## Channing Arndt and Finn Tarp

Print publication date: 2016

Print ISBN-13: 9780198744801

Published to Oxford Scholarship Online: January 2017

DOI: 10.1093/acprof:oso/9780198744801.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2019. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see www.oxfordscholarship.com/page/privacy-policy).date: 22 March 2019

# Multidimensional Assessment of Child Welfare for Tanzania

Chapter:
(p.215) 14 Multidimensional Assessment of Child Welfare for Tanzania
Source:
Measuring Poverty and Wellbeing in Developing Countries
Publisher:
Oxford University Press
DOI:10.1093/acprof:oso/9780198744801.003.0014

# Abstract and Keywords

Identifying trends in living standards in Tanzania has been a subject of considerable interest. Analysis of a household budget survey conducted in 2007 revealed consumption poverty rates approximately similar to the rates calculated from a comparable survey conducted in 2001. This stagnation in consumption poverty occurred despite relatively high published rates of economic growth over the same period and little change in measured inequality. Price inflation over the same period as measured by the household budget survey also differed drastically from inflation rates derived from the published consumer price index (CPI) and the GDP deflator. The growth–poverty–inequality conundrum alongside the wide divergences in measured inflation provoked a great deal of analysis. More recently in 2015, the World Bank published a poverty assessment based on a household budget survey conducted in 2011/12 and found a reduction in consumption poverty of about six percentage points.

# 14.1 Introduction

Identifying trends in living standards in Tanzania has been a subject of considerable interest. Analysis of a household budget survey conducted in 2007 revealed consumption poverty rates approximately similar to the rates calculated from a comparable survey conducted in 2001 (Government of Tanzania 2009). This stagnation in consumption poverty occurred despite relatively high published rates of economic growth over the same period and little change in measured inequality. Price inflation over the same period as measured by the household budget survey also differed drastically from inflation rates derived from the published consumer price index (CPI) and the GDP deflator (Adam et al. 2012). The growth–poverty–inequality conundrum alongside the wide divergences in measured inflation provoked a great deal of analysis.1

More recently, the World Bank (2015) published a poverty assessment based on a household budget survey conducted in 2011/12. This recent assessment focused heavily on comparisons of the results from 2011/12 with the data available from the 2007 survey and found a reduction in consumption poverty of about six percentage points. In the companion volume to this book, Arndt et al. (2016a) draw upon this and other analyses to assess growth and poverty for Tanzania, and Arndt et al. (2017) conduct a macroeconomic assessment of the growth–poverty relationship using a structural model. (p.216) They find that the six percentage point reduction in poverty from 2007 to 2011/12 lies at the optimistic end of a reasonable range.

The assessment of consumption poverty trends in Tanzania over this most recent period (2007–11/12) has been substantially complicated by changes in the data collection methods employed in 2011/12 compared with all earlier surveys. In their poverty assessment, the World Bank (2015) also took the opportunity to apply a series of methodological changes to the computation of the nominal consumption aggregate and the poverty lines. These differentials rendered the analyses of the 2011/12 survey non-comparable with published analyses from 2007 and earlier. In order to account for these differences, the World Bank (2015) took a series of steps to revise 2007 data and calculations.

The revisions to the 2007 data were considerable. World Bank (2015: 2) reports that ‘consumption per adult rose by almost one-third’. The poverty line was also adjusted upward substantially, leaving the measured poverty rate at the national level essentially at the same value as reported in previously published assessments. Nevertheless, the issue of achieving comparability in data and methods clearly dominates any analysis of consumption poverty trends over the 2007 to 2011/12 period.

Rather than enter this fray, the work presented here seeks to analyse welfare trends from a multidimensional perspective, relying on data from four Demographic and Health Surveys conducted over the period 1991/2 to 2010. The chapter is structured as follows. Section 14.2 provides a brief review of multidimensional poverty measures. Both the first-order dominance (FOD) method and the Alkire–Foster (AF) approach are considered. Section 14.3 presents the datasets employed and the choices made to derive a set of comparable indicators. Section 14.4 presents results, including a comparison across the FOD and AF approaches. A final section concludes by highlighting the need for a collection of poverty tools to fully capture the complex nature of poverty dynamics.

# 14.2 Multidimensional Poverty Measurement

## 14.2.1 First-Order Dominance

The FOD methodology and implementation are described in detail in Chapters 3 and 4. They highlight that FOD analysis is an approach to comparing populations using multiple, binary welfare indicators without imposing weighting schemes or making assumptions about preferences for each indicator. Briefly, multidimensional welfare comparisons are based on the simple criterion that it is better to be not deprived than deprived in any indicator. FOD comparisons of population A and B result in one of three (p.217) outcomes: population A dominates population B; population B dominates population A; dominance is indeterminate. Indeterminate outcomes occur when two populations are too similar or too different for definitive comparisons to be made (without further information or assumptions). For example, when comparing two individuals using three binary indicators with outcomes (0,1,0) and (1,0,1), dominance cannot be established because we do not assume it is better to be not deprived in any given dimension. The same logic can be extended to populations.2

Dominant outcomes are binary and thus provide no information about the extent of domination. To mitigate this shortcoming, we draw bootstrap samples from the surveys considered and conduct FOD analysis for each sample.3 The share of dominant outcomes for each pair of populations across all bootstrap samples can be interpreted as a probability of domination. Thus, while the welfare indicators are ordinal in nature, the application of bootstrap sampling produces probabilities of one population performing better than another. Probability of net domination across all bootstraps is used to rank areas. The probability of net spatial domination of area i is defined as the sum of the probability that i dominates each other area minus the sum of the probability that each other area dominates i. This probability of net spatial domination can be linearly transformed into an index that falls in the interval [−1,1] where higher values indicate that an area is better off. Analogously, bootstrap samples can be employed to calculate temporal net domination of a given area in time period t relative to time period s.

## 14.2.2 Alkire–Foster Approach

Next, we consider an alternative approach to multidimensional analysis, the Alkire–Foster (AF) approach developed by Alkire and Foster (2007). The method is well known for its application to the Multidimensional Poverty Index of the United Nations Development Programme (UNDP) which assesses welfare in over one hundred countries (see for example, Alkire and Santos 2010). This section provides a brief overview of the methodology. Alkire et al. (2015) provide a recent and comprehensive discussion of an array of multidimensional poverty measures.

The AF approach to multidimensional analysis aggregates information obtained from a set of binary welfare indicators into a single index that captures both the incidence and intensity of multidimensional poverty. The process of defining this index can be described in two steps: identification and (p.218) aggregation. Identification is achieved in what Alkire and Foster (2007) refer to as a dual cut-off method. First, as with FOD, the approach begins with a set of binary welfare indicators, where in each dimension an individual is deemed to be deprived or not deprived according to a dimension-specific threshold. Second, an across dimension cut-off must be specified to distinguish the poor from the non-poor. In this context, the cut-off (k) identifies the poor as those with a weighted deprivation count greater than a cutoff level k. This provides a poverty headcount (H). When weights are equal across dimensions, k can be expressed as a number of dimensions such that individuals who are poor in k or more dimensions are considered poor.

Identification of the poor (via H) provides no information about the intensity of poverty. If an individual with a weighted deprivation count greater than k (i.e. one who is defined as poor in the multidimensional sense) becomes poor in an additional dimension, the multidimensional headcount ratio would not reflect this increase in the intensity of poverty. Therefore, an additional aspect of poverty is introduced to reflect the intensity of poverty. Intensity is measured by the average weighted deprivation count among those who are identified as poor. The final AF poverty index is referred to as the adjusted headcount ratio (M0) and is expressed as the product of the multidimensional headcount ratio (H) and the average deprivation count among the poor (A),

$Display mathematics$

Thus, a change in M0 cannot be understood without considering both H and A. Though the method is sensitive to thresholds within and across dimensions as well as dimensional weights, the adjusted headcount ratio is simple to compute and convenient for comparisons across time and space.

## 14.2.3 Comparison of the FOD and AF Approaches

Two important differences between the FOD and AF methodologies could lead to dissimilar results. First, FOD results use information from the full distribution of outcomes whereas M0 is the product of two averages: H and A. For FOD, indeterminacy may result between two populations B and C when B outperforms C for all but a small segment of population B. In the same situation, AF is likely to clearly establish that population B outperforms C. Second, the use of weights allows the AF method to result in clear outcomes that may be indeterminate with FOD. As noted, because no assumptions are made about the relative importance of each dimension, FOD dominance cannot be established between pairs of welfare outcomes such as (0,1,0) and (1,0,1). However, with the AF method, the comparison is dependent upon how weights are assigned. For instance, with equal weighting the second pair (p.219) is clearly superior to the first. On the other hand, with a weighting scheme (0.2, 0.6, 0.2) the first outcome is associated with greater welfare.

Results derived from FOD rely on few assumptions and strict criteria for establishing dominance. Thus, when dominance is established, the result is quite robust. AF, on the other hand, applies a weighting scheme and cut-off levels that may influence results. Despite this potential for different conclusions, in a comparison across thirty-eight countries Permanyer and Hussain (2015) find that the methodologies align closely with a correlation coefficient of 0.95. Arndt et al. (2016b) similarly find high correlations using census data for Mozambique. For these analyses, the indicators and thresholds determining deprived and not deprived in each dimension, which both FOD and AF are obliged to specify, were the same.

# 14.3 Data and Indicators

In this analysis, data from four Tanzania Demographic and Health Surveys (TDHS) are used to define five binary welfare indicators that allow multidimensional welfare to be estimated using both the FOD and AF methodologies in two subpopulations of children.

## 14.3.1 Demographic and Health Survey

The 1991/2, 1996, 2004/5, and 2010 TDHS provide the data used in this analysis (National Bureau of Statistics and Macro 1993, 1997, 2005, 2011). The TDHS aims to provide estimates for the entire country, for urban and rural areas, and regions. The 1991/2 and 1996 TDHS samples were drawn in a three-stage design, with the goal of selecting 500 households each in Dar es Salaam and Zanzibar, and 300 households in the remaining regions. Using the 1988 census sampling frame, 357 enumeration areas (EAs) were first selected from wards/branches and then within wards/branches such that rural and urban EAs were selected proportionally within each region. In the third sampling stage, households were selected from complete household listings in each EA. The sampling design for the 2004/5 and 2010 TDHS involved two stages where in the first stage 475 clusters were selected from a list of EAs based on the 2002 census with eighteen clusters selected in each region except Dar es Salaam where twenty-five clusters were selected. In the second stage, households were then systematically selected from complete household listings in each EA.

From this micro data, we capture the non-monetary multidimensional nature of poverty by defining two population groups: school-age children at least seven and less than eighteen years old and young children less than five years old. The 7–17 sample includes 13,608, 11,472, 14,357, and 14,687 (p.220) children and the under-five sample includes 7287, 6080, 7461, and 7526 children for 1991/2, 1996, 2004/5, and 2010, respectively. In each population group, children’s welfare is examined over time and across regions. Spatial areas include the nation, urban/rural areas, and geographical zones. Larger sample sizes for the 7–17 population group also allow analysis of administrative regions.4

## 14.3.2 Indicators

Table 14.1. Welfare indicators for children aged 7–17 and children aged 0–4

Population

Indicator

Deprivation threshold

Children aged 7 –17

Water

Water is not from a pipe, tap, or well.

Sanitation

Sanitation facility is not a toilet or ventilated improved pit (VIP) latrine.

Alternative sanitation

Sanitation facility is not a flush toilet or latrine of any kind.

Housing

Floors are made of dirt, sand, dung, or planks.

Education

The child has not completed at least primary school or is not in school.

Information

The household does not have a radio or television.

Children aged 0–4

Water

Water is not from a pipe, tap, or well.

Sanitation

Sanitation facility is not a flush toilet or VIP latrine.

Education

The child’s mother has not completed at least primary school.

Housing

Floors are made of dirt, sand, dung, or planks.

Nutrition

The child is more than two standard deviations below the median of the reference population in at least one of the following anthropometric measures: weight for age, height for age, or weight for height.

Delivery

The child was delivered at home.

Source: Author’s definitions

For each population group we identify a set of five binary welfare indicators based on the Bristol Indicators (Gordon et al. 2003); the indicators are presented in Table 14.1.

Ideally, the sanitation threshold would be specified such that children using unimproved sanitation (e.g. uncovered latrines or no facilities) would be considered deprived. However, in 1992, 1996, and 2004 the TDHS does not distinguish between covered and uncovered latrines. In 2010, 73 per cent of school-age children used latrines and of these children 89 per cent used uncovered latrines. It is logical then to classify all latrines to be a deprivation. In section 14.4, we examine the sensitivity to the sanitation indicator choice by considering an alternative sanitation threshold where the use of any kind of latrine is not deemed to be a deprivation.

Browsing household surveys, the possibilities of examining a rich variety of deprivations appear to be great. However, both the FOD and the AF methodologies require that all indicators be non-missing for every individual or household in the sample. Care must be taken in constructing indicators that apply to the full population being examined. For instance, immunization histories seem to provide a useful measure of the health of children under five. Yet, children under the age of one would not be fully immunized and therefore should not be deemed deprived based on incomplete immunization records. Consequently, the sample would need to be restricted to children aged one to five rather than zero to five.

Women’s health indicators present similar difficulties. The Demographic and Health Surveys offer information on a wide range of family planning, fertility, and maternal health topics. However, these questions tend to be posed to a narrow range of women for whom these issues apply, and thus care must be taken to restrict the sample to the relevant population. For instance, maternal health issues would limit the population to not only (p.221) women of childbearing age, but also women who were pregnant in the recent past. Depending on sample sizes or analytical goals, necessary restrictions may render the inclusion of certain indicators impractical due to the concomitant restrictions on the sample.

## 14.3.3 Descriptive Statistics

Figure 14.1 presents mean deprivation trends for children aged 7–17 at the national and urban/rural level. Table 14.2 also reports deprivation at the zonal level for all indicators including the alternative sanitation indicator. Overall, Figure 14.1 exhibits positive signs of advancement in most indicators. School-age children make considerable progress in access to education and information with national deprivation in education reduced by more than half between 1992 and 2010. Similar trends are observed in rural and urban areas and in all zones.

Access to safe water follows the most variable pattern. Urban water deprivation is relatively low but increases over time from 9 to 14 per cent. While national and rural areas achieve gains over the entire period, welfare backslides somewhat between 2004 and 2010 to 29 per cent and 33 per cent, respectively. In the zones, only Western makes progress between each survey while Central, Eastern, and Southern Highlands deteriorate over the eighteen-year period.

Figure 14.1. Children aged 7–17 deprived by welfare indicator (per cent)

Source: Authors’ calculations based on the 1991/2, 1996, 2004/5, 2010 TDHS (National Bureau of Statistics and Macro 1993, 1997, 2005, 2011)

Table 14.2. Children aged 7–17 deprived by welfare indicator (per cent)

Water

Sanitation

Alternative Sanitation

1992

1996

2004

2010

1992

1996

2004

2010

1992

1996

2004

2010

Nation

36

35

26

29

97

97

93

89

13

13

13

16

Rural

44

41

30

33

98

99

98

97

17

16

17

20

Urban

9

8

12

14

92

91

76

60

1

1

3

3

Central

27

27

34

35

97

96

97

95

8

10

9

17

Eastern

14

16

15

22

96

94

84

78

3

5

2

2

Lake

50

36

41

28

97

100

93

87

20

19

18

21

Northern

42

50

24

39

97

97

92

91

13

21

12

18

S. Highlands

30

37

23

32

98

99

95

93

7

4

9

14

Southern

43

36

26

34

98

99

96

91

5

3

4

8

Western

43

41

23

22

96

96

98

93

17

12

23

24

Zanzibar

8

5

1

1

97

96

85

73

57

45

32

25

Source: Authors’ calculations based on the 1991/2, 1996, 2004/5, 2010 TDHS (National Bureau of Statistics and Macro 1993, 1997, 2005, 2011)

Housing

Education

Information

1992

1996

2004

2010

1992

1996

2004

2010

1992

1996

2004

2010

Nation

81

81

75

71

40

40

25

17

62

56

38

36

Rural

91

91

88

83

42

43

28

20

70

61

44

41

Urban

47

35

31

24

34

28

14

7

37

29

20

19

Central

87

85

84

87

44

41

29

23

70

59

49

52

Eastern

62

56

46

46

41

36

16

11

48

41

28

24

Lake

88

91

81

74

44

42

24

20

66

59

35

36

Northern

72

80

70

64

31

39

18

11

51

52

37

39

S. Highlands

87

88

83

79

42

45

25

14

75

63

43

47

Southern

86

83

76

69

36

40

29

14

67

64

46

34

Western

89

86

90

84

42

39

33

24

68

58

41

35

Zanzibar

66

63

44

34

44

34

24

14

44

33

18

25

Source: Authors’ calculations based on the 1991/2, 1996, 2004/5, 2010 TDHS (National Bureau of Statistics and Macro 1993, 1997, 2005, 2011)

(p.222) Urban areas progressed in terms of the housing and the primary sanitation indicator. Though access to urban sanitation improved by thirty-two percentage points over the study period, deprivation remained high at 60 per cent. In contrast, rural areas achieved little gains in either indicator with deprivations in sanitation and housing of 97 per cent and 83 per cent in 2010. Within the zones, Zanzibar, and Eastern zone follow urban patterns while the remaining zones generally mirror rural areas.

The vast majority of the population uses covered or uncovered pit latrines (83 per cent in 1992 and 73 per cent in 2010). The primary sanitation indicator classifies children using any pit latrines as deprived while the alternative sanitation indicator shifts this large percentage of children to being not deprived. As a result, deprivation in the alternative sanitation indicator (children with no sanitation facility) is extremely low. In contrast to the primary sanitation indicator, the percentage of children deprived in the alternative indicator increased at the national, rural, and urban areas, with more substantial increases in Central, Southern Highlands, and Western zones. Zanzibar is the only area to significantly reduce alternative sanitation deprivation.

Table 14.3. Children 0–4 deprived by welfare indicator (per cent)

Water

Sanitation

Housing

Education

Nutrition

1992

1996

2004

2010

1992

1996

2004

2010

1992

1996

2004

2010

1992

1996

2004

2010

1992

1996

2004

2010

Nation

35.89

35.92

27.89

29.53

97.45

97.52

95.84

91.53

82.97

81.67

80.70

75.64

52.89

44.71

42.11

40.51

54.41

54.56

47.41

41.21

Rural

43.08

42.08

30.50

32.63

98.89

98.44

98.87

98.05

91.49

91.05

91.11

86.58

57.15

48.63

46.45

44.89

55.79

56.98

49.69

43.50

Urban

8.56

7.39

16.34

16.37

92.00

93.24

82.46

63.77

50.63

38.24

34.62

29.06

36.68

26.59

22.88

21.88

49.17

43.33

37.33

31.46

Central

27.79

27.08

34.68

41.02

98.28

97.82

98.07

97.18

90.48

85.65

91.18

89.54

48.68

41.46

43.64

42.24

58.75

54.13

50.70

50.90

Eastern

19.23

15.01

18.44

20.21

96.82

94.18

91.54

81.90

62.72

53.72

50.81

45.57

46.29

36.98

32.86

32.87

54.31

54.83

36.04

34.04

Lake

46.35

37.44

44.61

28.35

98.13

98.82

95.18

91.09

90.35

91.02

85.12

79.63

60.52

50.07

40.22

42.42

48.72

49.58

43.74

36.76

Northern

41.65

48.04

25.58

40.73

94.48

97.19

94.21

92.39

73.58

80.26

76.70

70.74

39.33

37.98

34.31

33.98

52.83

55.14

46.06

43.64

S. Highlands

38.37

35.77

25.64

34.44

98.79

98.66

97.47

90.74

86.71

82.27

82.37

71.71

49.44

42.64

47.66

34.12

59.19

63.66

53.48

47.48

Southern

26.56

35.89

24.50

29.00

99.29

99.15

97.92

94.11

86.95

86.72

84.71

82.32

52.32

39.75

42.34

34.42

65.85

65.57

58.95

44.77

Western

42.91

45.32

20.91

23.64

97.55

97.06

98.28

95.39

90.71

89.02

90.85

88.35

66.86

54.26

47.44

51.71

49.28

47.16

49.27

38.18

Zanzibar

11.13

3.25

1.32

1.72

97.25

94.95

82.89

73.58

68.63

64.32

42.99

35.21

57.16

58.48

51.35

42.75

60.89

49.85

35.91

39.11

Source: Authors’ calculations based on the 1991/2, 1996, 2004/5, 2010 TDHS (National Bureau of Statistics and Macro 1993, 1997, 2005, 2011)

Table 14.3 presents mean deprivation levels for children under five. Deprivations in water, sanitation, and housing closely follow the levels and trends seen with school-age children. Deprivation in education for under-fives measures whether children’s mothers have completed primary school. Though (p.223) (p.224) (p.225) declining in every year, under-five education deprivation is greater than that of school-age children. Under-five nutrition, as evidenced by anthropometric measures, improved over the eighteen-year period. However, these figures remained high with 31 per cent and 41 per cent of urban and rural children nutritionally deprived in 2010. Though improvement occurred in all zones, as many as 51 per cent of children in Central were still nutritionally deprived in 2010.

# 14.4 FOD Results

## 14.4.1 Temporal FOD Comparisons

We begin by examining whether child welfare, as defined by our set of five indicators, improved between 1992 and 2010. FOD temporal analysis compares the performance of a given area between survey years and is reported as the average probability of net domination across 100 bootstrap iterations. Net probability of domination measures the probability that the welfare of an area improves between two years minus any probability of regression.

Table 14.4. Temporal net FOD comparisons, children 7–17 years (probabilities)

1996 FOD 1992

2004 FOD 1992

2010 FOD 1992

2004 FOD 1996

2010 FOD 1996

2010 FOD 2004

Static

Boot

Static

Boot

Static

Boot

Static

Boot

Static

Boot

Static

Boot

National

0.03

1

1.00

1

0.98

1

0.97

1

0.97

0.11

Rural

0.04

1

0.51

1

0.97

1

0.53

1

0.90

0.13

Urban

1

0.23

0.28

0.19

0.17

0.09

0.07

Central

0.13

0.09

0.09

0.07

0.03

0.02

Eastern

0.18

0.54

0.17

1

0.47

0.20

0.09

Lake

0.00

1

0.67

1

0.99

0.24

1

0.81

0.15

Northern

−0.21

1

0.57

1

0.51

1

0.86

1

0.88

−0.01

S. Highlands

0.03

1

0.66

1

0.82

1

0.61

1

0.67

0.09

Southern

0.03

1

0.49

0.33

1

0.66

1

0.72

0.03

Western

0.19

0.11

1

0.62

0.09

1

0.53

1

0.34

Zanzibar

0.17

1

0.92

1

0.92

1

0.82

1

0.81

0.00

Source: Authors’ calculations based on the 1991/2, 1996, 2004/5, 2010 TDHS (National Bureau of Statistics and Macro 1993, 1997, 2005, 2011)

Table 14.4 reports the temporal FOD outcomes for school-age children. Both the static results and bootstrap probabilities provide strong evidence of welfare progress at the national level and in rural areas from 1992 or 1996 to 2004 or 2010. In contrast, urban areas advance between 1992 and 1996 and then stagnate in the remaining years. National and rural stagnation between 1992 and 1996 is consistent with very little to no change in the percentage of children who are deprived in sanitation, housing, and education. Urban stagnation across most years and national and rural stagnation between 2004 and 2010 are directly associated with decreasing welfare in the water indicator. Among the zones, only the Central zone shows little to no signs of advancement during the study period. In line with substantial improvements in all indicators, Zanzibar exhibits the greatest probability of advancement among the zones.

Table 14.5. Temporal net FOD comparisons with the alternative sanitation indicator, children 7–17 years (probabilities)

1996 FOD 1992

2004 FOD 1992

2010 FOD 1992

2004 FOD 1996

2010 FOD 1996

2010 FOD 2004

Static

Boot

Static

Boot

Static

Boot

Static

Boot

Static

Boot

Static

Boot

National

0.05

0.39

0.05

0.39

0.02

0.00

Rural

0.06

1

0.42

0.08

0.27

0.00

0.00

Urban

0.18

0.03

0.00

0.00

0.00

0.01

Central

0.01

0.14

0.00

0.12

0.01

−0.08

Eastern

0.01

0.23

0.09

1

0.41

0.16

−0.01

Lake

0.06

1

0.35

0.32

0.01

0.06

0.00

Northern

−1

−0.27

0.40

0.16

1

0.77

1

0.63

0.00

S. Highlands

0.06

0.35

0.18

0.07

0.05

0.01

Southern

1

0.38

0.03

0.09

0.00

0.00

Western

0.24

0.04

0.13

0.00

0.00

0.02

Zanzibar

0.34

1

0.94

1

0.95

1

0.79

1

0.88

0.00

Source: Authors’ calculations based on the 1991/2, 1996, 2004/5, 2010 TDHS (National Bureau of Statistics and Macro 1993, 1997, 2005, 2011)

To evaluate the sensitivity of temporal outcomes to the sanitation threshold, FOD comparisons were re-estimated using the alternative sanitation indicator and reported in Table 14.5. Consistent with alternative sanitation deprivation increasing over time (Table 14.2), evidence of temporal advancement is drastically reduced. Notably, static advancement in national and urban areas disappears and only moderate bootstrap probabilities of welfare gains in national and rural areas remain between 1992 or 1994 and 2004. However, Zanzibar, where the alternative sanitation indicator improved in all years, exhibits strong probabilities of advancement.

Table 14.6. Temporal net FOD comparisons, children 0–4 years (probabilities)

1996 FOD 1992

2004 FOD 1992

2010 FOD 1992

2004 FOD 1996

2010 FOD 1996

2010 FOD 2004

Static

Boot

Static

Boot

Static

Boot

Static

Boot

Static

Boot

Static

Boot

National

0.06

1

0.69

1

0.97

0.35

1

0.89

0.16

Rural

0.04

1

0.23

1

0.88

0.02

0.40

0.17

Urban

0.07

0.05

0.03

0.00

0.01

0.03

Central

0.05

0.05

0.04

−0.05

−0.04

0.00

Eastern

0.30

1

0.37

0.39

0.12

0.15

0.06

Lake

0.04

1

0.26

1

0.90

0.02

1

0.75

0.28

Northern

−0.03

0.19

0.23

1

0.55

1

0.52

−0.02

S. Highlands

0.03

0.18

1

0.62

0.05

1

0.39

0.06

Southern

0.01

0.22

0.28

0.13

1

0.61

0.10

Western

0.10

0.05

1

0.46

0.01

1

0.27

0.04

Zanzibar

0.10

1

0.77

1

0.94

1

0.62

1

0.71

0.00

Source: Authors’ calculations based on the 1991/2, 1996, 2004/5, 2010 TDHS (National Bureau of Statistics and Macro 1993, 1997, 2005, 2011)

Finally, the temporal FOD results for children under five are reported in Table 14.6. Though the indicator trends for school-age children are generally (p.226) (p.227) (p.228) comparable to under-fives deprived in water, sanitation, housing, and education, the under-five temporal results demonstrate the strict nature of the FOD criteria. For example, children under five and school-age children advance in all indicators between 1996 and 2004 nationally and between 1996 and 2010 nationally and in rural areas. However, unlike outcomes for school-age children, in the under-five static case, 2004 does not dominate 1996 for the nation or rural areas. In both periods, the probability of domination is lower than that of school-age children. This example demonstrates that the FOD criteria demand progress not only on average, but throughout the distribution (see Chapter 11 for more a more detailed discussion of indeterminacy).

## 14.4.2 Spatial FOD Comparisons

In each year, FOD comparisons are made between all areas to determine the degree of domination of each area and zone. Values in the inner table represent the probability that the row area dominates the corresponding column area.5 Row averages measure the probability that the row population dominates all other populations, and column averages measure the probability that the column population is dominated by all other populations. In interpreting a population’s relative wellbeing, both row and column averages should be considered.

Table 14.7. 1992 Bootstrap spatial FOD comparisons, children 7–17 years (probabilities)

Area

National

Rural

Urban

C

E

L

N

SH

S

W

Z

Avg.

National

1

0.03

0.36

0.05

0.02

0.19

0.17

Rural

0.02

0.00

Urban

1

1

0.94

0.69

0.97

0.28

0.78

1

0.96

0.11

0.77

Central

0.01

0.10

0.07

0.02

0.07

0.08

0.04

Eastern

0.26

0.53

0.39

0.59

0.09

0.50

0.31

0.27

Lake

0.01

0.01

0.00

Northern

0.12

0.49

0.04

0.36

0.40

0.08

0.25

0.17

S. Highlands

0.10

0.03

0.02

0.01

0.04

0.02

Southern

0.01

0.00

Western

0.11

0.04

0.07

0.02

0.02

Zanzibar

0.17

0.29

0.27

0.07

0.39

0.01

0.11

0.35

0.22

0.19

Average

0.16

0.36

0.00

0.18

0.08

0.28

0.03

0.15

0.20

0.21

0.01

0.15

Note: Figures in bold indicate FOD in the static sample.

Source: Authors’ calculations based on the 1991/2 TDHS (National Bureau of Statistics and Macro 1993)

Table 14.8. 2010 Bootstrap spatial FOD comparisons, children 7–17 years (probabilities)

Area

National

Rural

Urban

C

E

L

N

SH

S

W

Z

Avg.

National

1

0.87

0.03

0.01

0.08

0.20

Rural

0.09

0.01

Urban

1

1

1

0.59

1

1

0.99

0.99

0.91

0.85

Central

0.00

Eastern

0.92

0.97

0.95

0.72

0.49

0.68

0.76

0.35

0.58

Lake

0.02

0.42

0.59

0.05

0.11

Northern

0.04

0.27

0.01

0.08

0.04

S. Highlands

0.02

0.43

0.54

0.02

0.05

0.17

0.12

Southern

0.02

0.34

0.04

Western

0.17

0.02

Zanzibar

0.94

1

0.98

0.02

0.94

0.11

0.39

0.43

0.99

0.58

Average

0.29

0.49

0.00

0.58

0.06

0.27

0.17

0.21

0.26

0.23

0.00

0.23

Note: Figures in bold indicate FOD in the static sample.

Source: Authors’ calculations based on the 2010 TDHS (National Bureau of Statistics and Macro 2011)

The 1992 and 2010 spatial comparisons for school-age children are presented in Tables 14.7 and 14.8.6 Within the tables, all domination in the static case (bold values) and significant bootstrap probabilities occur when urban areas, Eastern, Northern (1992), and Zanzibar dominate or when rural areas, Lake (1992), and Central (2010) are dominated. Column averages indicate that Southern Highlands, Southern, and Western zones also have moderate probabilities of being dominated in both years. Between the remaining areas, FOD is indeterminate or the probabilities of domination are quite low. Column averages for urban, Eastern, and Zanzibar and row averages for rural areas and Central increase considerably between 1992 and 2010, indicating a greater disparity between the welfare of the better-off and worst-off areas. In both years, the nation is nearly as likely to dominate other areas as it is to be dominated.

Table 14.9. 1992 Bootstrap spatial FOD comparisons, children 0–4 years (probabilities)

Area

National

Rural

Urban

C

E

L

N

SH

S

W

Z

Avg.

National

0.94

0.01

0.02

0.08

0.11

Rural

0.00

Urban

0.94

0.94

0.97

0.81

0.53

0.60

0.97

1

0.60

0.50

0.79

Central

0.14

0.04

0.04

0.02

0.02

0.03

Eastern

0.34

0.59

0.13

0.15

0.41

0.33

0.10

0.04

0.21

Lake

0.02

0.01

0.00

Northern

0.07

0.42

0.02

0.07

0.21

0.02

0.07

0.09

S. Highlands

0.02

0.01

0.00

Southern

0.00

Western

0.03

0.00

Zanzibar

0.01

0.04

0.03

0.02

0.17

0.03

Average

0.14

0.31

0.00

0.12

0.08

0.08

0.06

0.17

0.16

0.08

0.05

0.11

Note: Figures in bold indicate FOD in the static sample.

Source: Authors’ calculations based on the 1991/2 TDHS (National Bureau of Statistics and Macro 1993)

Table 14.10. 2010 Bootstrap spatial FOD comparisons, children 0–4 years (probabilities)

Area

National

Rural

Urban

C

E

L

N

SH

S

W

Z

Avg.

National

1

0.62

0.01

0.16

Rural

0.03

0.00

Urban

1

1

1

0.47

0.75

0.98

1.00

0.91

0.79

0.79

Central

0.00

Eastern

0.76

0.88

0.90

0.24

0.50

0.43

0.35

0.37

0.44

Lake

0.01

0.63

0.41

0.01

0.03

0.08

0.12

Northern

0.01

0.34

0.03

0.04

S. Highlands

0.02

0.41

0.01

0.04

Southern

0.18

0.57

0.02

0.01

0.08

Western

0.02

0.00

Zanzibar

0.29

0.70

0.53

0.01

0.19

0.08

0.04

0.06

0.40

0.23

Average

0.21

0.44

0.00

0.48

0.05

0.12

0.16

0.15

0.14

0.16

0.00

0.17

Note: Figures in bold indicate FOD in the static sample.

Source: Authors’ calculations based on the 2010 TDHS (National Bureau of Statistics and Macro 2011)

Tables 14.9 and 14.10 present the spatial results for children under five in 1992 and 2010. In 1992, significant domination occurs only when urban areas and Eastern dominate or rural areas are dominated. The remaining areas are (p.229) (p.230) (p.231) essentially indeterminate with very low probabilities of domination. In 2010, the number of instances of static domination increases and domination now also occurs when Zanzibar dominates and when Central is dominated. Eastern and Zanzibar’s row averages significantly increase between 1992 and 2010, indicating an increasingly greater welfare compared to all other areas. The probability that rural areas and Central are dominated, as indicated by column averages, also increases, suggesting that these areas are falling behind all other areas.

## 14.4.3 Spatial FOD Rankings

Net domination scores measure the average probability across all bootstrap samples that an area dominates all other areas less the probability that it is dominated by all other areas. Net domination can be interpreted as the probability of domination, and allows areas to be ranked. Zonal rankings based on school-age children are reported in Table 14.11 (for zones) and Table 14.12 (for regions).7 It is worth noting that the difference in net domination scores is often insufficiently large to distinguish between differences in welfare outcomes and variability introduced through random bootstrapping. To avoid misinterpreting rankings within the tables, shading identifies clusters with similar net domination scores. Within these clusters, ranks cannot be established with confidence.

Table 14.11. Spatial FOD ranking and probability of net domination by zone and year, children 7–17

1992

1996

2004

2010

Rank Change

Domination

Rank

Domination

Rank

Domination

Rank

Domination

Rank

Eastern

0.26

1

Eastern

0.57

1

Eastern

0.73

1

Eastern

0.56

1

0

Zanzibar

0.20

2

Zanzibar

0.55

2

Zanzibar

0.54

2

Zanzibar

0.55

2

0

Northern

0.16

3

Central

−0.04

3

Northern

0.17

3

Northern

−0.04

3

0

S. Highlands

−0.08

4

Western

−0.10

4

Southern

−0.16

4

S. Highlands

−0.04

4

0

Central

−0.08

5

Northern

−0.12

5

Lake

−0.20

5

Lake

−0.15

5

−3

Western

−0.11

6

S. Highlands

−0.24

6

S. Highlands

−0.26

6

Southern

−0.16

6

−1

Southern

−0.14

7

Lake

−0.26

7

Western

−0.34

7

Western

−0.17

7

1

Lake

−0.21

8

Southern

−0.35

8

Central

−0.49

8

Central

−0.55

8

3

Note: Rankings within shaded groups are highly sensitive to small perturbations and should be interpreted with caution.

Source: Authors’ calculations based on the 1991/2, 1996, 2004/5, 2010 TDHS (National Bureau of Statistics and Macro 1993, 1997, 2005, 2011)

Across all four years, Zanzibar and Eastern outperform all areas with the probability of domination more than doubling between 1992 and the remaining years (Table 14.11). Though a number of zones seem to change rank from year to year, these changes are not robust due to small differences in the probabilities of domination. For example, Lake appears to improve from last to fifth, but given probabilities in 2010, a rank of fifth and seventh cannot be distinguished with confidence. However, the decline in Central province is robust. Not only was Central ranked last in 2010, but it has a probability of being dominated 0.38 greater than the seventh ranked zone, Western. The gap between the best-performing and worst-performing zones widened considerably from a range spanning [−0.21, 0.26] in 1992 to [−0.55, 0.56] in 2010.

Table 14.12. Spatial FOD ranking and probability of net domination by region and year, children 7–17

1992

2010

Rank Change

Domination

Rank

Domination

Rank

Dar es Salaam

0.64

1

Zanzibar (Urban)

0.74

1

−1

Zanzibar (Urban)

0.60

2

Dar es Salaam

0.58

2

1

Kilimanjaro

0.20

3

Zanzibar (Rural)

0.28

3

−1

Zanzibar (Rural)

0.09

4

Kilimanjaro

0.25

4

1

Tanga

0.02

5

Pemba

0.15

5

−4

Mbeya

−0.01

6

Coast

0.15

6

−4

Rukwa

−0.02

7

Mbeya

0.09

7

1

Tabora

−0.02

8

Mwanza

0.04

8

−3

Pemba

−0.04

9

Morogoro

0.00

9

−9

Coast

−0.04

10

Iringa

0.00

10

−9

Mwanza

−0.05

11

Ruvuma

−0.03

11

−3

Singida

−0.06

12

Mara

−0.03

12

−9

Arusha & Manyara

−0.07

13

Shinyanga

−0.07

13

−3

Ruvuma

−0.07

14

Tabora

−0.09

14

6

Lindi

−0.08

15

Tanga

−0.13

15

10

Shinyanga

−0.10

16

Arusha & Manyara

−0.15

16

3

Kgoma

−0.12

17

Kgoma

−0.21

17

0

Morogoro

−0.12

18

Lindi

−0.21

18

3

Iringa

−0.13

19

Mtwara

−0.21

19

−1

Mtwara

−0.15

20

Singida

−0.22

20

8

Mara

−0.16

21

Rukwa

−0.25

21

14

Dodoma

−0.16

22

Kagera

−0.28

22

−1

Kagera

−0.16

23

Dodoma

−0.41

23

1

Note: Rankings within shaded groups are highly sensitive to small perturbations and should be interpreted with caution.

Source: Authors’ calculations based on the 1991/2, 2010 TDHS (National Bureau of Statistics and Macro 1993, 2011)

Table 14.12 reports regional rankings in 1992 and 2010. In both years Zanzibar urban, Dar es Salaam, Kilimanjaro, and Zanzibar rural are the highest-ranked regions, with Zanzibar urban and Dar es Salaam decisively first and second. Consistent with strong temporal advancement, Zanzibar urban’s net domination widens in 2010. In 1992, the remaining nineteen (p.232) (p.233) regions have net domination scores falling in a narrow range between 0.02 and −0.16. Though many of the rank shifts between 1992 and 2010 rely on small differences in net domination scores, a few regions stand out. Pemba and Coast improve four places to ranks of fifth and sixth. Mororno, Mara, and Iringa all climb nine positions. Shidiga, Tanga, and Rukwu fall eight, ten, and fourteen places. Finally, Dodoma is decisively last in 2010.

## 14.4.4 Alkire–Foster

The AF approach provides, as noted, an alternative method for evaluating multidimensional poverty using the same set of binary indicators. In this analysis, a child is identified as multidimensionally poor when deprived in two or more equally weighted indicators. Recall that M0 = HA, and thus the adjusted headcount ratio reflects the proportion of children who are multidimensionally poor (H) multiplied by the average intensity of deprivation among poor children (A).

Table 14.13. Multidimensional poverty in two dimensions

Child Population

1992

1996

2004

2010

change

% change

7–17

Nation

M0

0.61

0.60

0.49

0.45

−0.16

−26.2

H

0.89

0.88

0.82

0.77

−0.12

−13.5

A

0.69

0.68

0.60

0.59

−0.10

−14.7

Urban

M0

0.38

0.31

0.24

0.17

−0.21

−54.2

H

0.65

0.57

0.47

0.33

−0.32

−49.0

A

0.58

0.54

0.51

0.52

−0.06

−10.3

Rural

M0

0.68

0.66

0.57

0.53

−0.15

−22.3

H

0.96

0.95

0.93

0.89

−0.07

−7.2

A

0.71

0.69

0.61

0.59

−0.12

−16.3

0–4

Nation

M0

0.63

0.61

0.57

0.54

−0.10

−15.3

H

0.92

0.91

0.89

0.85

−0.07

−7.5

A

0.69

0.67

0.64

0.63

−0.06

−8.4

Urban

M0

0.43

0.36

0.33

0.27

−0.17

−38.7

H

0.75

0.67

0.61

0.48

−0.26

−35.3

A

0.58

0.54

0.54

0.55

−0.03

−5.2

Rural

M0

0.69

0.67

0.63

0.60

−0.09

−12.5

H

0.97

0.96

0.96

0.94

−0.03

−2.9

A

0.71

0.69

0.65

0.64

−0.07

−9.9

Source: Authors’ calculations based on the 1991/2, 1996, 2004/5, 2010 TDHS (National Bureau of Statistics and Macro 1993, 1997, 2005, 2011)

(p.234) Table 14.13 reports M0 and its components, H and A, for school-age children and children under five who are deprived in two or more dimensions. Nationally, the adjusted headcount ratio for school-age children has declined over the eighteen-year study period from 0.61 in 1992 to 0.45 in 2010. The proportion of school-age children who are multidimensionally poor fell twelve percentage points to 77 per cent and the intensity of poverty fell ten percentage points to 59 per cent. Thus, the decline in M0 can be attributed roughly equally to incidence and intensity.

Rural areas experienced a similar reduction in the adjusted headcount ratio for schoolage children, which fell from 0.68 in 1992 to 0.52 in 2010. However, rural gains were driven primarily by a reduction in the intensity of poverty, which dropped twelve percentage points compared to only a seven-point decline in the headcount ratio. The proportion of schoolage children suffering two or more deprivations remained extremely high at 89 per cent in 2010. In contrast, the large reduction in the urban index from 0.38 to 0.17 was primarily due to a reduction in the poverty headcount, which at 33 per cent in 2010 was nearly cut in half over the study period. Moreover, the intensity of urban poverty declined by only six percentage points.

At the national, urban, and rural levels across years, a similar pattern occurs in children under five. However, all three measures, M0, H, and A, are higher and decline less compared to outcomes for school-age children. This disparity in gains between the two populations of children is consistent with FOD temporal results.

Figure 14.2. Relative contributions to the adjusted headcount ratio, M0, for children aged 7–17 by year

Source: Authors’ calculations based on the 1991/2, 1996, 2004/5, 2010 TDHS (National Bureau of Statistics and Macro 1993, 1997, 2005, 2011)

Figure 14.3. 2010 relative contributions to the adjusted headcount ratio, M0, for children aged 7–17, by area

Source: Authors’ calculations based on 2010 TDHS (National Bureau of Statistics and Macro 2011)

(p.235) Figures 14.2 and 14.3 explore each indicator’s relative contribution to the school-age adjusted headcount ratios over time and by aggregate areas, respectively. The most notable aspect of these figures is how stable the contribution of each indicator is over time and space. Nonetheless there are several subtle observations to be made. Between 1992 and 2010, the relative (p.236) contribution of education and information to poverty declined while the impact of sanitation increased (Figure 14.2). Across all three areas, sanitation and housing are the biggest contributors to poverty (Figure 14.3). Sanitation and information have a relatively greater impact on urban poverty while housing has a relatively greater influence on rural poverty.

## 14.4.5 Comparisons

Table 14.14. Multidimensional poverty in two dimensions by zone and region, children 7–17 years

1992

2010

Change

M0

H

A

Rank

M0

H

A

Rank

M0

H

A

Rank

Zanzibar

0.48

0.79

0.61

1

0.23

0.45

0.51

1

0.25

0.34

0.10

0

Eastern

0.48

0.77

0.63

2

0.30

0.55

0.55

2

0.18

0.22

0.08

0

Northern

0.56

0.86

0.65

3

0.45

0.75

0.60

3

0.11

0.11

0.06

0

S. Highlands

0.65

0.91

0.71

5

0.45

0.78

0.58

4

0.19

0.13

0.13

1

Lake

0.68

0.95

0.71

8

0.47

0.81

0.58

5

0.21

0.15

0.14

3

Western

0.67

0.94

0.71

7

0.50

0.86

0.58

6

0.17

0.07

0.13

1

Southern

0.65

0.94

0.69

6

0.51

0.85

0.60

7

0.15

0.09

0.10

−1

Central

0.63

0.91

0.70

4

0.57

0.89

0.64

8

0.07

0.01

0.06

−4

Zanzibar (Urban)

0.29

0.54

0.54

2

0.06

0.13

0.46

1

0.23

0.41

0.09

1

Dar es Salaam

0.26

0.54

0.48

1

0.11

0.23

0.48

2

0.15

0.30

0.01

−1

Zanzibar (Rural)

0.46

0.80

0.57

3

0.28

0.56

0.50

3

0.18

0.24

0.07

0

Pemba

0.57

0.88

0.65

5

0.35

0.67

0.52

4

0.22

0.21

0.13

1

Kilimanjaro

0.50

0.87

0.58

4

0.36

0.68

0.53

5

0.14

0.19

0.04

−1

Coast

0.61

0.94

0.65

9

0.39

0.76

0.52

6

0.22

0.18

0.13

3

Mwanza

0.64

0.93

0.69

15

0.42

0.75

0.57

7

0.22

0.18

0.12

8

Iringa

0.69

0.93

0.74

20

0.43

0.73

0.59

8

0.26

0.20

0.15

12

Mbeya

0.60

0.87

0.69

8

0.43

0.79

0.55

9

0.17

0.09

0.14

−1

Morogoro

0.64

0.92

0.70

16

0.44

0.76

0.58

10

0.20

0.16

0.11

6

Ruvuma

0.63

0.93

0.67

12

0.46

0.81

0.57

11

0.16

0.12

0.10

1

Mara

0.71

0.98

0.73

23

0.47

0.82

0.57

12

0.24

0.17

0.15

11

Tanga

0.62

0.91

0.69

10

0.48

0.76

0.63

13

0.15

0.15

0.05

−3

Shinyanga

0.68

0.92

0.74

18

0.48

0.83

0.58

14

0.20

0.09

0.17

4

Arusha & Manyara

0.57

0.82

0.70

6

0.48

0.79

0.61

15

0.09

0.03

0.10

−9

Tabora

0.59

0.92

0.64

7

0.49

0.90

0.55

16

0.09

0.02

0.09

−9

Rukwa

0.63

0.92

0.68

13

0.53

0.85

0.62

17

0.10

0.07

0.06

−4

Mtwara

0.70

0.97

0.72

21

0.53

0.85

0.62

18

0.17

0.12

0.10

3

Kagera

0.70

0.96

0.73

22

0.53

0.89

0.60

19

0.17

0.07

0.13

3

Singida

0.65

0.91

0.71

17

0.54

0.89

0.61

20

0.11

0.03

0.10

−3

Kgoma

0.68

0.97

0.70

19

0.55

0.89

0.61

21

0.14

0.08

0.09

−2

Lindi

0.63

0.93

0.68

14

0.57

0.96

0.59

22

0.06

−0.03

0.09

−8

Dodoma

0.62

0.90

0.69

11

0.58

0.90

0.65

23

0.04

0.01

0.04

−12

Source: Authors’ calculations based on the 1991/2, 1996, 2004/5, 2010 TDHS (National Bureau of Statistics and Macro 1993, 1997, 2005, 2011)

Zonal and regional values of M0, H, and A and the associated rankings for school-age children are reported in Table 14.14. As was seen in the FOD rankings (Tables 14.11 and 14.12), large groups of zones and regions are grouped in relatively tight ranges of M0. For example, in 1992 regions ranked five through twenty-three had M0 values falling in the range 0.57 to 0.71. Despite the very different approaches to comparing areas, FOD and AF produce similar spatial rankings. Zonal rankings based on the adjusted headcount ratio are nearly identical in 2010 to rankings based on net domination scores (Table 14.11). The notable exception is that Central is ranked last over the entire period based on the AF methodology, but declined over time with FOD (Table 14.11).

AF and FOD regional ranks are also remarkably similar, especially given the tight range of net domination scores and adjusted headcount ratios. In 2010, the top six regions have nearly the same rankings (Zanzibar (urban), Dar es Salaam, Zanzibar (rural), Pemba, Kilimanjaro, and Coast). The remaining regions follow a similar pattern, with Dodoma ranked last in both approaches. While the dynamics between 1992 and 2010 diverge between the approaches, some similarities remain, such as the widening gap between Zanzibar and Eastern zones and Zanzibar urban and Dar es Salaam regions—a gap most likely driven by greatly improved water quality in Zanzibar compared to other areas.

Table 14.15. Correlation between FOD spatial domination score and M0

1992

1996

2004

2010

Children 7–17

Zone

0.97

0.99

0.96

0.97

Regions

0.96

0.98

0.98

0.98

Children 0–4

Zone

0.86

0.81

0.82

0.86

Source: Authors’ calculations based on the 1991/2, 1996, 2004/5, 2010 TDHS (National Bureau of Statistics and Macro 1993, 1997, 2005, 2011)

Table 14.15 reports the correlations between M0 and a transformed FOD net domination index by year and by level of aggregation.8 Spatial correlations across regions/zones for the population of school-age children are strikingly high and range between 0.96 and 0.99. This result is consistent with correlations reported in Permanyer and Hussain (2015) and Arndt et al. (2017). The correlations are somewhat lower in the population of children under five falling in the range 0.81 to 0.86.

Despite some similarities, FOD and AF also generate numerous dissimilar temporal outcomes, in contrast to the spatial analyses. The AF method indicates welfare improved in every year for the nation, urban areas, and rural (p.237) (p.238) areas in both populations of children (Table 14.13). The per cent declines in the urban index was more than double that of rural areas. Welfare gains indicated by M0 were driven by both reduced poverty headcounts, H, and reduced intensity, A. Though, in both populations of children, FOD indicates national and rural welfare are likely to have improved over the entire period, advancement between individual years is less conclusive, particularly in the under five years of age sample (Table 14.6). In contrast to AF outcomes, FOD provides evidence of urban advancement only in the school-age population between 1992 and 1996 (Table 14.4) and not at all in the under-five population.

Why the big temporal difference? The FOD criteria are strict and require advancement throughout the distribution of welfare states. Regression in a subset of the population may lead to indeterminate results. On the other hand, advancement using the AF method is based on average headcount and intensity values. If a subset of the population fails to advance, M0 may still indicate the population as a whole is advancing. As noted in the discussion of FOD temporal results, temporal stagnation is likely to be associated with periods of regression in the water indicator and stagnation in the sanitation, housing, and education indicators. Given the equal weights applied in the AF method, the periodic lack of advancement in these indicators was offset by gains elsewhere, allowing advancement in the adjusted headcount measure.

# 14.5 Conclusion

Poverty analysis in Tanzania highlights the need for careful consideration of multiple welfare measures. With uncertainty surrounding consumption poverty estimations, multidimensional welfare analyses provide useful opportunities to supplement and cross-check these estimations.

In this chapter, we considered the FOD and AF approaches to multidimensional welfare analysis. In the Tanzanian context, the use of several methods shines a light on the limitations of any one approach to fully capture the complicated interactions of the many factors determining welfare.

(p.239) The FOD and AF approaches provide similar stories across areas and most notably the large urban rural disparities that have increased between 1992 and 2010. The two methodologies result in remarkably similar rankings of zones and regions. These rankings suggest a widening gap between the best- and worst-performing areas and indicate that the majority of areas lie in a tight range in the middle.

In contrast, despite employing the same set of welfare indicators, the approaches do not provide a clear and simple story of welfare dynamics. AF outcomes reflect the overall trend of indicator advancement with great improvements in the adjusted headcount index across all years, particularly in urban areas and for the school-age population. FOD, however, suggests periods of advancement and stagnation.

The national-level and rural areas appear to achieve robust welfare gains; however, these results are sensitive to the population of children considered as well as how the sanitation indicator is defined. FOD outcomes also highlight the failure of several indicators to improve, particularly, urban water, which deteriorated, and rural sanitation, which stagnated (or deteriorated if considering the alternative indicator). As a result of deterioration in urban water access, urban areas exhibit few signs of advancement. Furthermore, FOD provides no evidence of advancement between 2004 and 2010.

These results contrast with the adjusted headcount index of AF and consumption poverty figures, which indicate the greatest gains occur in urban areas and, in the case of consumption poverty, the greatest poverty reduction occurs between 2007 and 2011. Nonetheless, rather than conflict, the two multidimensional approaches complement one another by highlighting different aspects of poverty dynamics. While AF focuses on population averages, FOD identifies advancement or regression found throughout the population. In a sense, the approaches provide upper (AF) and lower (FOD) bounds on welfare advancement in Tanzania over the eighteen-year period.

References

Bibliography references:

Adam, C., D. Kwimbere, W. Mbowe, and S. O’Connell (2012). ‘Food Prices and Inflation in Tanzania’, International Growth Centre Working Paper, London and Oxford, July.

Alkire, S. and J. Foster (2007). ‘Counting and Multidimensional Poverty Measurement’, OPHI Working Paper 7. Oxford: University of Oxford.

Alkire S., J. Foster, S. Seth, M. Santos, J. Roche, and P. Ballon (2015). Multidimensional Poverty Measurement and Analysis. Oxford: Oxford University Press.

Alkire, S. and M. Santos (2010). ‘Acute Multidimensional Poverty: A New Index for Developing Countries’, OPHI Working Paper 38. Oxford: University of Oxford.

(p.240) Arndt, C., L. Demery, A. McKay, and F. Tarp (2016a). ‘Growth and Poverty Reduction in Tanzania’, in C. Arndt, A. McKay, and F. Tarp (eds), Growth and Poverty in Sub-Saharan Africa. Oxford: Oxford University Press, 238–62.

Arndt, C., M. A. Hussain, V. Salvucci, F. Tarp, and L. P. Østerdal (2016b). ‘Poverty Mapping Based on First-Order Dominance with an Example from Mozambique’, Journal of International Development, 28: 3–21.

Arndt, C., V. Leyaro, K. Mahrt, and F. Tarp (2017). ‘Growth and Poverty: A Pragmatic Assessment and Future Prospects’, in C. Adam, P. Collier, and B. Ndulu (eds), Tanzania: Policies for Prosperity. Oxford: Oxford University Press.

Atkinson, A. B. and M. A. Lugo (2010). ‘Growth, Poverty and Distribution in Tanzania’, International Growth Centre Working Paper 10/0831, London and Oxford, November.

Demombynes, G. and J. G. Hoogeveen (2007). ‘Growth, Inequality and Simulated Poverty Paths for Tanzania, 1992–2002’, Journal of African Economies, 16: 596–628.

Gordon, D., S. Nandy, C. Pantazis, S. Pemberton, and P. Townsend (2003). Child Poverty in the Developing World. Bristol: Policy Press.

Government of Tanzania (2009). Household Budget Survey 2007—Tanzania Mainland. Dar es Salaam: National Bureau of Statistics.

Hoogeveen, J. and R. Ruhinduka (2009). ‘Poverty Reduction in Tanzania since 2001: Good Intentions, Few Results’. Paper prepared for the Research and Analysis Working Group.

Kessy, F., O. Mashindano, A. Shepherd, and L. Scott (eds) (2013). Translating Growth into Poverty Reduction: Beyond the Numbers. Dar es Salaam: Mkuki na Nyota.

Mashindano, O., K. Kayunze, L. da Corta, and F. Maro (2011). Agriculture Growth and Poverty Reduction in Tanzania 2000–2010: Where Has Agriculture Worked for the Poor and What Can We Learn from this? Chronic Poverty Research Centre, Working Paper No. 208, June.

Mkenda, A., E. Luvanda, and R. Ruhinduka (2010). ‘Growth and Distribution in Tanzania: Recent Experience and Lessons’, Interim Report to REPOA.

National Bureau of Statistics (NBS) and ICF Macro (2011). Tanzania Demographic and Health Survey 2010. Dar es Salaam: NBS and ICF Macro.

National Bureau of Statistics (NBS) and Macro International Inc. (1993). Tanzania Demographic and Health Survey 1991/1992. Columbia, MD: NBS and Macro International.

National Bureau of Statistics (NBS) and Macro International Inc. (1997). Tanzania Demographic and Health Survey 1996. Calverton, MD: NBS and Macro International.

National Bureau of Statistics (NBS) and ORC Macro (2005). Tanzania Demographic and Health Survey 2004-05. Dar es Salaam: NBS and ORC Macro International.

Permanyer, I. and M. A. Hussain (2015). ‘Multidimensional Poverty Indices and First Order Dominance Techniques: An Empirical Comparison of Different Approaches’, EQUALITAS Working Paper 35. Spain: EQUALITAS.

Osberg, L. and A. Bandara (2012). ‘Why Poverty Remains High in Tanzania: And What to Do About It?’, Special Paper 12/3. Dar es Salaam: REPOA.

World Bank (2007). Tanzania: Sustaining and Sharing Economic Growth (CEM and Poverty Assessment). Washington, DC: World Bank (1 March).

(p.241) World Bank (2012). Spreading the Wings: From Growth to Shared Prosperity (PREM Issue 2). Washington, DC: World Bank.

World Bank (2013). Tanzania Economic Update: Raising the Game (PREM Issue 4). Washington, DC: World Bank.

World Bank (2015). Tanzania Poverty Assessment. Washington, DC: World Bank.

## Notes:

(1) Examples include Atkinson and Lugo (2010); Demombynes and Hoogeveen (2007); Hoogeveen and Ruhinduka (2009); Kessy et al. (2013); Mashindano et al. (2011); Mkenda et al. (2010); Osberg and Bandara (2012); and World Bank (2007, 2012, 2013).

(2) See Chapter 11 for a discussion of indeterminate outcomes.

(3) Bootstrap sampling follows the same stratified cluster sample design used in the DHS sampling. Samples are drawn with replacement.

(4) The region of Manyara was created from Arusha in 2002. To maintain consistency throughout the survey, these regions are combined. To achieve minimum sample sizes, Pemba North and Pemba South are combined and Zanzibar North and Zanzibar South are combined into Zanzibar rural.

(5) Note that bootstrap sampling introduces a degree of randomness into the results and care must be taken in interpreting very small probabilities or small differences in probabilities.

(6) For both populations of children, spatial tables generally follow the trend seen between 1992 and 2010 and are therefore not presented.

(7) Zonal rankings for children under five are not presented. The results are similar to rankings for school-aged children but have a larger number of areas with net dominations scores too similar to distinguish with confidence.

(8) In order to facilitate comparisons with M0, the net domination score was transformed to a range of [0,1] such that low values are associated with higher welfare rates.