We must stop meating like this: Why it’s time for a better measure of meat consumption

We urgently need to reckon with the impact that eating meat has on the climate, human health, biothreats, animal welfare and much more. But delivering a proportionate response depends on knowing how much meat we are eating. And that, as this blog sets out, isn’t entirely clear.

World leaders didn’t seem to have much appetite for discussing the sustainability of the global food system at COP26. In fact, the furthest meat consumption got up the agenda was the outrage over serving burgers at the summit, likened by one activist to “serving cigarettes at a lung cancer conference”. Failing to discuss meat was a missed opportunity: we urgently need to reckon with the impact that eating meat has on the climate, human health, biothreats, animal welfare and much more. But delivering a proportionate response depends on knowing how much meat we are eating. And that, as this blog sets out, isn’t entirely clear.

A tale of two surveys

The UK government’s Family Food Survey (FFS) suggests that per-person meat consumption peaked in 1980 at 1160g purchased (for consumption in the home) each week. Since then, it has declined by 19% to 961g per week. Over those four decades, as Figure 1 shows below, dietary habits have shifted quite considerably too. Purchase of red meat has fallen by 64%, but we are buying around 50% more poultry, facilitated by the industrialisation of chicken farming. We have also changed the way we eat: the volume of ready meals, convenience, and takeaway meats bought has gone up 200%.

If you take the view that eating less meat makes reaching our climate ambitions more likely, there are some positives you can take from the Family Food Survey data, but they come with a heavy dollop of caution. The overall trend in meat-eating is down and we are eating noticeably less red meat – which is associated with some of the largest environmental (and health) externalities.

However, since reaching its lowest point in 2015, consumption has ticked back up by 3.5%. OECD data, set out in Figure 2, shows a similar pattern. Does this represent a full-blown reversal of dietary trends? That seems unlikely. What we can say though is that we are certainly not heading in the right direction fast enough: the Committee on Climate Change (CCC) has set a 20% reduction target to be reached by 2030.

From a meat-reduction perspective, this isn’t a particularly positive picture. So how do we reconcile it with this recent headline?

“UK public now eating significantly less meat”BBC News

The BBC story references a study in The Lancet[1], based on Oxford University scientists’ analysis of the National Diet and Nutrition Survey (NDNS). It shows that between 2008-2019 meat consumption fell by a remarkable 17%, down to 86.3g. By comparison, the FFS for the same period found only a 4% reduction, to 137g per day (at home). If we scale this up, the difference between the two survey estimates is the equivalent of 73 steaks a year.[2]

Those are stark discrepancies. What exactly should we make of them?

Firstly, while the level of meat consumption varies between the two datasets, the overall trend of the last decade is downwards. There is also no disagreement about the direction of travel for specific product categories, with red and processed meat broadly down and white meat more popular. The discrepancy is primarily a matter of scale.

Explaining the differences involves getting into the methodological weeds of the two measures. Whereas the FFS asks participants to keep a diary of food purchases over a two week period, NDNS participants record food consumption over four days, including approximating their portion sizes and leftovers. Put another way, the FFS provides a proxy for meat consumption by telling us what is in the fridge, whereas the NDNS tells us how much meat is on the plate.

Food waste could be part of the equation, since the FFS doesn’t capture what participants actually consume, only what they buy. Underreporting consumption is another factor. Shifting cultural norms and a growing perception that meat eating is a ‘deviant’ behaviour could lead to participants playing down how much they eat (one academic study has suggested women could be more likely to underreport than men). Plausibly, the NDNS could be prone to more underreporting since participants have to estimate how much they eat, whereas the FFS statistics are based on weight/volume of products purchased – though the FFS itself notes that underreporting may be an issue.

Should we be worried about variations in meat consumption estimates?

In short, yes, and there is an eyebrow-raising line in the Oxford study:

“…the small decrease reported in the NDNS and Family Food Surveys and the inherent uncertainty in these estimates, together with the apparent increase in meat available for consumption[3], precludes a firm conclusion that meat intake is convincingly in decline in the UK.”

If we can’t say for certain what is happening to meat consumption, I think there are three reasons to be concerned:

  1. Without reliably knowing the trajectory of meat-eating trends, how are policymakers to assess what a proportionate response is to help us reach net zero? If we adopt the FFS measure – which suggests some stagnation over the last five years – ‘harder’ interventions (such as a “meat tax”) might be more justifiable. If we adopt the NDNS measure – which paints a much more positive picture – we might be reaching for less restrictive tools (such as behaviour change interventions). Given the toxicity of ‘meat politics’, an accurate picture of consumption is vital to understanding the scale of the problem and justifying the appropriate intervention.
  2. Without reliable estimates of consumption, how are we to know the success of an intervention? If, for example, the government were to support the development of alternative meat products (something the SMF has called for) we would want to evaluate the impact this policy choice has on animal meat consumption. Lessons could be learned here from the development of more reliable alcohol consumption statistics, prone to inconsistencies in the past, following Scotland’s introduction of minimum unit pricing.
  3. Meat consumption statistics are a hot topic in the media. Stories based on consumption-based measures, like the BBC article highlighted above, can rather clumsily be linked to climate impact. These stories miss the point: it is the number of animals reared and slaughter that matters, not how many of them we eat.

Whether we look at the FFS, NDNS, or any other meat consumption measure, one thing seems adequately clear: we are not going fast enough in the right direction. And reaching the CCC’s 2030 20% target will not happen without government interventions. But exactly what those interventions should be, and how effective they are, depends on better measurement of meat. Improving data-collection on meat consumption should be the first step towards any new meat policy.


  • [1] The Oxford University study has several noteworthy conclusions, including no significant difference in meat consumption between men and women (as a proportion of energy intake) and between socioeconomic gradients, contradicting previous research. There has also been a 3% point increase in the number of non-meat eaters over the last decade, younger people have increase their consumption over time (compared to those born before 1999) and millennials eat more meat than those born in other decades.
  • [2] Estimate based on a 255g beef rump steak sold from Tesco
  • [3] A reference to data from the Agriculture and Horticulture Development Board suggesting that meat available for consumption in the UK has increased by 3g per person per day (adjusted for population size).



Related items:

Page 1 of 1