Tracking change: A standard of living index for Matlab

Abstract

This paper is intended to show how an index measuring changes in living standards can be constructed by applying a fundamental economic concept on household expenditure data. Data from a small scale panel survey on 240 households in the four clusters of the Matlab research project is used to show the methodology can be put into practice. The first cluster of villages are those which are covered by both BRAC's Rural Development Programme (RDP) and ICDDR,B's Maternal Child Health and Family Planning (MCH,FP). The second cluster of villages is where only BRAC operates in isolation the third where only ICDDRB implements its activities, and the fourth cluster is where neither organization works. However, inferences on the merits of BRAC and ICDDRB's inputs will be more relevant when the methodology is applied to the large expenditure survey currently being carried out by the BRAC-ICDDRB Matlab research team since individual indices can be created for each cell. The basic theoretical premise underlying this study is that price indices can be used as standard of living indices using an economic tool called 'revealed preference' theory. This theory uses actual household expenditures per person to yield insights into changes in household welfare as consumers 'reveal their preferences' through their expenditure decisions. After data cleaning the items used were rice, daal, vegetable, onion, sugar/guur, chicken, duck, spinach, edible oil, sari and children's clothes for 215 households. The index was computed for the four cells and the results suggest that there has been a rise in living standards in the comparison cell and no change in the other three cells. However these results need to be interpreted with extreme caution and the following qualifications are in order. Firstly the results are based only on two years of data and changes in living standards ought to be monitored only over a longer time period. More importantly the sample size is relatively small. Thirdly the bundle of goods used in this preliminary index ought to be widened to encompass more expenditure items. The issue of a relative scale with which to place the figures in context also needs to be deferred due to the fact that this is only the first of a series of indices which will be constructed using annual panel data. This research paper therefore intends to present a methodological framework from which inferences on standard of living can be made but does not intend to draw conclusions regarding the merits of the various development interventions due to the limited nature of the data. The paper also by no means intends to have the final word on the multifaceted 'standard of living' issue and recognizes that many facets of a household's quality of life may have been ignored by using only one indicator (expenditure) as the basis for comparison.