28 October 2015

The Voices of the Hungry Speak! What Do They Say?

The measurement of hunger is devilishly difficult.  And the world does it very badly.  FAO has come in for its fair share of criticism, providing the "undernourishment" figures which are based on food supply data, modified imperfectly, for purchasing power.  These are the "795 million" who go to bed hungry each night as reported on in the State of Food Insecurity (SOFI) and used in various hunger indicators, including the Global Hunger Index.

So it is to FAO's enormous credit that they have spearheaded an imaginative new initiative, the Voices of the Hungry programme.  The initiative was described in the Global Nutrition Report of 2014, and now it has some results, coming soon.

The initiative is called "Voices of the Hungry" because it involves asking people 8 questions aimed at assessing their "food insecurity experience". The questions have simply yes/no answers.

The simplicity of the module allows it to be included in Gallup World Poll questionnaires, administered to individuals of 15 years and older, in 147 nationally representative surveys.  Sample surveys are typically 1000 individuals although in larger countries the samples are several times larger (e.g. China and India).  Interviews are face to face, although in medium and high income countries with more than 80% telephone coverage, the interviews are done by phone.  So raw scores can range from 0-8, with ad-hoc, but reasonable, adjustments for maximum and minimum scores.  Thresholds are established for moderate hunger (FI-moderate) and for severe hunger (FI-severe).

But how can these experiences be converted into scores that allow comparisons across countries? A global standard is developed and then each country's score is adjusted to the global standard.  This part of the report is hard to follow and some worked examples would have really helped.  The assumptions behind the procedures are thoroughly tested, and, for nearly all countries, they hold.

So what are the results?

The report stresses that these are preliminary, nevertheless they are fascinating.

1. The country rankings are fairly closely related to the country rankings using the FAO undernourishment score.  Out of 21 potentially related country indicators, the VoH moderate food insecurity (FI-moderate) prevalence measure is most strongly correlated (Spearman Rank) with under 5 mortality rates.  The country rank on the FAO undernourishment score is the 11th most strongly correlated indicator at  0.757.   The link with under 5 stunting is the 15th most strongly ranked at 0.666

2. More interestingly, the levels of food insecurity are very different from the undernourishment prevalence estimates.  For example, South Africa has an undernourishment rate of less than 5% which signals not much food insecurity, but the FI-moderate for South Africa is 41.2% and FI-severe is 21%.  A very different picture.  Haiti and South Sudan both have FI-severe rates above 70% and this may not be too surprising given conditions in these contexts, but the high level of FI severe for Malawi (56.1% versus an undernourishment % of 20.7) is more surprising and the low level of FI-severe for Ethiopia (12.1% versus an undernourishment % of 32.0 ) might be a surprise to many.

3. The indicator FI-moderate sheds light on hunger in high income countries, which is really useful too.  FI-moderate for the US and the UK is around 10% (about 40 million people!) and only slightly lower than Indonesia (13%). The social democracies of Denmark, Germany, Japan, Netherlands, Norway, Sweden and Switzerland have FI-moderate rates below 5% although the authors say there were too few complete cases to allow robust estimation.

So, the VoH data are as yet fully robust, but are really welcomed.  They are collected face to face, at individual level, in a timely way (annually), and are comparable across countries.  I think they will grow to be a very useful way of measuring food security for all countries--and remember, the SDGs apply to all of us.

Some will be wary of Gallup's involvement. Are they doing this to sell products? To advance their bottom line? Perhaps. But they are doing it transparently.  This will obviously have to be watched.
I was an enthusiast of this proposed project back in 2011 and so I am pleased to see it come to initial fruition.  Well done to FAO, Gallup and DFID for having the imagination to see it -- and to do it.

No comments: