CHAPTER SIX
3.0 LITERATURE REVIEW
Thus income alone should be relegated in the measurement of poverty especially in African rural settings and this chapter lends its support to John Williamson's (2003)38 call for a minimum asset bundle (cited in Carter et al. 2005:31). We also posit a minimally adequate asset level (MAAL)/threshold in Chapter Eight as the antithesis of the Minimal Income Question39. While income is easy to use in the measurement of poverty, by the same token, it should not be considered sacrosanct to the extent of relegating critically important dimensions of poverty into obscurity on the poverty lexicon and epistemology.
Assets do matter in the study of poverty.
the following psychological and social effects: household economic stability; decreased economic strain on households; educational attainment; decreased marital dissolution;
decreased risk of intergenerational poverty transmissions; increased health satisfaction among adults; increased property values; decreased residential mobility; and increased levels of civic participation.
Importantly, assets provide both individual and community-wide benefits. One can infer that it is these positive psychological and social effects that underpin and buttress social cohesion within communities. Given this precept, it is possible to reproduce conditions necessary for social inclusion that allow the 'poor to do better' steeped in an asset-focused redistributive project.
According to Naschold (2005:3)40 assets are important on the following grounds; first, the economic well-being of a household is dependent on its stock of assets. '[Thus] in a dynamic sense, it is the accumulation of assets which over time enables households to earn enough income to move out of poverty'. Second, asset levels fluctuate less from day to day than income and thus are closer to a measure of well-being. Carter and May (2001), (cited in ibid,5) argue that assets can be interpreted as measuring the underlying, or structural, well being of a household, whereas income, and to a lesser extent consumption, contain a much larger amount of stochastic variation. These are compelling reasons to experiment with asset-based measures of poverty given their potential to measure the underlying, structural, well-being of a household. According to Naschold (2005:5) the (IFPRI) Pakistan Rural Household Survey and the IFPRI/Addis Ababa Universityl University of Oxford Ethiopian Rural Household Survey (ERHS) used assets such as land, agricultural and financial assets in the analysis of household asset dynamics.
There are methodological problematics associated with asset-based measures one of which is the unit of measurement. Lybbert, Barrett, Desta and Coppock (2004) as cited in Carter et al. (2005:25) aggregated heterogeneous livestock into 'tropical livestock units' using a weighting system that allowed them to aggregate sheep and goats with larger
animals such as cattle and camels. Itis here argued that cost or values can be attached to assets without recourse to equivalisation. In the world of business and commerce, assets which include livestock and poultry, are given a value without recourse to equivalisation.
Thus, even if the goal is to carry out some international comparisons, purchasing power parities (PPP) can still be helpful in this instance in spite of their well documented shortcomings.
Carter et al. (2005) support the asset-based approach to poverty analysis insofar as it enables 'to distinguish deep-rooted, persistent structural poverty from poverty that passes naturally with time due to systemic growth'. Carter et al. (2005:24) argue that identification of the asset-poverty line makes it possible to distinguish structural from stochastic poverty transitions. Based on this reasoning, Carter et al. (2005) further contend that a natural disaster or other economic shock that destroys assets and thereby knocks a household below the dynamic asset poverty threshold could have permanent effects. The presumption that Carter et al. (2005) make is that the poor necessarily have some asset holdings. However, it can be argued that it may be a case of the poor not having any asset against their name. Some poor people do not even have a hoe, let alone a donkey, to their name. While Carter et al. (2005) also observe that the poor's marginal returns to assets are low, militating against their leap forward even in good times, this only goes to substantiate the fact that the poor's asset base is intrinsically fragile at best and non-existent at worst.
Because of their paucity with regard to assets, it is axiomatic that asset poverty is associated with a paralysis to participate in the society, that is, social exclusion and poverty traps bifurcated in various ways. Naschold (2005:2) notes that poverty traps and long term poverty could be eliminated if every household could be lifted above the minimum welfare threshold and safety nets ensured that they remain there, thus, one-off social expenditure will have benefited households in the current and future periods. Thus, asset redistribution should ordinarily be linked to a theory of justice. Naschold (2005:3) further contends that even in the absence of multiple equilibria and poverty traps, there may well be a case for helping the poor to escape poverty through redistributive policies.
This hammers home the importance of such an asset as land which defines the space in which people can operate.
Carter et at. (2005:27) advocate, in the study of asset-based measures of poverty, using a mixture of methods - qualitative and quantitative methods (triangulation). Triangulation is supported by Adato, Carter and May,forthcoming, as cited in Carter et at. (2005:27).
Qualitative analysis can be especially valuable in identifying historical causes of structural transitions that predate initial surveys (ibid, 27). It is also important, for instance, in identifying complementarities of assets and their substitutes including their aggregation in different contexts and cultures of communities and nations hence we lend our weight.
4.0 BRIEF APPRAISAL OF CARTER ET AL. (2005) FGT EXTENSIONS The FGT4! class of decomposable, single period poverty measures are
1 N
(z _u)a
Po.= -
L/i --'
N i=! Z
where N is the sample size, z is the scalar-valued poverty line, u is the flow-based measure of welfare (income or expenditures), I is an indicator variable taking value one if
Uj < Z and zero otherwise, and a is a parameter reflecting the weight placed on the severity of poverty. Po (i.e., a = 0) yields the headcount poverty ratio - share of a population falling below the poverty line - while PI and P2 yield the poverty gap - the money metric measure of the average financial transfer needed to bring all poor households up to the poverty line - and the squared poverty gap, a more distributionally- sensitive indicator of severe poverty.
Carter et al. (2005) recast the FGT measure as follows:
Pa=
~ tIps
Iil+s(Z -U il+s(Ail+s(Ail+s-1 .."AilJa
N i=! s=o Z
Carter et al. (2005:28) argue that this measure integrates the FGT formula with a dynamic framework which measures the discounted stream of expected welfare levels relative to the poverty line using the structural poverty mapping from assets to income or expenditures, Ut (At), and the expected asset dynamics, At+s(At+s-l ," " At), They then generate a dynamic generalization of the FGT class of poverty measures, Pa as follows:
where ~€ [0,1] represents a discount factor and the other variables are natural extensions in the time domain of the static FGT formula. This dynamic version of the FGT measure would rely on familiar flow-based welfare measures such as income or expenditures, but would exploit the structural mapping between assets and flow welfare measures and underlying asset dynamics within the local economy to distinguish effectively between those who are poor but predictably improving and those who are poor and stuck indefinitely in a low equilibrium trap. For the present period, s=O, one could use either realized expenditures or income, Uit or predicted structural expenditures or income, u(Ait) depending on whether or not one believes the true stochastic portion of current income dominates measurement error (Carter et al. 2005),
Carter et al. (2005) also posit a second measure which focuses on the gap between households' current asset holding and those necessary to move them currently and permanently above the asset poverty line. The class of decomposable asset-based poverty measures, Pa,is described by the following formula:
They state that P~ measures employ two terms and associated indicator variables to capture each household's distance from the static and dynamic asset poverty lines, respectively. The first indicator variable,
11-
= 1 if Ai <cd
and zero otherwise, reflects whether the household's ex ante asset stock falls below the static asset poverty line. This captures the asset transfers necessary to make a structurally poor household non-poor in the current period. The second indicator, 1/*= 1 ifcd
< A* and zero otherwise, reflects the possibility that the transfer necessary to make a household structurally non-poor today might not prove sustainable if the prevailing asset dynamics are such that an asset stock ofcd
naturally degrades over time, so that the household would predictably be poor again in short order even after augmentation of its initial endowments tocd.
In this case, they argue that the household's future asset poverty is reflected by the discounted distance betweencd
and A*. They assert that this class of dynamic asset poverty measures will count as asset poor those who are presently structurally poor as well as those who are structurally non-poor but who one would expect to become structurally poor in time.They argue that it reflects the strategic policy questions associated with the Washington Consensus of: what asset accumulation is necessary to lift people out of poverty (i.e. what is