Dear Thomas
Thank you for your detailed criticisms. As you know, some of the issues you raise have been addressed before, especially in SOFI 2012, the first SOFI issued after I arrived here. You are quite right that the decision to improve the methodology was made before my arrival, but SOFI 2012 explicitly showed the implications for hunger estimates of different assumptions about calorie requirements. For some time, FAO (wrongly in my view) focused on food availability, ignoring access. The new estimates sought to give more attention to access as well as other factors previously ignored or wrongly assumed. We have also considerably improved the database, with more information from consumption surveys. While it is generally a bad idea to change criteria in the middle of monitoring exercise, the new PoU estimates are better than the old ones, in my humble opinion, although I will be the first to concede that limitations and problems remain.
You suggest that FAO should not have changed the methodology despite all the criticisms of the PoU, including from inside. The way in which the original methodology, developed by the Indian statistician Sukhatme during the 1960s, had been applied in the following years was felt to be inappropriate in several regards, and we showed the implications of the major changes made to the methodology and assumptions in the technical appendix. My colleagues believed that it was FAO’s responsibility to improve the methodology given the responsibility FAO has to report on the MDG1c indicator.
Using an explicitly low cut-off – related to the minimum dietary energy needs for a sedentary lifestyle – allows us to credibly improve the reliability of estimates. As dietary energy requirements vary considerably in any population, and we know very little about such variations in different contexts, we can much more reliably estimate the proportion of people who do not even get this minimal dietary requirement. We have made the database public, to allow others to make estimates using different benchmarks and methodologies. While we have been criticized for changing the methodology and assumptions, very little use has been made of the data available to come up with alternative estimates.
You correctly point out that the hunger trends are not consistent, but as you also correctly point out, the price spikes were in (early) 2008 and 2012 (one can add 2010-11 as well). My colleagues have long conceded that their provisional estimates for 2009 were wrong due to wrong assumptions regarding likely income contractions following the 2008 financial crisis and higher food price levels in many countries, overly influenced by the early pessimistic views on the extent and depth of the financial crisis of 2009 in developing countries, and the earlier cereal price spikes. What we now know is that developing countries continued to experience sustained growth in 2009 and 2010, and also that the impacts of such high international producer prices were mitigated in many countries by food export restrictions, food subsidies, etc., explaining the non-correspondence between international producer and domestic consumer food prices.
This is why we should be concerned about the recent trend to eliminate food price subsidies together with fossil fuel subsidies. Of course, such subsidies are mainly for dietary energy rich foods, which have contributed to the recent rapid rise in overweight, obesity and diet-related non-communicable diseases.
At the Second International Conference on Nutrition last November, an important consensus was achieved on addressing all types of malnutrition including hunger (inadequate dietary energy intake), ‘hidden hunger’ (micronutrient [vitamins, minerals, trace elements] deficiencies) as well as diet-related non-communicable diseases (typically, but not always associated with obesity).
However, I do not believe that serious nutritionists will insist on a single indicator conflating dietary energy undernourishment with other nutrient deficiencies. As Richard Jolly’s colleague Robert Chambers reminds us, the PoU does not fully consider the standard four dimensions of food insecurity: besides food availability and access, stability and utilization are also important, with the latter especially relevant for nutrition. As he correctly notes, a comprehensive and sustained effort to overcome undernutrition must also include other dimensions besides public health and food narrowly understood including hygiene and sanitation.
I have previously indicated, in various fora, that we at FAO are quite aware of the limitations of the PoU approach. This is the reason why FAO’s Voices of the Hungry project has been developing a Food Insecurity Experience Scale (FIES), which we believe will effectively complement the PoU estimates by providing a cost-effective tool for more nuanced assessment of the incidence of food insecurity at different levels of severity (therefore, not just of hunger) across the world. Consumption surveys are very expensive to conduct and most developing countries are unable to undertake them at representative scale regularly. I am afraid you seriously underestimate the costs of undertaking credible alternatives. This is no trivial matter as the monitoring costs estimates for the post-2015 SDGs are much greater than existing ODA flows!
Finally, I have to reject the presumption by some people, whom I otherwise respect, that civil servants cannot be trusted to report honestly as institutional self-aggrandizement imperatives will always prevail. Institutional self-aggrandizement is undoubtedly a problem for all institutions, not just public ones. Such blanket criticisms are often also made of national and international statistical reporting. International reporting depends on national reporting, and while there may be occasional misreporting for political reasons, such claims are easy to make, but often reflect ignorance of where most data comes from, and naiveté about how much the so-called data revolution can do to improve international development monitoring.
If we report good progress, we are accused of exaggerating progress for political reasons. On the other hand, if we report shortfalls, we are accused of deliberately doing so to secure more funding to keep ourselves in business. Sometimes, the same people make both criticisms as and when it suits them. I am ready to concede that these are not imaginary problems, but such criticisms often reflect a contempt for civil servants honestly trying their best to do their work, which is not helpful if we expect governments and international organizations to be part of the solution.
There is much we need to do to improve international development monitoring, especially during this period, and I do hope that all concerned will base their suggestions and criticisms on the modest progress which has been made in recent years, not least because of earlier criticisms made.
jomo
FAO (ES-ADG, Room B532), Vialle delle Terme di Caracalla, 00153 Roma, Italy. Office: [email protected] +39-0657053566 Websites: www.fao.org, www.jomoks.org, www.ideaswebsite.org
Thank you for your detailed criticisms. As you know, some of the issues you raise have been addressed before, especially in SOFI 2012, the first SOFI issued after I arrived here. You are quite right that the decision to improve the methodology was made before my arrival, but SOFI 2012 explicitly showed the implications for hunger estimates of different assumptions about calorie requirements. For some time, FAO (wrongly in my view) focused on food availability, ignoring access. The new estimates sought to give more attention to access as well as other factors previously ignored or wrongly assumed. We have also considerably improved the database, with more information from consumption surveys. While it is generally a bad idea to change criteria in the middle of monitoring exercise, the new PoU estimates are better than the old ones, in my humble opinion, although I will be the first to concede that limitations and problems remain.
You suggest that FAO should not have changed the methodology despite all the criticisms of the PoU, including from inside. The way in which the original methodology, developed by the Indian statistician Sukhatme during the 1960s, had been applied in the following years was felt to be inappropriate in several regards, and we showed the implications of the major changes made to the methodology and assumptions in the technical appendix. My colleagues believed that it was FAO’s responsibility to improve the methodology given the responsibility FAO has to report on the MDG1c indicator.
Using an explicitly low cut-off – related to the minimum dietary energy needs for a sedentary lifestyle – allows us to credibly improve the reliability of estimates. As dietary energy requirements vary considerably in any population, and we know very little about such variations in different contexts, we can much more reliably estimate the proportion of people who do not even get this minimal dietary requirement. We have made the database public, to allow others to make estimates using different benchmarks and methodologies. While we have been criticized for changing the methodology and assumptions, very little use has been made of the data available to come up with alternative estimates.
You correctly point out that the hunger trends are not consistent, but as you also correctly point out, the price spikes were in (early) 2008 and 2012 (one can add 2010-11 as well). My colleagues have long conceded that their provisional estimates for 2009 were wrong due to wrong assumptions regarding likely income contractions following the 2008 financial crisis and higher food price levels in many countries, overly influenced by the early pessimistic views on the extent and depth of the financial crisis of 2009 in developing countries, and the earlier cereal price spikes. What we now know is that developing countries continued to experience sustained growth in 2009 and 2010, and also that the impacts of such high international producer prices were mitigated in many countries by food export restrictions, food subsidies, etc., explaining the non-correspondence between international producer and domestic consumer food prices.
This is why we should be concerned about the recent trend to eliminate food price subsidies together with fossil fuel subsidies. Of course, such subsidies are mainly for dietary energy rich foods, which have contributed to the recent rapid rise in overweight, obesity and diet-related non-communicable diseases.
At the Second International Conference on Nutrition last November, an important consensus was achieved on addressing all types of malnutrition including hunger (inadequate dietary energy intake), ‘hidden hunger’ (micronutrient [vitamins, minerals, trace elements] deficiencies) as well as diet-related non-communicable diseases (typically, but not always associated with obesity).
However, I do not believe that serious nutritionists will insist on a single indicator conflating dietary energy undernourishment with other nutrient deficiencies. As Richard Jolly’s colleague Robert Chambers reminds us, the PoU does not fully consider the standard four dimensions of food insecurity: besides food availability and access, stability and utilization are also important, with the latter especially relevant for nutrition. As he correctly notes, a comprehensive and sustained effort to overcome undernutrition must also include other dimensions besides public health and food narrowly understood including hygiene and sanitation.
I have previously indicated, in various fora, that we at FAO are quite aware of the limitations of the PoU approach. This is the reason why FAO’s Voices of the Hungry project has been developing a Food Insecurity Experience Scale (FIES), which we believe will effectively complement the PoU estimates by providing a cost-effective tool for more nuanced assessment of the incidence of food insecurity at different levels of severity (therefore, not just of hunger) across the world. Consumption surveys are very expensive to conduct and most developing countries are unable to undertake them at representative scale regularly. I am afraid you seriously underestimate the costs of undertaking credible alternatives. This is no trivial matter as the monitoring costs estimates for the post-2015 SDGs are much greater than existing ODA flows!
Finally, I have to reject the presumption by some people, whom I otherwise respect, that civil servants cannot be trusted to report honestly as institutional self-aggrandizement imperatives will always prevail. Institutional self-aggrandizement is undoubtedly a problem for all institutions, not just public ones. Such blanket criticisms are often also made of national and international statistical reporting. International reporting depends on national reporting, and while there may be occasional misreporting for political reasons, such claims are easy to make, but often reflect ignorance of where most data comes from, and naiveté about how much the so-called data revolution can do to improve international development monitoring.
If we report good progress, we are accused of exaggerating progress for political reasons. On the other hand, if we report shortfalls, we are accused of deliberately doing so to secure more funding to keep ourselves in business. Sometimes, the same people make both criticisms as and when it suits them. I am ready to concede that these are not imaginary problems, but such criticisms often reflect a contempt for civil servants honestly trying their best to do their work, which is not helpful if we expect governments and international organizations to be part of the solution.
There is much we need to do to improve international development monitoring, especially during this period, and I do hope that all concerned will base their suggestions and criticisms on the modest progress which has been made in recent years, not least because of earlier criticisms made.
jomo
FAO (ES-ADG, Room B532), Vialle delle Terme di Caracalla, 00153 Roma, Italy. Office: [email protected] +39-0657053566 Websites: www.fao.org, www.jomoks.org, www.ideaswebsite.org