Contrasting Monetary and non-monetary measures of poverty in developing countries: a survey

While the study of poverty is popular amongst academia, there is yet to be a consensus on what defines such a phenomenon. Basing their arguments on different conceptual frameworks, some argue for a definition in monetary terms, as a lack of income or consumption, while others support a non-monetary approach focusing on shortcomings in diverse dimensions (health, education, nutrition … ). Does it really matter that researchers do not agree on a definition of a poverty? Probing the empirical literature comparing monetary and non-monetary poverty in developing countries, this survey argues that different definitions of poverty lead to different populations identified as poor. Indeed, cross-tabulating poverty headcounts, it finds that a sizable portion of the non-monetary poor are excluded by the monetary approach. Moving to a subgroup level, it finds nonmonetary poverty amongst the monetary richest, and considerable disagreements with regards to poverty risks. In addition, these significant differences are not resolved when switching to a dynamic setting through panel studies. Policy implications are important: the risk mistargeting and misevaluation of policies is high if we do not consider both approaches concomitantly without one taking precedence over the other.

I-Indeed, if monetary and non-monetary multidimensional measures identified the same population as poor, there would not be any issue with using income or consumption exclusively. However, empirical literature has identified a significant mismatch in identifying the poor : focusing on monetary measures might disregard important poverty dynamics and lead to mistargeting.
The empirical investigation of this mismatch has taken many routes. Some have compared the two measures on a macro-level, pointing that positive income trends do not necessarily lead to improvements in non-monetary dimensions (Bourguinon et al (2010); Alkire et. al (2014)). Other have focused on cross-country comparisons: Dreze and Sen (2013), compared India and Bangladesh development trajectories between 1990 and 2011 and found that while India exhibited a higher progression on GDP per capita, Bangladesh far outperformed it on child mortality, immunization, and female literacy.
Another trend in empirical literature has been to compare monetary and non-monetary measures of poverty from the same dataset. This has been allowed by the growing amount of non-monetary indicators captured in the national household surveys (Alkire, 2018). In a survey of high-income countries, Perry (2002) has found that monetary and non-monetary measures only present a modest overlap in identifying poor. For example, amongst the population identified as poor by an income poverty line in New Zea land, only 40% would also be identified as poor by a non-monetary measure of poverty: 60% of non-monetary poor would not be targeted by poverty alleviation policies (which are based on income measures). Boarini (2006) has found comparable results in a survey of the mismatch in OECD countries. However, it is to be noted that these findings stem from high income countries, with specific definitions of deprivation and with a relative definition of poverty (usually 60% of median income).
Some have argued that the mismatch might be due to considering the two measures in a static setting, and thus contrasted monetary and non-monetary indicators in a dynamic setting using panel data (Bradshaw and Finch, 2003, Whelan, 2004, Tran, 2015. Indeed, it can be argued that static comparisons are only a snapshot of poverty and do not disregard the variety of poverty status with time as a variable. However, these empirical investigations have all found that the considerable mismatch, while lowered for some (Whelan, 2004), is still present in a dynamic setting.
To solve this issue, some have argued for using both measures (Bradshaw and Finch, 2003;Bag and Seth, 2018) while other have tried to combine both in an index, with mixed results for policy information (Evans, 2020).
This survey attempts to gather evidence on this empirical effort. As opposed to Boarini (2006) and Perry (2002), it will mainly focus on low to median income countries. These countries have traditionally relied on absolute definitions of poverty (with a po verty line based on basic caloric needs) and thus may present different insights on the mismatch. In addition, this survey includes empirical work that has used a new non-monetary measure of poverty: the multidimensional poverty index. This aggregate introduced by Alkire (2011) has been gaining traction in the field of poverty alleviation, with multiple developing countries adopting it as an official poverty measurement with support from Oxford Poverty and Human Development Initiative (OPHI) (see Alkire,20 15). In a research agenda outlined by OPHI in 2018 (Alkire, 2018), the mismatch between monetary and non-monetary multidimensional poverty is designed as an under researched area for developing countries, while the same phenomenon has received considerable attention in developed countries. This review addresses this gap in the literature.
Based on key findings from said literature, we argue that monetary and non-monetary measures of poverty are inherently different and present a different perspective on poverty and as such, they should be both used as an aid to policy making.
This first part will introduce the topic, while the second part will delve into the conceptual frameworks guiding monetary and non-monetary measures of poverty. The third part will focus on presenting the empirical literature selected and the review process studying while the fourth part will introduce key findings from this corpus. Finally, the fifth part will discuss policy implications of said findings, the reviews' limitations and conclude.

II-Conceptual Framework
Poverty measurements do not take place in a vacuum (Suppa, 2016 ): they are entrenched in conceptual frameworks that tell poverty narratives which explain what exactly should be measured and why (Gamboa, 2020). Given that poverty analysis is a popular field, numerous conceptual frameworks compete and dictate which tool to use: either monetary, non-monetary or a mixture of both.
Each of these frameworks present with advantages and disadvantages. This part will try to assess the ones most invoked in the literature: the monetary approach, the capability approach, and the participatory approach. Contrasting these three frameworks will he lp shed light on the drivers behind the important mismatch in identifying the poor, which is attested by empirical literature.

The Monetary Approach: a. Origin and Definition
As pointed by Ruggeri Laderchi (2003), the monetary approach has been the dominating conceptual framework for poverty measurement since the start of poverty studies. Indeed, this approach goes as far back as the 19th and early 20th century with the works of Booth and Rowntree who studied poverty in England. Rowntree (1902) is credited with the first scientific study of poverty: he defined a poverty line according to a standard adequate diet then added a needed budget for clothing and rent. What is more, the monetary approach is still used by most national governments and international development agencies in defining their poverty alleviation plans (Sumner, 2007).
What exactly constitutes a monetary approach? Under this approach umbrella, are held together a multitude of methods which all share the use of a monetary medium (either income or consumption) as the central tenant to the quest of identifying the poor.
Indeed, the first step when determining poverty rates using the monetary approach is setting a poverty line (Haughton and Khandker, 2009). There is no consensus on how to determine such a line, but two main methods have emerged. One is concerned with the cost of acquiring a basket of food deemed necessary. This cost is imputed using human needs in daily calories multiplied by local market prices. The advocates of this method generally add a multiple to this poverty line, to account for non-food items. Indeed, even, the poorest individuals must dedicate some expenditure to non-food items (Banerjee and Duflo, 2007). When cost of food items is not available, some have argued for the food energy intake method (Haughton and Khandker, 2009): expenditure (or income) is regressed against food consumption to determine the expenditure (or income) level at which food needs are satisfied. This level of expenditure (or income) is used as a po verty line.
Once the issue of poverty line is settled, two mediums can lead to identifying the poor: either measuring all sources of income or recording consumption while affixing prices to any non-market goods. These are usually measured at the household level and then translated into per capita indicators through equivalence scales (Haughton and Khandker, 2009) b. Advantages The main argument for the use of the monetary approach for identifying the poor is its adequacy with micro-economics theory (Ruggeri Laderchi, 2003). Indeed, it satisfies the underlying assumptions of microeconomics by being linked to utility maximization theory: any consumer will maximize its utility (in a homogeneous manner) and consumption (or income) will reflect the value placed on a commodity. This suggests that while monetary measures of poverty are not direct measures of a shortfall in welfare, they are reliable proxies: monetary means will eventually be converted to outcomes or direct measures of poverty. If we take the example of a poor person who suddenly sees its income rising, it is expected that its access to health services will also improve.
Another argument for the use of the monetary approach holds that monetary instruments may be direct measures of poverty. Indeed, Atkinson (1989) argues for a universal basic income as a minimum right. Like basic education, a basic income would in itself constitute a vector of welfare.
Besides these core arguments, technical reasons may also explain the dominance of the monetary approach. Indeed, Sumner (2007) points that development agencies and governments aim for easy-to gather and cost-effective poverty measures. Assembling monetary indicators may seem less costly and more straightforward. Besides monetary indicators are perceived as more "rigorous". This goes to a general finding in development studies research, where quantitative measures are favored (White, 2002).

c. Disadvantages
Arguments for the monetary approach are based on assumptions that may not hold true in real life.
First, it posits that individuals are rational and driven by utility maximization (Johannsen et. al 2007). However, Mullainathan and Shafir (2013) show that scarcity can change individual's behavior turning them into irrational agents. They find that poverty "reduces a person's cognitive capacity more than going one full night without sleep" (Mullainathan and Shafir 2013, p. 13). In addition, Banerjee and Duflo (2011) found that some income poor people prefer to buy a television or a cellphone rather than food or cheap immunization for their children: few would argue that this constitutes a rational behavior.
Second, it assumes that individuals' preferences are homogeneous (Thorbecke, 2007) But can we really attest that individuals' utility functions are essentially the same? What about socio-demographic and cultural variations?
Third, it posits that utility only streams from market goods (Thorbecke, 2007), thus ignoring any positive or negative externalities streaming from non-market goods. This is particularly problematic in countries with important social welfare programs: how to account for free education and health?
Fourth, it does not account well for the chronicity of poverty. Spicker (2007) explains that chronicity is important because transitory poverty does not account for total poverty.
Given that Income and Consumption are more volatile compared to non -monetary measures, they might not be a good proxy for identifying the chronic poor. This is in line with Clark and Hulme (2005) argument that flows, such as income, are more time variant than stocks, such as housing conditions or educational attainment. It is based on a flowfund theoretical framework developed by Georgescu-Roegen (1971) elaborated to analyze the economy. Funds are the agents that are used, but not consumed, in the economic process. They are transformative agents such as labor power, land, and capital.
Flows are the elements transformed by the funds (income for example). Building on this framework, Carter and Barrett (2006), showed that while studying poverty, focusing on either flow or fund makes a considerable difference. An individual may be structurally poor because he does not own any "fund" to produce a "flow" of income above poverty line. But he may well be just transient poor since he owns a fund but only failed but the flow production failed temporarily.
Fifth, it does not account for non-trade off cases in poverty. Indeed, the monetary approach assumes trade-offs between the different poverty dimensions. But this is not always the case. For example, increasing consumption does not substitute for a deterioration in health dimension, say losing a child. Hence, Scheidel (2012) argues that the monetary approach is in the realm of "weak sustainability". A "strong sustainability" approach would account for non-trade off cases by building on several types of capital: in health, education, environment… Finally, another argument against using income and consumption in identifying the poor is that monetary tools are more subject to measurement errors. Measuring consumption and income is more likely to include errors of response (Evans, 2020), than relying on direct measurement of goods or deprivations.

a. Origin and Definition
Recalling the works of the field pioneers, one might argue that their approach was not exclusively monetary. Indeed, Booth's study of the east end of London, though mainly concerned with monetary poverty, also documented multidimensional poverty in a sense, because he reported on such metrics as the nature and regularity of employment (Ruggeri Laderchi, 2003). Rowntree (1902) defined a "secondary poverty" category, for those in "obvious want and squalor". These are the ones that fall above the monetary poverty line but still present with what Rowntree saw as direct expressions of poverty he could not ignore.
However, Amartya Sen was the first to introduce a competing framework to conceptualize poverty based on non-monetary measures: the capability approach. He sees development as the bettering of each individuals' capabilities, not as the maximization of utility (Sen, 1985;1999). At the basis of this approach is the idea of functionings. They are the "doings and beings" that an individual values: being well-nourished or welleducated for example.
Sen argues that functionings are as diverse as our societies. The variability even goes to the level of individuals and he defines the set of all functioning vectors that one individual can achieve as a capability set. Freedom is also at the core of the capability approach, as Sen outlines that an individual may or may not choose a certain functioning or capability set: the difference between a starving person and a fasting person is that one chooses not to eat, while the other has nothing to eat. As such, poverty is defined within this framework as a capability deprivation resulting from a lack of one or more functionings. Given the diversity of functionings, the capability approach is adamant on the multidimensionality of wellbeing.
This does not mean that unidimensional measures such as consumption have no place in this framework. Indeed, monetary measurements play a role in this CA as a mean to achieve certain characteristics: money can be used to purchase a certain amount of calories (Lancaster, 1966(Lancaster, , 1971. These characteristics may be converted into functionings depending on the intrinsic constitution of the individual (its metabolism, diseases…), socio-economic variables and the environment (different climates). In addition, public goods and non-market goods play an important role as they help individuals achieve a certain level of functioning. (see Figure 1).
How should we measure poverty then? According to the capability approach, we should measure poverty directly in terms of functioning achievements. In other words, we should investigate shortfalls in the realization of minimal needs for different dimensions: the achievement of basic education, access to health etc.

b. Advantages
The Capability approach presents with numerous advantages.
First, its underlying assumptions are more realistic: it acknowledges market failures and public goods and how they influence wellbeing. For example, a large social welfare program has a considerable influence on individuals well-being and its consequences would be measured in a capability-based poverty assessment. It also holds that individual's preferences are heterogenous (Thorbecke, 2007;Kuklys and Robeyns, 2004) and thus is more flexible than the monetary approach, leaving room from sociodemographic and cultural variability.
Second, it accounts better for the complexity and multidimensionality of poverty (Kakwani and Silber, 2008). As Amartya Sen puts it: "Human lives are battered and diminished in all kinds of different ways, and the first task … is to acknowledge that deprivations of very different kinds have to be accommodated within a general overarching framework" (2000). Indeed, it is not likely that poverty depends only on monetary measures: Kahneman and Deaton (2010) found that a larger income does not necessarily improve emotional wellbeing. Other non-monetary dimensions play a role, and the capability approach puts them at its core.
Third, it looks directly into outcomes of poverty instead of resource inputs. Resource inputs might be less accurate since they might not be converted into wellbeing functionings due to of externalities or individual's own choices. Thus, it is closer to the real situation of the poor and improves the fit between statistics and reality.

c. Disadvantages
When researchers try to operationalize the capability approach, a few issues arise.
First, there is the issue of how to define a universal set of capabilities to be assessed.
Indeed, Sen does not provide a list of universally essential capabilities, to leave room for variations across societies. Nussbaum (2000;p74) has tried to define these minimal capabilities, hoping for something that is not "the mere projection of local preferences but it fully international and a basis for cross-cultural attunement". These minimal capabilities include: the right to "Play", "respect for and living with other species" and political "control over one's environment". It may be argued that these are the projection of local western preferences and not universally essential capabilities. However, there seems to be a consensus on at least three dimensions: health, nutrition and education (Dreze and Sen, 1995).
Second, researchers have struggled with measuring capabilities, as potential outcomes that may not have been realized yet (Ruggeri Laderchi et al., 2003). One might argue that Third, issues of segregation of the poor and aggregation of dimensions arise. Indeed, and as with the monetary approach, the choice of a poverty line might seem arbitrary (Ruggeri Laderchi et al., 2003). In addition, the challenge of aggregation arises beca use of the multidimensionality of poverty according to this framework: how to account for various kinds of poverty in a single index? One can argue for a dashboard style reporting of all deprivations, but aggregation has its advantages as it allows for cross-country comparisons and inform on the severity of poverty.
Last, many critics point out that the capability approach might be very demanding in terms of data and thus inoperative (Sugden, 1993;Comim, 2008). Some scholars continue thus to advocate for monetary measures as proxies for poverty measurement (Meyer and Sullivan; even though these measures might be lacking in precision given the different conversion factors that lead them to well-being.
3. The Participatory Approach: a. Origin and Definition As pointed above, both the monetary approach and the capability approach have been criticized for their "arbitrariness" in defining poverty and poverty cutoffs. This might be to their externality to the poverty phenomenon: the views of the poor themse lves on what constitutes poverty are not solicited. This posture is deliberate, as external measurements of poverty are perceived to be more objective (Ruggeri Laderchi, 2003). On the contrary, the participatory approach argues for a deliberately subjective definition and measurement of poverty.
Pioneered by Chambers (1994Chambers ( , 1997 this approach was first rolled out in small projects with initial workshops comprising the poor: a subjective definition of poverty would emerge as well as tools to evaluate poverty severity. Poverty is thus equated with a shortfall in subjective life satisfaction. This approach was scaled up by the World Bank and by 1998, half of funded projects included a participatory assessment. A multitude of tools have been devised to operationalize this approach. These include wealth and well-being ranking, participatory mapping and modelling and seasonal calendars. Assessment tools can be used flexibly according to local context and no rigid methodology is devised (Ruggeri Laderchi, 2003).
In addition, scholars have explored other empirical adaptations of this approach. While previously described poverty assessments tools are based on direct assessments of life satisfaction, Kahneman (2011) choose to measure individuals affect and whether they experience stress or disappointments to draw inferences on the person's subjective state.
Chemin (2013) bases its measurements on biological manifestations such as cortisol levels to determine said subjective state.
While non-monetary tools tend to take the center stage in this approach, income or consumption have also been invoked as important poverty dimensions by themselves, as thus used in poverty assessments. For example, Mitra (2014) while devising a multidimensional poverty index for Nepal, adds consumption poverty as a dimension (with 25% weight) because monetary poverty has emerged as important in a national subjective poverty assessment survey.

b. Advantages
As explained in the previous section, the poverty approach presents with the m ain advantage that it is internally based and thus might present with a more realistic assessment of poverty. It is also less rigid and draws on variety of assessment tools, devised in collaboration with the poor.
In addition, the use of the participatory approach might be conducive to more cooperation in development projects (Ruggeri Laderchi, 2003): the poor feel empowered and thus adoption rates of policies go up.

c. Disadvantages
However, this approach is also mined with serious flaws.
First, internal assessment of poverty might be biased. Sen (1999) argues that due to large chronic deprivations, individuals start to adapt their expectations, thus "lowering their gaze". This adaptation leads them to assess in a more positive light objectively terrible living conditions. On the other hand, some well-off individuals, and by the same adaptation mechanism, might "raise their gaze" and produce a negative life assessment (Stewart, 2014). Social conditioning and limited information play a key role here, as well-off people might be isolated from poor people and vice-versa.
Second, while the participatory approach pretends for subjectivity and internality, it is always external observers who conduct the entire process of measurement of poverty.
External bias might thus still be present. Third, da Cuhna (1997) argues that the participatory approach reinforces existing social stratifications and thus might worsen the exclusion of lower castes. Indeed, given that poverty assessment takes places in the context of the community, opposition to current doxa might not be encouraged for fear of retribution.

III-Methodology
As we have seen in the previous section, scholars rely on conceptual frameworks to determine which tools to use when identifying poverty. While the monetary approach calls for the use of monetary tools, the capability approach favors non-monetary ones while the participatory approach gives the poor precedence in choosing how to assess their own poverty. These differing frameworks eventually lead to discrepancies in poverty assessments. As mentioned in the introduction, the study of this mismatch has taken many empirical routes, one of which is comparing both monetary and nonmonetary poverty within the same dataset. This paper is concerned with surveying these comparisons with a focus on developing countries, in order to draw generalizations.
To do so, a total of 16 studies focusing on the study of the or between monetary and nonmonetary poverty in developing countries were selected. These studies have used a multitude of indicators and methods to study the mismatch itself. This section will introduce the methodology used to select these studies.

Review process
Using multiple keywords ("monetary poverty", "multidimensional poverty", "non - This goes to show the difference in analyzing poverty between developed and developing countries, as more advances economies tend to use income as the main monetary tool to assess poverty with a 60% relative poverty line (Boarini, 2006). Indeed, given the importance of the informal sector in developing countries, income may not reflect the state of monetary poverty as well as consumption. Following Meyer and Sullivan (2003) we know that an individual may refrain from reporting an illegal source of income, but the effect of that income stream would still appear in his consumption patterns.
With regards to non-monetary indicators, while earlier work tends to focus on single dimensions (Baulch and Masset, 2003) or Principal Component Analysis derived weighted dimensions (Klasen, 2000) most of the selected studies use multidimensional poverty indexes (MPI) based on Alkire-Foster methodology (Alkire and Foster, 2011).
Following the Human Development Index, the MPI measures deprivations in the three main consensual poverty dimensions which themselves cover ten indicators: health (child mortality, nutrition), education (years of schooling, child enrolment) and standard of living (electricity, drinking water, sanitation, flooring, cooking fuel and assets). A dual cut-off method is used: within each indicator a household is considered deprived or not given a specific cutoff; then a household is considered poor if deprived in more than a third of all weighted indicators. This methodology presents with numerous advantages as it is decomposable (it is possible to extract which dimension contributes the most to poverty) and multidimensional. Besides, it is also adaptable and allows for local variations, in accordance with Sen's original desire to not put forth a set of rigid poverty dimensions. Indeed, besides the global MPI which was described previously, national governments and researchers have tweaked the dimensions and cutoffs according to their own socio-economic context: This is the case in Nepal where Mitra (2014) has chosen to disregard access to drinking water, as this does not constitute a major hindrance in a country rich of freshwater resources.  Only one study finds different overall trends over time: Bruck and Kebede (2013), analyzing the case of Ethiopia, find similar decreasing trends for two time periods while at the third interval consumption poverty seems to go up, while multidimensional poverty keeps decreasing.
If monetary and non-monetary poverty, analyzed from the same dataset, seem to move in the same direction and sometimes present with similar headcounts, then an argument can be made for using monetary tools only. However, the next sections will show that this similarity rapidly breaks down with a more careful analysis.

However, the crosstabulation of matching headcounts reveals a significant mismatch between monetary and non-monetary poverty:
To  Looking at exclusion errors numbers, it appears that every study documents a significant mismatch between monetary and non-monetary poverty. What these numbers mean is that using the monetary approach exclusively would result in disregarding between 8.5% (Nepal) and 75.4% (China) of the non-monetary poor. This is the case in Nepal as well (Mitra, 2014) which has an 8% average exclusion error despite including consumption as a dimension in its multidimensional measure.
In addition, there does not seem to be a clear ordering of countries according to their development level, though we tend to see higher exclusion errors in less developed countries (Ethiopia, Uganda) than in emerging countries (India, Brazil, Indonesia) with the notable exception of China.
It also appears that inclusion error is less important than exclusion error. Taking the example of Uganda, Levine (2012) finds that if only consumption is used a measure for poverty, 44% of non-monetary poor would be excluded. If we use a multidimensional index of poverty (which does not include consumption or income as a dimension), only 5% of consumption poor would be disregarded.
One might argue that the mismatch may be resolved with different cutoffs applied to measures, but its multidimensional index includes consumption with a 25% weight. Klasen (2000) finds a high correlation between both measures in South Africa but signals that this might not be generalizable because of the specific apartheid legacy: government policies propelled favored groups' consumption and non-monetary indicators while black people were ignored for both. Overall, these correlations seem to be weak, especially when considering monetary tools as a reliable proxy for direct measurements of poverty.

Population Subgroup analysis
Another way to look at the mismatch between monetary and non-monetary poverty is to look more closely at subgroups of the population. This analysis will help us further our understanding of the differences between both measures and grasp their potential for informing policy. The empirical literature has used different tools to probe the mismatch at the subgroup level, amongst which poverty risk analysis and quintile analysis. The   We have now seen how monetary and non-monetary poverty present with a significant mismatch and disagreements on poverty risks. However, these analyses were performed in a static setting and chronicity is an important feature of poverty. Given that monetary measures such as income and consumption are flow variables and thus more prone to volatility than fund variables, it may be argued that the mismatch between both types of poverty would dissolve eventually. The only way to verify this hypothesis is through a panel study following the same individuals for a certain period.
The selected literature includes such panel studies, and a general finding is that the mismatch between monetary and non-monetary poverty is not improved when switching from static to dynamic settings. Indeed, Baulch and Masset (2003), using Chi-square tests of independence, find that the distribution of consumption poverty is as different from that of food and education poverty in both dynamic and static setting. Also in Vietnam, Tran (2015) finds that among those who escaped consumption poverty, only 17% also rose from non-monetary poverty. Considering a representative Chinese panel of 1537 households for 9 time-periods, Alkire and Fang (2018) found that only half of the chronic monetary poor were also chronic non-monetary poor. Looking at Ethiopian panel data, Seff (2016) finds that consumption poverty does not predict multidimensional poverty and vice-versa: 41% of people whose multidimensional measure worsened saw their consumption poverty improve while 53% of people whose multidimensional measure improved, saw their consumption poverty worsen.
In addition, our hypothesis that monetary measures present with disproportionally higher volatility compared to non-monetary measures does not seem to hold true. While, nonmonetary poverty does present with slightly more stability, it is also surprisingly dynamic. This is the case in Ethiopia, where Bruck and Kebede find that only 6 4% of individuals are chronic non-monetary, a slight increase from the 53% consumption poor.
These surprising dynamics might be due to people relocating, health deterioration or generational change.
Furthermore, monetary and non-monetary measures seem to present with different properties in a dynamic setting, notably with regards to shock predictability. Indeed, Seff (2016) finds that Ethiopian households who fell into non-monetary poverty were 15% more likely to have experienced a shock between the two waves of panel. This predictive ability is not found in consumption poverty.

V-Conclusion
In the previous section, and based on included literature, we have found that although monetary and non-monetary measures may present with similar overall trends, a more careful analysis reveals an important mismatch between populations identified as suffering from poverty. Moving to subgroup analysis, this mismatch is clearer, with nonmonetary poverty present amongst the richest monetary quintile, and important disagreements on poverty risks between both measures: being poor in one measure is associated with different independent variables than for the other measure. Changing our gear to a dynamic setting, we find that the mismatch issue is not resolved, and the theory of convergence over time does not hold.
What policy implications can be drawn from these findings? Finally, it should be noted that the study of the mismatch in a dynamic setting has only been carried out in a few studies given the scarcity of panel data. Further research is needed to assess how the two measures differ in the long-term and how they react to targeted policies. Until then, it would be safer to use both measures concomitantly in all policy formulation and assessment, without one taking precedence over the other, to get a complete picture of the poverty phenomenon. Figure 1 Inulin-type fructan with β-2,1-bonds.

Figure 3
Schematic ow sheet of a conventional bioethanol production; Wet distillers grain (WDG) Figure 4 Schematic diagram of AOAC method 997.08 (l.) and 999.03 (r.); Saccharides S (glucose and fructose). Response surface plots for the fructan content (l.) and HMF formation (r.) both based on stillage dry matter depending on temperature in °C and acid concentration in wt% (H2SO4) for a hydrolysis time of 80 min and a subsequent neutralisation with CaCO3.