Dear esteemed colleagues,
Jomo kindly brought to my attention the interesting discussion that followed his post on the release of the latest SOFI assessment, and in particular the comments (well.. the harsh criticisms) by professor Thomas Pogge.
Let me start my stating upfront that I fully share what seems to be everyone's major concern in this particular discussion: namely, that data and information on the extent and distribution of malnutrition, and in particular of severe under nutirition, are as reliable and detailed as possible so that they can be effectively used to guide intervention. I share it to the point that I have made it since 2011 the main objective of my professional activity.
I have always believed that criticisms are essential to advance knowledge, and participated in debates from both ends: making and receiving even harsh criticisms. But I also believe that - to be constructive - criticisms should be well informed.
By reading the intervention that professor Pogge makes in response to Jomo's post, however, I cannot avoid noticing that there is still an important, continuing confusion on the meaning and the role that the threshold level used to compute FAO estimates of the Prevalence of Undernourishment (PoU) plays.
I thought I had addressed extensively in the past in private converssations with professor Pogge, but I realize now that I must have been unable to fully convey it and therefore I feel obliged to try again.
In short, professor Pogge seems to continue believing that FAO numbers imply a severe underestimation of the extent of undernourishment because of the threshold level used to compute the estimates. He seems to be (wrongly) interpreting the use of a single, national value of the Minimum Dietary Energy Requirement (MDER) value of around 1800 kcal/day as a cut-off point that separates those who are undernourished from those who are not, as if it was equivalent to the statement that, if an adult is consuming 1800 kcals or more, then he or she must be adequately nourished. Then, because the 1800 kcal value is obtained by making reference to levels of physical activity that correspond to a sedentary lifestyles, he has an easy shot at raising the example of the rickshaw puller to exemplify what he thinks to be a fundamental drawback of the FAO method.
The essence of the confusion, I think, stems from mistakenly interpreting the FAO method as if it would equally apply to an individual, rather than to a population, and by failing to recognize the possibility to control for the various aspects that determine the energy requirements of indivduals.
If we could afford the luxury of being able to assess sufficiently well the principal charactersitics that determine dietary energy requirements (age, gender, body mass, physical activity level, health status, metabolic efficiency and a few environmental aspects such as temperature and oxygen
concentration) we could take a large enough sample of individuals, to be considered representative of the population we are interested in, determine what each of them need, compare it to what they eat, and determine whether who is eating enough (a point made also by Bill Gates in a recent issue of Development Asia avilable here https://www.scribd.com/doc/182243503/Beyond-the-MDGs-What-will-the-Global-Development-Agenda-look-like-after-2015).
But the point is that, unless we are interested in assessing the individual conditions of all members in a population, we do not need to do it. Economy of information when possible, I believe, is a value. As we should not waste food, we should also not waste resources to collect information that is not necessary for the purpose at hand. That is what the (proper) use of statistical inference affords us to do.
I am sure professor Pogge is aware of this, and indeed he is not suggesting that we use the "headcount" approach skey=tched above as the regular basis for conducting the annual, worldwide SOFI assessment. Most appropriately, he suggests that we do it on an experimental basis, taking a few examples, and check whether or not the statistical shortcut leads us to the proper destination.
The good news is: we have done it. In the few cases where the details of the information contained in nationally representative sample surveys of food consumption allowed us to do it (thanks to the generosity of those who funded those extremely expensive surveys) we have verified that the method pioneered by P.V. Sukathme in 1960 and perfected by Loganaden Naiken in the early 1990's is indeed the best way to do a sufficiently reliable analysis with the kind of information we have for most countries.
What seems strange to many commentators (the use of the MDER) is simply the way in which we control for the relevant information that we do not have at individual level in the food consumption surveys, and that we derive from information on the demographics of the popualtion. That's all.
Of course, there can be ways to improve and refine our estimate of the MDER, when information is available not only on the gender age pyramid, but also on the distribution of body masses and physical activity levels exist, but the point remains that it is the MINIMUM of the range of requirements in the group that should be used. Otherwise valuable researchers have fallen in the past in the error of using the AVERAGE energy requirement in a population as a threshold to assess food inadequacy, only to be disappointed to find values stubbornly close to 50%, which they have mistakenly interpreted as the failure of intervention to generate progress over time. If they would have applied the same approach to affluent populations, they would have found the same, an error pointed out by Sukhatme himself back in 1978 (see P. V. Sukhatme, Economic and Political Weekly, Vol. 13, No. 31/33, Special Number (Aug., 1978) where he reminded readers that it would be similar to an:"assessment carried out in Great Britain by the late Sir Arthur Bowley. He found that 50 per cent of the population in UK were below the average need for UK and he concluded that some 50 per cent of the population must be undernourished. I need hardly add that notwithstanding the eminence of Sir Arthur Bowley, the conclusion was rejected by the Government of UK. This was 40 years ago when the concept of requirement had hardly developed to a point to grasp its full implications. But to adopt the same method today that was rejected as inapplicable in the UK decades ago, is to ignore the knowledge we have gained in understanding the concept of physiological requirement."
I hope that we can continue discuss the way in which we can jointly contribute to improve the quality and reliability of the many assessments that are being conducted, but I also sincerely hope that we could focus more on real issues and not continuing to question the theoretical and empirical validity of the FAO method for annual assessment and national level.
As Jomo said, there are needs to improve the assessment, and we are doing it, starting by recognizing that food insecurity is more than just the sufficiency of dietary energy intake. We hope that the academic and policy communities will continue to pay close attention to what we do, asking questions when things are not sufficiently transparent and clear, but stop implying that we are all a bunch of hypocrits ready to sell our dignity and professional deontolgy to the need to please politicians. That, professor Pogge, is personally offensive, and I hope you understand why I felt the need to intervene in this very valuable debate.
Carlo
Carlo Cafiero, PhD
Project Manager of “Voices of the Hungry”
Statistics Division (ESS)
Food and Agriculture Organization of the United Nations (FAO)
Viale delle Terme di Caracalla, 00153 Rome, Italy
Tel: +39 06 570 53895
Skype: carlocafiero
http://www.fao.org/economic/ess/ess-fs/voices/en/
Jomo kindly brought to my attention the interesting discussion that followed his post on the release of the latest SOFI assessment, and in particular the comments (well.. the harsh criticisms) by professor Thomas Pogge.
Let me start my stating upfront that I fully share what seems to be everyone's major concern in this particular discussion: namely, that data and information on the extent and distribution of malnutrition, and in particular of severe under nutirition, are as reliable and detailed as possible so that they can be effectively used to guide intervention. I share it to the point that I have made it since 2011 the main objective of my professional activity.
I have always believed that criticisms are essential to advance knowledge, and participated in debates from both ends: making and receiving even harsh criticisms. But I also believe that - to be constructive - criticisms should be well informed.
By reading the intervention that professor Pogge makes in response to Jomo's post, however, I cannot avoid noticing that there is still an important, continuing confusion on the meaning and the role that the threshold level used to compute FAO estimates of the Prevalence of Undernourishment (PoU) plays.
I thought I had addressed extensively in the past in private converssations with professor Pogge, but I realize now that I must have been unable to fully convey it and therefore I feel obliged to try again.
In short, professor Pogge seems to continue believing that FAO numbers imply a severe underestimation of the extent of undernourishment because of the threshold level used to compute the estimates. He seems to be (wrongly) interpreting the use of a single, national value of the Minimum Dietary Energy Requirement (MDER) value of around 1800 kcal/day as a cut-off point that separates those who are undernourished from those who are not, as if it was equivalent to the statement that, if an adult is consuming 1800 kcals or more, then he or she must be adequately nourished. Then, because the 1800 kcal value is obtained by making reference to levels of physical activity that correspond to a sedentary lifestyles, he has an easy shot at raising the example of the rickshaw puller to exemplify what he thinks to be a fundamental drawback of the FAO method.
The essence of the confusion, I think, stems from mistakenly interpreting the FAO method as if it would equally apply to an individual, rather than to a population, and by failing to recognize the possibility to control for the various aspects that determine the energy requirements of indivduals.
If we could afford the luxury of being able to assess sufficiently well the principal charactersitics that determine dietary energy requirements (age, gender, body mass, physical activity level, health status, metabolic efficiency and a few environmental aspects such as temperature and oxygen
concentration) we could take a large enough sample of individuals, to be considered representative of the population we are interested in, determine what each of them need, compare it to what they eat, and determine whether who is eating enough (a point made also by Bill Gates in a recent issue of Development Asia avilable here https://www.scribd.com/doc/182243503/Beyond-the-MDGs-What-will-the-Global-Development-Agenda-look-like-after-2015).
But the point is that, unless we are interested in assessing the individual conditions of all members in a population, we do not need to do it. Economy of information when possible, I believe, is a value. As we should not waste food, we should also not waste resources to collect information that is not necessary for the purpose at hand. That is what the (proper) use of statistical inference affords us to do.
I am sure professor Pogge is aware of this, and indeed he is not suggesting that we use the "headcount" approach skey=tched above as the regular basis for conducting the annual, worldwide SOFI assessment. Most appropriately, he suggests that we do it on an experimental basis, taking a few examples, and check whether or not the statistical shortcut leads us to the proper destination.
The good news is: we have done it. In the few cases where the details of the information contained in nationally representative sample surveys of food consumption allowed us to do it (thanks to the generosity of those who funded those extremely expensive surveys) we have verified that the method pioneered by P.V. Sukathme in 1960 and perfected by Loganaden Naiken in the early 1990's is indeed the best way to do a sufficiently reliable analysis with the kind of information we have for most countries.
What seems strange to many commentators (the use of the MDER) is simply the way in which we control for the relevant information that we do not have at individual level in the food consumption surveys, and that we derive from information on the demographics of the popualtion. That's all.
Of course, there can be ways to improve and refine our estimate of the MDER, when information is available not only on the gender age pyramid, but also on the distribution of body masses and physical activity levels exist, but the point remains that it is the MINIMUM of the range of requirements in the group that should be used. Otherwise valuable researchers have fallen in the past in the error of using the AVERAGE energy requirement in a population as a threshold to assess food inadequacy, only to be disappointed to find values stubbornly close to 50%, which they have mistakenly interpreted as the failure of intervention to generate progress over time. If they would have applied the same approach to affluent populations, they would have found the same, an error pointed out by Sukhatme himself back in 1978 (see P. V. Sukhatme, Economic and Political Weekly, Vol. 13, No. 31/33, Special Number (Aug., 1978) where he reminded readers that it would be similar to an:"assessment carried out in Great Britain by the late Sir Arthur Bowley. He found that 50 per cent of the population in UK were below the average need for UK and he concluded that some 50 per cent of the population must be undernourished. I need hardly add that notwithstanding the eminence of Sir Arthur Bowley, the conclusion was rejected by the Government of UK. This was 40 years ago when the concept of requirement had hardly developed to a point to grasp its full implications. But to adopt the same method today that was rejected as inapplicable in the UK decades ago, is to ignore the knowledge we have gained in understanding the concept of physiological requirement."
I hope that we can continue discuss the way in which we can jointly contribute to improve the quality and reliability of the many assessments that are being conducted, but I also sincerely hope that we could focus more on real issues and not continuing to question the theoretical and empirical validity of the FAO method for annual assessment and national level.
As Jomo said, there are needs to improve the assessment, and we are doing it, starting by recognizing that food insecurity is more than just the sufficiency of dietary energy intake. We hope that the academic and policy communities will continue to pay close attention to what we do, asking questions when things are not sufficiently transparent and clear, but stop implying that we are all a bunch of hypocrits ready to sell our dignity and professional deontolgy to the need to please politicians. That, professor Pogge, is personally offensive, and I hope you understand why I felt the need to intervene in this very valuable debate.
Carlo
Carlo Cafiero, PhD
Project Manager of “Voices of the Hungry”
Statistics Division (ESS)
Food and Agriculture Organization of the United Nations (FAO)
Viale delle Terme di Caracalla, 00153 Rome, Italy
Tel: +39 06 570 53895
Skype: carlocafiero
http://www.fao.org/economic/ess/ess-fs/voices/en/
