Why Does the Engel Method Work? Food Demand, Economies of Size and Household Survey Methods
2002; Wiley; Volume: 64; Issue: 4 Linguagem: Inglês
10.1111/1468-0084.00023
ISSN1468-0084
Autores Tópico(s)Fiscal Policy and Economic Growth
ResumoOxford Bulletin of Economics and StatisticsVolume 64, Issue 4 p. 341-359 Why Does the Engel Method Work? Food Demand, Economies of Size and Household Survey Methods John Gibson, John Gibson Department of Economics, University of Waikato, Private Bag 3105, Hamilton, New Zealand. E-mail: [email protected] *I am grateful to Ron Bewley, Angus Deaton, Les Oxley, Peter Phillips, Scott Rozelle, Christopher Scott, Jeffrey Williams, Fred Zimmerman and an anonymous referee for helpful advice. Data for this research are part of a World Bank poverty assessment for Papua New Guinea, for which financial support from the governments of Australia (TF-032753), Japan (TF-029460), and New Zealand (TF-033936) is gratefully acknowledged. All views in this paper are those of the author and should not be attributed to the World Bank.Search for more papers by this author John Gibson, John Gibson Department of Economics, University of Waikato, Private Bag 3105, Hamilton, New Zealand. E-mail: [email protected] *I am grateful to Ron Bewley, Angus Deaton, Les Oxley, Peter Phillips, Scott Rozelle, Christopher Scott, Jeffrey Williams, Fred Zimmerman and an anonymous referee for helpful advice. Data for this research are part of a World Bank poverty assessment for Papua New Guinea, for which financial support from the governments of Australia (TF-032753), Japan (TF-029460), and New Zealand (TF-033936) is gratefully acknowledged. All views in this paper are those of the author and should not be attributed to the World Bank.Search for more papers by this author First published: 16 October 2002 https://doi.org/10.1111/1468-0084.00023Citations: 51 Anand and Harris (1994) review the empirical performance of various welfare indicators and conclude that the food share has little merit as an overall welfare indicator. The respondent was usually the senior female, especially for food expenditures. However, for certain items like alcohol and tobacco, the senior male usually provided the information. The principle of asking the respondent best informed to answer questions regarding each type of consumption is widely used by recall surveys. With an average food share of 0.5, a 20 percent decrease in per capita food expenditures isimplied. It is perhaps no coincidence that the two countries in Deaton and Paxson's sample with the least puzzling results (Britain and France) collect household expenditures using the diary method, while the other countries gather data by recall. A household with two people makes an average of 25 non-food purchases per fortnight while a household with 10 people averages 50 non-food purchases per fortnight. This result is also found by the field survey described below. The foods whose expenditures recalled by respondents were lowest, compared with expenditures reported in diaries, were the foods that were most frequently purchased, and this correlation is significant at the p<0.01 level. An alternative hypothesis, that recall errors reflect respondent fatigue, is not supported because there was no significant association between the placement of a food in the questionnaire and the size of the recall error. For example, the "intergenerational family" (the head and their spouse, the parents and parents-in-law of the head, children of the head and their spouses, and grandchildren of the head) comprises 78 percent of the residents of the average household in urban areas of Papua New Guinea. A unit increase in the logarithm of household size raises the share of residents who are not part of the intergenerational family by 14 percentage points. By rewriting β ln (x/n1−σ) as β ln x−(1−σ)β ln n it is clear that β ln (x/n)+βσ ln n=β ln (x/n1−σ). Rodgers, Brown, and Duncan (1993) report evidence that errors in survey responses are not independent of either the true values of the variable being measured or of other explanatory variables in the causal model. These values were drawn afresh at each experiment. Values for household size were rounded to the nearest integer, with the minimum constrained to n = 1. The mean food budget share generated by these assumptions was 0.50 (standard deviation = 0.16). The means and standard deviations of the generated x, n, and wf variables match the data collected in the field. Deaton and Parxon use IV to deal with random measurement errors in ln (x/n) that might bias the γ coefficient because of the correlation between ln (x/n) and ln n. Illiterates were typically the elderly, so sometimes other family members recorded in the diaries on their behalf or interviewers visited more frequently than usual (roughly every second day) and did the recording. This is also true when the unweighted data are used. For the recall sample, the OLS coefficient on household size is −0.1000 and the IV coefficient is −0.0928. For the diary sample, the OLS coefficient is −0.0061 and the IV coefficient is −0.0237. The size classes are used to give smoother estimates because of the small sample size. For the same reason, the poverty rates are calculated from all areas of Papua New Guinea rather than just from the capital city. Other methods based on estimation of a system of demand equations (for example, see Lancaster et al., 1999) may be too involved for many applied economists who just want to estimate an equivalence scale as an input into some other modeling task. Read the full textAboutPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Citing Literature Volume64, Issue4September 2002Pages 341-359 RelatedInformation
Referência(s)