The importance of household asset diversity for livelihood diversity and welfare among small farm colonists in the Amazon
2005; Taylor & Francis; Volume: 41; Issue: 7 Linguagem: Inglês
10.1080/00220380500170899
ISSN1743-9140
Autores Tópico(s)Income, Poverty, and Inequality
ResumoAbstract Taking a small farm colony on the Amazon frontier as a case study, this article examines the relationships among household assets, livelihood diversity and welfare. The findings show that: (1) few households diversified into non-agricultural income sources, but those that did also had agricultural incomes comparable to households primarily reliant on agriculture; (2) distinct household assets influence the extent of agricultural and non-agricultural diversity, implying that households with combinations of specific assets were best able to diversify their livelihoods, and (3) while specific types of household assets influence household welfare, livelihood diversity does not exert an additional effect on welfare. A key issue that emerges is that different arrays of assets are important for agricultural and non-agricultural diversity as well as for household welfare, implying that households need diverse assets for diverse livelihoods as well as better welfare. Notes This research was supported by a grant from the US National Science Foundation (SBR-9511965), and the author thanks Adilson Serrão and Alfredo Homma of EMBRAPA/CPATU for support in Brazil, Charles Wood and Robert Walker for support in the US, research team members André Caetano, Roberto Porro, Fabiano Toni, Célio Palheta, Rui Carvalho, and Luiz Guilherme Teixeira, and two anonymous referees of the journal as well as the people of Uruará, for engaging discussions of the issues pursued here. Errors contained herein are the author's responsibility. 'First opportunity' sampling raises questions about sampling bias. Brazilian researchers familiar with the Transamazon corridor found distributions on key variables in the sample (age of household head, length of residence, number of cattle, land area deforested, etc.) to be as they expected. Comparisons between the survey data and agricultural and demographic census data for the same year for Uruará [IBGE, Citation 1998 a; Citation 1998 b] yielded similar figures. For example, rural family sizes were the same (5.6 people), as was the percentage of land under primary forest (65 per cent). Both sources also agreed on the primary agricultural products (for example, rice, beans, corn, manioc among annuals, and cocoa, black pepper, and coffee among the perennials). I conclude that bias is minimal, though likely not absent. In a handful of cases (n = 10), annual and/or perennial crops were unidentified and so were excluded. Agricultural incomes refer to gross incomes and are based on the reported production of each product (in kg), multiplied by the proportion sold, multiplied by 1996 prices in Uruará. Proportions sold ran as follows: rice 55 per cent, beans 40 per cent, corn 39 per cent, manioc 36 per cent, pineapples 91 per cent, sugar cane 91 per cent, tomatoes 93 per cent, watermelons 64 per cent, bananas 73 per cent, cocoa 94 per cent, coffee 65 per cent, oranges 82 per cent, black pepper 97 per cent, coconuts 88 per cent, cupuaçu 71 per cent, mangoes 34 per cent, guaraná 75 per cent. Note that the first four proportions were based on the Uruará survey, whereas the others were based on 1996 agricultural census data for the state of Pará. Cattle income is based on the off-take rate of 12 per cent reported in the 1996 agricultural census for Uruará and assumptions about kg per head slaughtered. The research team observed, but did not record other livestock. This will generate some downward bias in agricultural diversity estimates, but the 1996 agricultural census shows that in Uruará, 84 per cent of agricultural income from livestock was from cattle. One could similarly object that milk production from dairy cattle is excluded. This is another oversight of the Uruará survey. Prices per kg come from the 1996 agricultural census for Uruará and ran as follows: rice R$0.28, beans R$1.52, corn R$0.32, manioc R$0.43, pineapples R$1.08, sugar cane R$0.30, tomatoes R$0.55, watermelons R$2.16, bananas R$2.10, cocoa R$0.85, coffee R$1.01, oranges R$0.08, black pepper R$1.36, coconuts R$0.22, cupuaçu R$0.81, mangos R$0.26, guaraná R$3.86, and cattle, R$2.45. At the time, R$1 roughly equaled US$1. Agricultural incomes are strictly monetary incomes, and do not reflect autoconsumption. Initial wealth is computed as the sum of the products of z-scores for these 13 indicators, multiplied by their respective weights from a principle components analysis. urban house, 0.74; brick walls, 0.48; electricity, 0.63, generator, 0.52; gas stove, 0.63; sewing machine, 0.54; refrigerator, 0.73; radio, 0.48; television, 0.77; satellite dish, 0.68; bicycle, 0.54; and car, 0.50. Initial agricultural capital is computed using the same methods as initial wealth. Factor weights from PCA were as follows: chainsaw 0.79, cocoa dryer 0.50, tractor 0.59. Labour-saving capital is computed using the same methods as the other indexes. Factor weights from PCA were as follows: chainsaw 0.60, insecticides 0.65, fungicides 0.63, and herbicides 0.62. I do not include a variable for tenure status because land titles are required for credit, and credit exerted stronger effects on the outcome variables in this article. I exclude a model of whether a household had agricultural income, because it was weak and unenlightening. Also excluded are separate models of income from annuals, perennials and cattle, as these were broadly similar to the total income model. I exclude models for third party contributions since they were so rare (n = 1 out of 261) and for 'other' contributions since these were ill-defined. Housing quality is computed using the same methods as the other indexes. Factor weights from PCA were as follows: well 0.44, wood walls 0.85, adobe walls – 0.88, and electricity 0.46. Durable goods wealth is computed using the same methods as the other indexes. Factor weights from PCA were as follows: second house in town 0.54, gas stove 0.62, sewing machine 0.55, refrigerator 0.75, television 0.84, satellite dish 0.71, bicycle 0.45, car 0.56. Additional informationNotes on contributorsStephen G.L. Perz This research was supported by a grant from the US National Science Foundation (SBR-9511965), and the author thanks Adilson Serrão and Alfredo Homma of EMBRAPA/CPATU for support in Brazil, Charles Wood and Robert Walker for support in the US, research team members André Caetano, Roberto Porro, Fabiano Toni, Célio Palheta, Rui Carvalho, and Luiz Guilherme Teixeira, and two anonymous referees of the journal as well as the people of Uruará, for engaging discussions of the issues pursued here. Errors contained herein are the author's responsibility.
Referência(s)