Income and non-income Pro-Poor Growth in Colombia from 1996 to 2995: a multidimensional approach

Transcription

1 Income and non-income Pro-Poor Growth in Colombia from 1996 to 2995: a multidimensional approach Adriana Cardozo, Melanie Grosse This version: August 24, 27 Abstract By using the multidimensional approach of pro-poor growth, this paper explores if economic growth in Colombia was benecial to the poor from 1996 to 25. By using income and non income indicators, we draw growth incidence curves and derive pro-poor growth measures to give a broader view on how poverty evolved before, during and in the aftermath of the 1999 economic recession. Results show that income growth was pro-poor since the pro-poor growth rate was higher than the mean growth rate (.94 versus.4), a dierence that is too small and varied considerably by areas and regions. The multidimensional approach of pro-poor growth showed a very mixed picture depending on the variable and level of disaggregation. Education improved nearly everywhere, asset accumulation increased more in richer households and quality of housing and access to services improved more in middle to upper middle classes. Furthermore, we nd evidence that using income or consumption makes a much stronger dierence than suggested by the current pro-poor growth discussion which treats both equally (mainly because either income or consumption are available in the data sets). JEL Classication: D, I, O1, O12. Key words: Pro-Poor Growth, Inequality, Welfare Measurement, Multidimensionality of Poverty, Latin America, Colombia University of Göttingen, Department of Economics, Platz der Göttinger Sieben, 77 Göttingen, Germany, acardoz@uni-goettingen.de; mgrosse@unigoettingen.de. The authors would like to thank the participants of the Network on Inequality and Poverty, Colombian Chapter for helpful comments and discussion.

2 Abstract Based on data from Colombian household surveys (HHS) as well as quality of life surveys (QLS), this study evaluates whether growth in Colombia was pro-poor from 1996 to 25. The precise objective is to examine in a multidimensional approach how the poor experienced the negative eects of this crisis as compared to the non-poor, investigating whether they shared equally in the benets of growth.the novelty of the study is the twofold analysis of poverty in Colombia by using the pro-poor growth measurement approach developed by Ravallion and Chen (2), which focuses on the income dimension of poverty, and its extension to the non-income dimension developed by Klasen (26). Income growth incidence curves (GIC) are drawn for the period 1996 to 25 by using the yearly available household surveys, while non-income growth incidence curves (NIGIC) are calculated from 1997 to 2 by using the corresponding Quality of Life Surveys (QLS). 1 Household survey data allow for subdividing the income analysis into three sub-periods 1996 to 1999, 1999 to 22, and 22 to 25 thereby presenting a better picture of the recession's eects, as well as of how fast incomes recovered from it. Calculations are done cumulatively as well, taking 1996 as a base year and comparing each of the subsequent years with it to see how income growth evolved over the period as a whole. Results show that by interpreting the aforementioned methodology strictly, income growth in Colombia from 1996 to 25 was pro-poor since the pro-poor growth rate (PPGR) was higher than the growth rate in mean (GRIM) (.94 versus.4). Nevertheless both growths rates and the dierence are very small and the methodology does not indicate anything about how large this distance needs to be to conclude that growth was truly benecial to the poor relative to the non-poor. A more cautious conclusion would be that growth was very low and was not averse to the poor (relative to the non-poor), leading to an almost unchanged incidence of poverty in comparison to the level observed in Subdivision of results into smaller periods gives a more precise picture. From 1996 to 1999, the economic slowdown aected the poor much more severely than the non-poor. In both urban and rural areas, incomes fell more drastically for the extreme poor. Between 1999 and 22, all GIC have a positive slope, and growth was pro-poor as shown by the PPGR, a result strongly inuenced by a statistical eect of income moving back to the levels observed before the economic crisis. In the period 22 to 25, the economic recovery showed stronger gains, and income growth was propoor in urban and rural areas as well as for the nation as a whole. Regional comparisons show that it is only in the Pacíco region that growth was clearly anti-poor, while in the others, the national results hold: low income growth rates and small dierences between the GRIM 1 The QLS provide an alternative source of comparison between income and consumption as well as a richer dataset for calculating non-income growth incidence curves.

3 and the PPGR. Contrasting income and consumption growth between 1999 and 2 by using the QLS, we nd strong dierences between both monetary measures. While income is pro-poor in most of the cases in the relative and weak absolute sense, consumption data is nearly always pro-poor in the weak absolute sense but never in the relative sense. This questions the "equal treatment" of income versus consumption as "equally suitable" for pro-poor growth analysis. Turning to non-income pro-poor growth, we nd a very mixed picture depending on the level of disaggregation. The overall picture is quite positive but with some interregional dierences. Education improves nearly everywhere and in most of the cases this progress is propoor. Asset ownership, housing quality and access to basic services also improved, but not in the relative sense, meaning that the percentiles below the poverty line did not benet more compared to the richest. The health status shows progressed as well, but also with some worrisome exceptions. Only "life" in Colombia, measured by some life satisfaction and subjective welfare variables, seems to be have improved all-over in the period of analysis.

4 1 Introduction As a central topic in current development debates, analysis of the relationship between poverty reduction and economic growth has driven to the measurement of "pro-poor growth" by which researches try to narrow down this discussion to the concrete measurement of the growth eects on poverty. Departing from an implicit consensus of the fact that initial conditions like inequality, assets, natural resources and institutions have an important inuence on the poverty reducing ability of growth, the pro-poor growth analysis focuses on explaining in how far poverty changes can be explained by growth and by which channels growth might drive to more eectively reducing it. By using the most recent developed methodologies on pro-poor growth measurement, this study takes the Colombian data to examine whether economic growth reduced poverty between 1996 and 25 analyzing both the income and non-income dimensions of poverty. The well-known but often ignored multidimentional approach of poverty (Sen 1998) is taken here into account by using Klasen (26) extension to Non-Income Growth Incidence Curves (NIGIC). The income dimension is analyzed applying Ravallion and Chen s (2) methodology for drawing Growth Incidence Curves (GIC) and calculating pro-poor growth rates. The main advantage of these methodologies is that they allow to track changes in consumption, progress in education, health, access to services and housing quality, along the whole distribution of income, providing more accurate information than just looking at usual averages. The period of analysis chosen is of high interest given that the country experienced a complete economic cycle. The expansion that started in 1992 ended with a drastic output contraction in 1999, and was followed by a slow recovery process since then. Poverty reduction achievements by 1999 reversed and unemployment escaleted. For the income dimension, the data sources are the Encuesta Nacional de Hogares (ENH) 1996 to 2, and the Encuesta Continua de Hogares 1

5 (ECH) 21 to 25, conducted annually by the National Administrative Department of Statistics (DANE). 2 In order to have a broader picture of the recession's eects as well as of how fast incomes recovered from it, we divided the period into three sub-periods 1996 to 1999, 1999 to 22, and 22 to 25. For the non-income dimension, we use the Encuesta de Calidad de Vida (ECV) from 1997 and 2 (the only years in which they were conducted) similar to Livings Standard Measurement Surveys (LSMS). This type of survey has detailed questions concerning both income and consumption, but also on non-income dimensions of welfare such as household characteristics, service availability, education, health, life satisfaction and subjective welfare. The paper is divided into ve sections. Section 2 reviews the stylized facts on economic growth, poverty and progress in education, health, access to basic services and housing quality. Section discusses the denition and measurement of pro-poor growth, both for the income and non-income dimension and gives a more detailed overview of the data and indicators used. Section 4 analyzes the resulting GIC and NIGIC and Section 5 concludes. 2 Colombia: stylized facts on economic growth poverty and non-income indicators 2.1 Economic growth From 1991 to 25 Colombia had a complete economic cycle: an expansion from 1991 to 1994, contraction of the product starting 1996 and up to 1999 when the crisis hit its peak, and from 21 onwards economic recovery. The beginning of this cycle was marked by the reforms introduced in 1991, including state modernization, a constitutional reform, and key changes in the labor, nancial, and exchange rate markets. Barriers to foreign direct invest- 2 Although the main objective of these household surveys is the construction of labor market indicators, they are also used to calculate the incidence of poverty since they are the only ones providing yearly information on household incomes. 2

6 ment as well as to capital exports were removed and the exchange market liberalized by eliminating controls to the foreign currency trade (Parra and Salazar 2). The liberalization eorts fostered two important eects: a surge of foreign capital and a continuous increase in public spending due to the obligations stated in the new constitution. Taken together, these eects helped appreciation of the real exchange rate, channeling the new incoming external resources and investment mainly to the non-tradable real estate sector (Tenjo 2). Starting in 1994, growth slowed down due to a mix of a weaker aggregate demand, political uncertainty, falling coee revenues, recession in the neighboring country of Venezuela and increasing interest rates. By that time, the boom in real estate markets began to wane, housing prices fell, and borrowers began facing diculties in paying their debts due to the previously generated asset mismatch (Tenjo 2). After a short reactivation in 1997, capital ows to the country collapsed as result of the diculties facing international nancing markets due to the Asian and Russian crisis. The reversal of ows led to a breakdown of the credit channel and evidenced the already existing fragilities in the domestic private sector. To defend the existing target zone from a speculative attack, the Central Bank increased active interest rates contributing to the sharp deterioration in portfolio indicators and balance sheets. Higher interest rates and higher risk perception drove to a credit crunch and the recession reached its peak at the end of 1999 when real GDP fell by 4, percent (Nuñez 26). The immediate eects of the crisis were bankruptcies, reorganizations of rms, and a dramatic increase in unemployment, that reached 2,1 per cent in This period coincided with a marked escalation of internal conict after a failed process of peace talks between the government of Andrés Pastrana ( ) and the FARC (Revolutionary Armed Forces of Colombia). Parallel to this, confrontations between paramilitaries and guerrillas worsened considerably, as did drug-related violence. A

7 From 2 to 22, economic growth remained low (2.1 percent), unemployment high, and public spending as well as credit to the private sector constrained. Increased condence due to the slowdown of the internal con- ict, favorable international conditions given by high oil and commodity prices as well as low interest rates in developed countries, revived capital ows raising private investment and easing credit. Annual growth averaged 4.6 percent in 2-25, private consumption accelerated, and unemployment decreased to 12 percent by 25. Similarly to the boom experienced in the early nineties, one of the most dynamic sectors in this reactivation phase has been construction. 2.2 Poverty and Inequality From 1978 to 1995, Colombia managed to reduce the incidence of poverty by ten percentage points. This reduction, however, reversed during the crisis of 1999 with poverty almost returning to 1988 levels (Figure 2). The impact of the recession on households was visible in continuosly falling per capita income and in high unemployment rates. Income losses began in 1996 and ended up in a 9 percent drop in real household income by 21 (Figure ). The u-shaped per capita income graph gives evidence of recovery since 22 also with regard to the incidence of poverty, which decreased to 49 percent by 25. All poverty indicators (poverty incidence, gap, and severity of the Foster, Greer, and Thorbecke (1984) family) show a similar trend, increasing up to 1999, briey slowing down from 2 to 21, rising again in 22, and improving since then. By 25, poverty and inequality indicators at an aggregate level as well as real income had returned to the levels of the early to mid-nineties, or even did better (Table 4). The urban/rural divide shows that at least during the recession, the change in policy has taken place under the administration of President Alvaro Uribe (in oce since 22), whose eorts have concentrated on ghting the guerrillas directly while negotiating the disarmament of paramilitary groups. 4

8 poverty and indigence in rural areas increased much more than in urban areas, widening the gap between both areas. Although rural poverty decreased by more than 1 percentage points from 1999 to 2, it began escalating again since then (Table 4 and Figures 4 and 5). Falling rural wages and employment in non-agricultural activities as well as decreasing agricultural productivity explain this trend. 4 In the labor market, indicators show that occupation conditions deteriorated sharply during the crisis. Unemployment moved from 1 percent in 1997 to 17 percent in 1999 and 16 percent in 22. Although unemployment began decreasing in 22 in line with economic cycles, in 25 it was still higher than in 1996 with lower utilization of the labor force, higher underemployment (from 17.1 percent in 1997 to 7.4 percent in 25) and higher part-time employment (Figure 6). Weak improvements in real incomes, especially for skilled workers, and decreasing formal wage employment explain this trend (Farné et al. 26). 5 Income inequality in Colombia had decreased in the 196s and 197s but reversed after 1988 (Nuñez et al. 22). This latter trend intensied from 1996 to 1999 when the Gini coecient increased to 6. Although improving to.58 in 25, Colombia's income inequality is extreme in the international context, although relatively moderate in comparison to the Latin American one, where it ranks just above the median (Cardozo, 26). 2. Evolution of education indicators, health and households infrastructure Important decentralization and reorganization reforms of the 1991 constitution to the education system, contributed to an increase of 2 percent in the gross enrollment rate from 1994 to 24, explained mainly by higher 4 It is important to note that rural income is on average percent of urban income, a proportion that saw no major changes from 1996 to According to ocial statistics, unemployment in rural areas is much lower than in urban (7.1 percent versus 1.6 percent in 25) but as explained above, low incomes help maintain high poverty rates. 5

9 enrollment in public schools. Special improvements were observed in rural areas, particularly concerning secondary education which increased from,6 percent in 1996 to 55.1 in 24. In spite of this progress, secondary rural education is still low and far away from the net enrollment rate of 76,8 percent in urban areas (Sanchez, 22). In general, reforms helped reduce illiteracy, had positive eects on enrollment and increased the amount of teachers in the public sector. However, quality of education, and particularly of public education, continued to be very low and even weakened, showing dramatic dierences compared to private schools. 6 Studies referring to net enrollment rates by income quintiles, indicate that both the dierence of primary net enrollment between the lowest and highest quintiles as well as of secondary enrollment dropped. Although poorer households increased their demand for secondary public education, the gap in secondary net enrollment rates between the poorest and the richest quintiles remained above percentage points in 24. (Sanchez, 22). The eects of the crisis years on education, can be seen in lower growth rates of enrollment between 1998 and 1999 as well as in the higher demand of middle income households for public education. In these years net enrollment of the poorest quintiles decreased, specially in secondary education (Barrera et. al, 26). Instead of increasing coverage, inequality in access to tertiary education remained high, with only six percent of the years old students enrolled in 2 belonging to the rst quintile of income. New entrance to tertiary education was also hardly aected by the crisis, when the number of new entrants from 1997 to 1999 declined by 19 percent (World Bank 2) In the helth sector, indicators point at a major increase in coverage after reforms undertaken in 199. Coverage moved from 2 percent of the total 6 Rodríguez (25) argues that decentralization had an adverse eect on the quality of public education reected by an increasing gap of about 1.2 points annually between private and public school students in secondary ocial examinations. 6

10 population in 199 to 2 in 1995 and 75 in 24 (Sanchez, 26). Although the goal of Law 1 of 199 was to achieve universal health coverage by 2 as well as equality of access for all individuals, quality of services still depends on the purchasing power of the households. Dierentiation of health aliation into two types of regimes, the subsidized and the contributive, helped generate resources for incorporating poor households into the health system, specially in rural areas (GES, 25). Nevertheless inequality is still visible: only 48.1 percent of population in the rst quintile of income is covered and 8.7 percent of the fth is incorporated (Campos et al, 24). Beyond this, González (21) arms that public spending in health favors the richest forth and fth quintiles while other authors underline important regional dierences in access to health services favoring urban areas. Regarding infrastructure and basic services, analysis using non-income indicators like the Index of Unmet Basic Needs 7 and the Quality of Life Index 8 show signicant improvements in welfare from 1985 to 24. According to the former indicator, poverty decreased between 1985 and 2 when the percentage of population lacking one of these basic needs decreased by about 5 percent. Similarly, the living conditions of the Colombian population reected by the Quality of Life Index improved considerably since 1985, particularly in the non-urban areas. Although eects of the crisis are not directly visible in these indicators, which measure infrastructure conditions in the long run, they are evident in the stagnating pace of improvements in the second half of the nineties (PNUD, 25). In general, coverage and access to basic services, grew very slowly from 7 The Index of Unmet Basic Needs mainly captures infrastructure conditions at the urban level, including variables like inadequacy of housing conditions, homes without basic sanitary facilities, crowding, school absenteeism, and economic dependency. It assumes that people are poor if they lack one of their basic needs, and extremely poor if they lack two or more. 8 This index is a multidimensional indicator, ranking from to 1, with the latter representing the highest possible welfare. It captures in a single measure variables corresponding to quality of housing, access to public services, education as a measure of human capital, and the size and composition of the household. 7

11 1997 to 2, specially sewer and waste management services. According to Ramírez (26), access to electricity increased from 9,8 percent in 1997 to 95.7 percent in 2, but still with a lower coverage in rural zones (8,1 percent in 2). Access to water shows higher dierences, with 97,6 percent of urban households connected to public water supply in 2 and 5,5 in rural areas. Coverage of sewer service is 9 percent for urban households but only 16 for rural, while waste management services coverage is 74 percent for urban households and 16 percent for rural. Regional dierences are visible in sewer services and water supply, where the departments located in the Atlantic and Pacic regions (without Valle del Cauca) show the lowest rates. Dening and Measuring Pro-Poor Growth.1 Dening Pro-Poor Growth There are basically two ways of classifying studies on PPG: rst, according to their approach (weak or strong), and second, according to specic features of the measurement methodologies (complete/full or partial). For the weak (also called general) approach, any growth path leading to poverty reduction is considered pro-poor, so any positive income growth is dened as being propoor. In contrast, the strong (also called strict) approach considers growth to be pro-poor only when both poverty and inequality decrease. This approach is based on the identity that decomposes reductions in poverty into changes in mean income or growth eect, and changes in the distribution of income, called the distributional eect (Klasen 26; Kakwani, Khandker, and Son 24; Datt and Ravallion 1992). The strong approach to pro-poor growth can be further subdivided into (strong) relative or strong absolute. The relative approach focuses on proportional changes in income between poor and non-poor and considers growth pro-poor when relative inequality decreases. This is only possible if incomes of the poor rise by a higher proportion than incomes of the non-poor. For the strong absolute approach, 8

12 growth is pro-poor if absolute income gains of the poor are as much or more than those of the non-poor, meaning that absolute inequality (dened as the absolute dierence in income between the poor and non-poor) decreases. 9 One important critique of the relative approach is that in a recession, large income drops among the rich can give a pro-poor picture even if the poor are not gaining at all. Similarly, a pro-rich distributional shift during a period of overall economic expansion may result in large absolute gains for the poor without the growth path being considered pro-poor (Ravallion, 24). Furthermore, and as explained by Klasen (26) and shown with a simple nice numerical example, pro-poor growth in the strong absolute approach is almost impossible to achieve in practice, given that absolute income gains of the poor are usually much lower than those of the non-poor. However, as shown by Klasen (26), this concept makes good sense when analyzing the non-income dimension of poverty. 1.2 Measuring Pro-Poor Growth Using GIC When using the absolute approach mentioned above, the growth incidence curve (GIC) graphs the rate of growth of real income (or real expenditure) for each percentile of the distribution between two periods of time. A curve 9 An example given by Ravallion (24) illustrates the dierence between changes in relative and absolute inequality better. Consider only two households: a poor one with an income of USD 1, and a non-poor one with an income of USD 1, in the rst period. After an income increase of 1 percent for both households in the second period, the poor household earns USD 2, while the non-poor one earns USD 2,. In this case, the distance from each household to the mean remains unchanged and thus relative inequality does not change. According to the strict approach, growth would have been neither pro-poor nor anti-poor. But since the absolute dierence between the two households increases from USD 9, to USD 18,, absolute inequality rises sharply and growth can be considered anti-poor in the strict-absolute sense. 1 Pro-poor growth measures can be classied into partial or complete types. Partial measurement uses neither a concrete measure of poverty nor a poverty line, while complete measurement requires a poverty line to compare dierent growth paths and the degree to which they are pro-poor. The growth incidence curve developed by Ravallion and Chen (2) and the poverty growth curve proposed by Son (2) can both be categorized under the partial type of measurement. In contrast, the indices of McCulloch and Baulch (2), Kakwani and Pernia (2), and Ravallion and Chen (2) are categorized under the full approach because the growth processes are judged from a rate or an index of propoor growth that requires dening a poverty line (Kakwani, Khandker, and Son 24). 9

13 below zero (the x axis) at all points of the distribution indicates that all households suered income losses. The contrary indicates income gains for all percentiles and consequently a poverty decrease compared with the initial period. An upward-sloping curve indicates that rich households (the richer income percentiles) beneted more than others, while a downward-sloping curve indicates the poor beneted more, giving evidence of pro-poor growth in a relative sense (i.e., that relative inequality has fallen). The GIC is formally derived from the following equations: y t (p) = F 1 t (p) = L t (p) t with y t (p) > (1) GIC : g t (p) = y t y t(p) 1 (p) 1 (2) g t (p) = where p is the corresponding percentile, F 1 t L t (p) L t 1 (p) ( t + 1) 1 () is the inverse of the cumulative distribution function at the p th percentile (which gives the income of that percentile), L t (p) is the Lorenz curve (with slope L t (p)) and t = t 1 t 1 is the growth rate in the mean (GRIM) of income (or expenditure) per capita. The GIC can be dened as the growth rate in income of the p th percentile as shown in equation (2) or as shown in equation () after replacing (1) into (2). If all percentiles exhibit the same growth rate, then the Lorenz curve does not change, inequality remains unchanged and g t (p) = t in equation () for all p. Should the ratio between the growth rate of the p th percentile to the mean increase over time (i.e., yt(p) t > y t 1(p) t 1 ), then the growth rate of the p th percentile is higher than the mean growth rate, g t (p) > t. Following this, inequality falls if g t (p) is a decreasing function for all p (Ravallion 21). The graphical analysis of the GIC would not demand using a poverty line to determine whether growth was benecial to the poor. However this is only possible when the slope of the curve has a clear trend. First-order 1

14 dominance of the distribution at date t over t 1 exists when the GIC is above zero for all percentiles, a conclusion that cannot be easily drawn if the GIC switches sign. In practice, the GIC often has dierent slopes at dierent points and switches along percentiles, making it impossible to draw clear conclusions. Based on the GIC, Ravallion and Chen (2) proposed the rate of propoor growth (PPGR) as the area below the GIC up to the selected poverty line of the initial period. This area equals total income growth of the poor. The PPGR is equivalent to the ordinary rate of growth times a distributional correction given by the ratio of the actual change in poverty over time (using the Watts index) to the poverty change that would have been observed if growth had not aected the income distribution (Ravallion, 24). If the PPGR is higher than the growth rate in mean (GRIM), growth is pro-poor, while the opposite result indicates that distributional changes negatively aected the poor. Formally this is dened as follows: where P P GR = g p t = dw t dt Z = 1 Ht g t (p)dp (4) H t Z Ht W t = z log[ ]dp (5) y t (p) is the Watts poverty measure, z is the poverty line and H t is the headcount ratio H at time t. For the aim of this paper, we calculate initially GIC by using the third quarter of Colombian household surveys "Encuesta Nacional de Hogares" (ENH) between 199 and 2 representative at the national level for the rural/urban divide, and for ve regions including the capital district Bogotá. Starting in 21, the methodology changed to a continuous one "Encuesta Contínua de Hogares" (ECH), which means that the information is collected year-round by dividing the sample size on a weekly basis (for details on methodological changes, see Lasso 22). For calculating poverty with the 11

15 ECH, the usual procedure is to aggregate information corresponding to the third quarter. For simplicity, we will call ENH and ECH short as household surveys (HHS). This study uses the nal corrected income variable produced by the DNP, which is the basis for ocial calculations of poverty. 11 When calculating standard poverty and inequality indicators, all information is taken into account. To draw the GIC, the rst and last two percentiles of the distribution are not shown to facilitate reading the curves, given that high uctuations in these percentiles may still be due to problems in the data rather than to true changes in income. 12 Calculation of real income is done using the implicit deators of the poverty line in its 25 version (called M 25), which updates the poverty lines by using the consumer price index for low-income groups 1 (for methodological details, see DNP 26).. The Non-Income Growth Incidence Curve (NIGIC) Most of the literature on pro-poor growth (including the theoretical discussion in Chapter.1) has focused exclusively on the income dimension. This focus to the income dimension of poverty (which is, thus, focusing on MDG 1), neglects all the theoretical discussion to take a broader view of poverty which considers the non-income dimensions as well (World Bank 2; Sen 11 Non-response and underreporting are important problems in Colombian household surveys. The National Planning Department (DNP) applies three correction steps: In the rst, they estimate missing income using human capital models based on Mincerian equations, in the second, they adjust incomes to the national accounts to correct for underreporting, and in the third, they also rectify under-reported income of homeowners (DNP 26).After all corrections, there remain very low and high uctuating incomes in the rst percentiles of the distribution as well as about 1 to 1.5 percent of households with an income equal to zero. 12 The use and extension of Ravallion's methodology reveals possible data weaknesses. As is well known, income surveys have more severe problems than expenditure surveys due to the fact that people tend to omit more information or underreport it. Drawing GIC places high demands on the surveys in terms of data quality. Furthermore it may also be a very useful tool to see how prices aect poor and rich households dierently (Grimm and Günther 25). 1 Poverty lines are available for each of the 1 metropolitan areas, for the rest of the urban areas, and for rural Colombia. 12

16 1998). The spirit of the MDGs considers improvements in non-income dimensions of welfare such as education, health, service availability as equally important as income growth. As shown by Grosse, Klasen, and Harttgen (25, 26) and Klasen (26) for Bolivia, it is possible to extend the pro-poor growth measurement to non-income variables such as education or health, by drawing non-income growth incidence curves (NIGIC). Following these authors we present two kinds of NIGIC, unconditional and conditional. In the unconditional case, we sort the percentiles on the x-axis according to the respective non-income variable, e.g. from the health-poor to the health-rich, and plot the growth (or progress) of health for these percentiles on the y-axis. In the conditional NIGIC, we sort percentiles according to per capita income, 14 but also plot the growth of the non-income indicator of these percentiles on the y-axis to see how their non-income growth is related to their income position. Furthermore, we will also extend the GIC and NIGIC, which show relative changes in their standard use, to absolute changes in income and non-income dimensions of welfare as well. This allows us to measure pro-poor growth using the three denitions of weak absolute, relative, and strong absolute introduced above. The absolute GIC is dened as: GIC absolute : c t (p) = y t (p) y t 1 (p): (6) For the NIGIC, the same graphical analysis holds: If the relative NIGIC is downward sloping growth has been pro-poor in the relative sense, and if it is above growth has been pro-poor in the weak absolute sense. If the absolute NIGIC is downward sloping growth has been pro-poor growth in the strong absolute sense (and an absolute NIGIC above also shows weak absolute pro poor growth). Analogously to deriving the PPGR from the GIC, Grosse, Klasen, and Harttgen (25) derive the pro-poor change, PPCH (as the area 14 It would also be interesting to use consumption instead of income as the monetary measure to see whether the conditional results change. 1

17 under the absolute GIC up to the headcount): P P CH = c p t = 1 H t 1 XH t and compare it to the change in mean (CHIM) 1 c t (p) (7) CHIM = t = t t 1 : (8) to have a single measure to judge upon absolute pro-poor growth: If the PPCH exceeds the CHIM growth is pro-poor in the strong absolute sense. For the non income dimension just described, we use the Colombian quality of life surveys (QLS) of 1997 and 2, which although includes more chaptes than the HHS, such as consumption, quality of living, health, life satisfaction and subjective welfare. We use also income derived from this survey (the nal corrected income variable: monthly household per capita income) and consumption. The latter (monthly household per capita consumption) was recalculated using the original data sources due to some changes in the questionnaire design from 1997 to 2. The consumption data used here include weekly, monthly, and quarterly expenditures on food, clothes, leisure (such as alcohol, tobacco, newspapers, etc.) but it excludes yearly spending on consumer durables such as furniture or journeys. 15 Consumption could not be corrected for agricultural home production because this information is not available in 2. We apply the GIC methodoloy to both income and consumption to compare results between them and to double check for inconsistencies vs. HHS results. For drawing the NIGICs we generated ve types of indexes on education, asset ownership, housing quality, health and life satisfaction. Education is measured as the "potential years of education" (which is the inverse of the educational gap) for each individual which is as well calculated for the household as a whole (average of the household) given that it allows to 15 Both income and consumption and the poverty lines are deated with the implicit deator of the poverty lines, methodology

18 include children younger than 5 years. Any member being currently enrolled is treated as having full potential years of education (which is equal to having no educational gap). The maximum is set to 16 years. The variable is generated irrespectively of the age of the individual since the range of age varies a lot within each grade, especially so at higher levels of schooling. Any member being currently not enrolled is treated as having completed his schooling career. 16 We also calculate the variable averaged over the household and only for adult members aged between 15 and 65. This excludes all young children most of which have full 16 potential years of education, since primary enrollment is very high in Colombia, and it excludes all elderly which are more likely to never have attended school. For other non-income dimensions, we calculate four simple composite welfare indices where each component enters with the same weight. First, we calculate an index of asset ownership consisting of consumer durables (short: durables index), ranging from having to 16 assets., 17. As mentioned above each asset enters the index with the same weight, i.e. 1 for having it and for not having it. Such durables should reect the "permanent income or wealth position" of a household beyond uctuations in income. Second, we calculate a house and service quality index consisting of dwelling characteristics (oor material, number of rooms, etc.) and access to services such as electricity, sanitation, etc., 18 ranging from 1 (not because each house has at least one room) to 19. Many variables reect the service availability of public infrastructure and others also the wealth position. Third, we calculate an index of life satisfaction and subjective welfare 16 This variable was also calculated for men and women separately to detect gender biases in education. Results point to no education gender biases in Colombia. 17 The list of variables included is: fridge, juicer, TV (black and white and/or color), radio, motorbike, car, oven, microwave, electric/hot shower, vacuum cleaner, washing machine, cable tv connection, satellite tv connection, video, computer. 18 The list of variables included is: number of rooms (entering from 1 to 8 which is limited to be the maximum), telephone line, electricity, high quality cooking material, high quality wall material, high quality oor material, natural gas connection, high quality water supply, sanitation, garbage collection service, high quality toilet, room with shower. 15

19 which includes a larger set of dierent variables than the two indices above. It ranges from to 1 and increases for example with variables like: living better today than ve years ago, feeling secure in the neighborhood, not having been a victim of violence in the last year, etc. 19 Again, each variable enters with the same weight, an assumption which might be more problematic compared to equal weights of each asset in the durables index. Fourth, we calculate a health index for each member of the household separately. We include the subjective health status, 2 whether the person goes frequently to health checkups, and whether she is covered by any social or health security system. 21 It ranges from 1 to 9 (not because the subjective health status is at least "1"). In addition to the health index at the individual level, the health index is also averaged over the household. It would be interesting to look at all variables that form the indices separately. However, since nearly all of them are dummy variables, the unconditional NIGIC cannot be applied to them; the conditional, however, could be, as shown by Grosse, Harttgen, and Klasen (26). Furthermore, one could think of more sophisticated methods such as principal components or factor analyses to derive the indices such as the "deprivation index" of Klasen (2) for South Africa. Last, all indixes could be integrated into one single "welfare" index or into two indixes, one more "monetary" and one more "human". 19 The list of variables included is: "life" is good, "life" is better than 5 years ago, "life" is better than childhood at the parents house, household lives economically well, household consumption is above subjective poverty line, household has/fears no "problems" with, e.g., work, family, partner, law, etc., neighborhood is safe, no member was victim of violence in the last year. 2 One should keep in mind that normally only one person in the household responds all the questions, so the "subjective" health status is not judged by each member himself but by the respondent of the whole interview. 21 The list of variables included is: health status (increasing from 1 to 4), annual preventive health checkup, being chronically ill, having been seriously ill in last year, being covered by, and paying for, social security or health system 16

20 4 Results 4.1 Dierent Time Spans, Urban/Rural Divide, and Regional Divide The GIC are drawn for the whole period and whole Colombia of 1996 to 25 and separately from 1996 to 1999, 1999 to 22, and from 22 to 25 to capture the growth eects on poverty for dierent time periods and on the urban and rural areas separately, thus those of the economic slowdown, crisis, and recovery periods on both areas. Calculations are undertaken as well taking 1996 as base year (cumulatively) and comparing each of the following years with it to see how income growth evolved in the period as a whole. For these dierent time spans, we use the HHS. The results can be compared to those GIC from the QLS which cannot be calculated for exactly the same period but only from 1997 to 2. Figure 8 shows results for Colombia as a whole from 1996 to 25. As can be seen, income gains were very close to zero (but always above) for almost all percentiles of the distribution except the rst ve to ten, giving evidence of the poor overall performance. While in urban areas the GIC is rather at, around zero (almost all percentiles had very low income growth rates, the poorest 5 percentiles mainly negative ones, and the PPGR is negative), in rural areas the curve has a clear negative slope indicating that growth was pro-poor in the relative and the weak absolute sense. This is conrmed by the PPGR, which is higher than the GRIM (Table 1). For total Colombia, the PPGR was just above the GRIM, inuenced by results in rural areas. The behavior in rural areas would suggest a fall in inequality, conrmed by decreases in all inequality indicators except Atkinson e = 2. The Gini coecient moved from.51 to.47 and the Theil entropy measure from.54 to.47. Urban inequality decreased from.57 to.55. The poverty gap and inequality among the poor fell in both areas as well as at the national level. The GIC for the period show that the economic slowdown 17

21 aected the poor much more than the non-poor (Figure 9). The positive slopes of the all three GIC, which are below the zero axis for almost all percentiles of the distribution, conrms that the poorer the household, the larger the impact of the economic slowdown on income. In urban areas, the income drop was larger for the extreme poor, where extreme poverty increased to 16 percent and the poverty gap widened. In rural areas, the rise in poverty indicators was larger, with per capita income falling more than in urban areas (7.9 percent versus 2.5 percent). Extreme poverty (which is three times that in urban areas) jumped to 48 percent of the population and the poverty gap to 49 percent. Between 1999 and 22, all GIC have a negative slope and growth was pro-poor in the relative and the weak absolute sense, as also shown by the PPGR. This result should be interpreted cautiously as it is strongly inuenced by a statistical eect of income moving back to the levels observed before the economic slowdown and by large income increases for the poor in rural areas (Figure 1). If compared with 1999, poverty indicators improved until 22, for the nation as a whole, urban, and rural areas, but if compared with the base year 1996 they almost all worsened. It is important to underline that although incomes of the poor population in rural areas increased up to 6 percent, the incidence of poverty decreased only slightly, from 78 percent to 75.1 percent. The poverty gap, in contrast, showed a larger decrease from 49. percent to 9.4 percent, indicating a decrease in the severity of poverty. In urban Colombia, the poverty gap remained close to 2 percent while the incidence of poverty decreased slightly. In the period 22 to 25, growth was pro-poor in both urban and rural areas and consequently at the national level. As can be seen in Figure 11, for Colombia as a whole and for urban Colombia, incomes of the extreme poor increased more than for the rest of the population. In contrast, the inverted u-shaped form of the rural GIC shows the poorest ve percentiles 18

22 of the population and the richest ten as exhibiting the worst growth. It was in years 2 and 24 that favorable international commodity prices helped the rural sector achieve higher incomes and reduce poverty, a dynamic that did not hold on into 25. Rural areas seem to have followed their own dynamics, which, although not completely isolated from domestic growth conditions, had dierent roots. In contrast, economic recession had a more direct eect on urban households, which had been major recipients of real estate credit in the boom period between 1991 and When incomes fell and the bubble in the real estate market burst, they were more directly aected than rural households, which are more dependent on external price conditions, demand from urban areas, and climate conditions. The results are consistent with Figure 12, where the PPGR and the GRIM are calculated for each year against 1996 (i.e., cumulatively). The cumulative PPGR was negative up to 2 and below GRIM (which was also negative), except in 21. These results give evidence of income losses for the entire population but more than proportionally for the poor in comparison to the non-poor. Starting in 2, growth becomes pro-poor due to income gains in rural areas (Table 2). Urban data show that both the GRIM and the PPGR were negative in all years. Growth was anti-poor in all years if compared with 1996, and only in 2 did the situation stop worsening and income losses diminish. In 25, the urban GRIM and the PPGR were very close and just above zero, indicating that the population began beneting from growth. In rural Colombia, growth was anti-poor until 21 and the GRIM was negative up to 22 while growth was favorable to the poor starting in 21. In order to identify regional dierences in the pattern of pro-poor growth, we draw GIC for the ve regions covered by the household surveys: Atlántico, Central, Oriental, Pacíco and Bogotá 22 (Figure 1). Atlántico on the 22 Since the 1991 Constitution, Colombia is divided into 2 departments (nine more than under the old constitution) and a capital district (Bogotá). Furthermore, the country is 19

23 Caribbean Coast is home to about 21 percent of Colombian population and the largest coastal cities in the country: Barranquilla (1. million), Cartagena de Indias (1 million) and Santa Marta (44, million). This region comprises departments whose economies are primarily based on agriculture and cattle farming (like Sucre and Magdalena) as well as departments with signicant mineral resources in La Guajira and Cesar, which incidentally is also one of the poorest departments in the entire country. In this region, the GRIM from 1996 to 25 was negative (.67) but the poorest 2 percent of the population had income growth rates up to 6 percent and the PPGR was also close to zero (.6) (Table ). The Central region accounts for 25 percent of the population and is mixed in its production structure as well. It combines important coee-producing departments (Caldas, Risaralda, and Quindio) with large industrial areas in Antioquia and around its capital city Medellín. It also consists of the departments of Tolima, Huila, and Caquetá, located in the south, which base their income mainly on agriculture. In this region, the GIC from 1996 to 25 was almost at, around a mean growth rate of.62, pointing to stagnation in incomes for almost all percentiles of the population. The small gains in growth were distributed almost equally among the population. In the Oriental region, growth was more favorable to the poorest percentiles and pro-poor overall. This region includes important industrial manufacturing and commerce activities around Cundinamarca that have evolved due to the close proximity to the capital city. The region that performed worst was Pacíco, where growth was anti-poor. This region encompasses extreme contrasts: it includes the poorest department divided into ve geographical regions. The departments are divided into "Municipios" which according to the 1991 Constitution are the smallest territorial entities with governmental and administrative functions. The LSMS household survey groups the departments into four regions plus Bogotá and does not include eight of the new departments, namely: Arauca, Casanare, Vichada, Guaínia, Guaviare, Vaupés, Amazonas, and Putumayo which represent percent of Colombian population. The QLS, however, do include these departments and the results for them are shown in the following chapter. They should be treated with some caution especially because the number of observations is smaller than for the other regions. 2

24 in the country (Chocó) as well as one of the most developed (Valle). In Bogotá, the capital district with 15 percent of the Colombian population, the PPGR was just slightly above the GRIM and both rates were close to zero. The change in slope of the GIC indicates that the population closest to the poverty line showed the worst income growth Income versus Consumption Turning to the QLS data on Colombia, the GIC using income from 1997 to 2 are similar to the ones derived from LSMS data from 1996 to 25 when looking at the slope which is negative (indicating relative pro-poor growth). This holds for national, urban, rural, and all regions except Pacica (Figures 15a-h, all including the moderate and extreme headcount as vertical lines). Similarities decrease when looking at weak absolute pro-poor growth because some GIC turn negative using the QLS data in contrast to LSMS data, especially in rural areas. However, this might be simply due to the dierent time span, 24 and agricultural recovery only started in 2 and 24 as mentioned above. The rst more interesting result is that consumption data gives a completely dierent picture for nearly all areas and regions (Figures 16a and b). With the exception of the Central region, all GIC using consumption data are above and positively sloped, indicating weak absolute pro-poor growth but relative anti-poor growth. When looking at absolute increases and using income, the general picture shows anti-poor growth in the strong absolute sense (which is, as explained above, to be expected) but there are some exceptions from this pattern. The absolute income GIC for rural areas is negatively sloped, with only the rst decile exhibiting positive income changes and but the others having increasingly heavier income losses. 25 This 2 Results for dierent time periods for the ve regions are not shown here, but available upon request. 24 This is currently under investigation using the 1997 and 2 LSMS data. 25 This is a case one could call a "pro-poor recession" when the poor (in this case those 21