosion 1104.5870.12331.769LFAL1100.4630.0131.000Flood frequency1109.8572.00020.000Drought severity11020.9226.80627.596????Table 2: The impacts of changes in economic structure on the changes in poverty rates and consumption (panel data analysis): Random effect regressions(1)(2)?Poverty rateConsumptionIncome per worker in agriculture-2.19012,297.2***(2.512)(3,559.6)Income per worker in industry0.428-288.2(0.458)(665.7)Income per worker in service-1.199**1,853.1**(0.537)(788.4)Share of working age population (%)-0.950**2,254.2***(0.452)(633.2)Employment rate (%)-0.966***1,195.3***(0.138)(207.7)% employed in industry-0.378*-180.4(0.210)(300.9)% employed in service-0.635***964.8***(0.126)(177.1)Constant1.785***-1,693.0***(0.228)(327.1)Observations193192R-squared0.550.61Number of districts110110Standard errors in parentheses.*** p<0.01, ** p<0.05, * p<0.1.Table 3: The impacts of changes in economic structure and gender disaggregated labor participation on the changes in poverty rates and consumption (panel data analysis): Random effect regressions(1)(2)Poverty rateConsumptionIncome per worker in agriculture-1.10712,856.6***(2.514)(3,531.6)Income per worker in industry0.189150.1(0.464)(666.1)Income per worker in service-1.302**2,007.4***(0.534)(774.1)Share of working age population among male (%)-2.077***3,462.5***(0.582)(801.5)Share of working age population among female (%)0.836386.7(0.749)(1,045.6)Employment rate of men (%)-0.267-1,011.5**(0.353)(499.8)Employment rate of women (%)-0.609*2,041.0***(0.324)(464.3)% employed in industry-0.327-130.6(0.208)(295.1)% employed in service-0.741***998.2***(0.130)(179.9)Constant1.540***-1,368.7***(0.241)(341.5)Observations193192R-squared0.570.63Number of districts110110Standard errors in parentheses.*** p<0.01, ** p<0.05, * p<0.1.Table 4: The impacts of changes in human capital, infrastructure and environments on the changes in poverty rates (panel data analysis): Fixed effect regressions(1)(2)Poverty rate Poverty rateShare of working age population (%)-0.0970.232(0.823)(0.873)Employment rate (%)-0.509**-0.451**(0.203)(0.188)% employed in industry0.2360.207(0.294)(0.294)% employed in service0.0690.176(0.237)(0.258)Fertility0.0480.029(0.063)(0.061)% adults who completed primary school-0.037(0.269)% adults who completed secondary school-0.451(0.467)Access to electricity (%)-0.423**-0.401**(0.164)(0.162)NDVI-0.0350.006(0.140)(0.143)NPP-0.060-0.066(0.091)(0.088)Precipitation-0.195***-0.212***(0.069)(0.070)Constant1.038*1.058*(0.563)(0.560)Observations220220R-squared0.7050.708Number of districts110110 Standard errors in parentheses.*** p<0.01, ** p<0.05, * p<0.1.Table 5: The impacts of changes in infrastructure and environments on the changes in agricultural income per worker and the shares of employmentAnnual change in Income per worker in agricultureAnnual change in % employed in agricultureAnnual change in % employed in industryAnnual change in % employed in serviceAnnual change in the share of working age population 0.007-0.389**0.0330.356***(0.015)(0.165)(0.108)(0.135)Annual change in the share of employed-0.0090.991***0.0010.009(0.008)(0.070)(0.047)(0.051)Annual change in the share of migrants-0.006-0.185***0.056**0.129***(0.006)(0.037)(0.027)(0.039)Annual change in the share of adults completed primary education-0.022*-0.116*0.0560.060(0.011)(0.064)(0.042)(0.051)Annual change in the share of households with access to electricity-0.206-6.503***5.529***0.974(0.298)(2.374)(1.875)(2.464)Annual change of NDVI-0.000-0.016*0.0020.015***(0.001)(0.009)(0.006)(0.005)Annual change of NPP0.0010.007**-0.001-0.006**(0.001)(0.003)(0.002)(0.003)Annual change of precipitation-0.002-0.0280.023*0.005(0.002)(0.020)(0.012)(0.018)Soil erosion-0.000***-0.0000.0000.000(0.000)(0.000)(0.000)(0.000)Distance to market3.70947.685*22.584-70.269**(4.978)(27.399)(22.475)(28.704)Road density1.634-5.962-7.135*13.097***(1.066)(5.039)(3.596)(4.905)LFAL-0.000-0.0010.0000.001(0.000)(0.001)(0.001)(0.001)Flood0.000-0.0000.0000.000(0.000)(0.000)(0.000)(0.000)Drought-0.0000.000***-0.000*-0.000***(0.000)(0.000)(0.000)(0.000)Constant0.000-0.004**0.0000.004***(0.001)(0.002)(0.002)(0.001)Observations101110110110R-squared0.4490.8760.3350.654Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1Figure 1: Poverty Rates by Region, PercentageSource: GLSS3, GLSS5, and GLSS7; GSS and World Bank staff calculations.Figure 2: Poverty Maps in 2000 and 201020002010 Sources: Coulombe 2005; GSS 2015.Figure 3: Flood, Droughts, and Soil ErosionNumber of floods from 1985 to 2011Drought severity between 1901 and 2008 Soil erosion rate (various years)Source: HDDFigure 4: Access to Roads, Markets, and ElectricityRoad density Distance to market (km) Source: HDD Source: HDDShare of households using electricity2010 Population and Housing CensusFigure 5: Correlation with poverty rates (2010)35814001568450053149502762250Significant at 1 percent level.00Significant at 1 percent level.Figure 6: Correlation with agricultural income and the percentage of employment in agriculture (2010) Figure 7: Correlation with the percentage of employment in industry and services (2010) Appendix: Table A1: Post estimation test resultsFixed effect (F test)Random effect (Breusch-Pagan Lagrange Multiplier test)Hausman testSelected modelTable 2 (1)Prob > F = 0.000Prob > chibar2 = 0.000Not conclusiveRandom effectTable 2 (2)Prob > F = 0.000Prob > chibar2 = 0.036Not conclusiveRandom effectTable 3 (1)Prob > F = 0.000Prob > chibar2 = 0.001Not conclusiveRandom effectTable 3 (2)Prob > F = 0.000Prob > chibar2 = 0.041Not conclusiveRandom effectTable 4 (1) Prob > F = 0.083Prob > chibar2 = 0.154Not conclusiveFixed effectTable 4 (2)Prob > F = 0.106Prob > chibar2 = 0.148Not conclusiveFixed effect ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download