Global climate policy impacts on livestock, land use, livelihoods, and food security

Golub A A, Henderson B B, Hertel T W, Gerber P J, Rose S K and Sohngen B (2013).  Global climate policy impacts on livestock, land use, livelihoods, and food security, PNAS

This is a really interesting, but complex paper, so I have taken some time to try and summarise it.  I’m very grateful for help and clarification from the authors themselves.

Its focus is on examining a range of possible policies geared at reducing GHG emissions from the agriculture and land use sectors.  The paper examines their a. effectiveness in reducing emissions, b. their impacts on livelihoods (ie. incomes) both of land owning households/farmers and of non-farmers/landless labourers  and c. on food security, measured in terms of energy and protein intakes.

The study looks at mitigation options in forestry (afforestation, avoided deforestation and forest management) as well as in agriculture (both crops and livestock, and covering mitigation of N₂O, CH₄ and fossil fuel related CO₂).  With livestock it distinguishes between ruminants, dairy, and non ruminants (ie. largely pigs and poultry).

The analysis is based on the development of a model that incorporates three components: one modelling production and trade – ie. all production/trade including agriculture(GTAP); one modelling different land uses (agroecological zone or AEZ) and one modelling GHG emissions from different sectors (GHG).

It takes a global perspective, and explores three  GHG  mitigation policy scenarios as follows:

Scenario A: the introduction of a GHG tax on all sectoral activities (including but not restricted to the land based sector), together with a carbon sequestration incentive (affecting forestry) in Annex 1 countries (ie. countries who have commitments to reduce their greenhouse gas emissions – rich countries)

Scenario B: the introduction of a. a GHG  tax in Annex I countries only plus a carbon sequestration incentive globally (ie. Annex I and non annex I countries who do  not have a legal requirement to reduce their emissions).  Annex II countries pay for the carbon sequestration of non Annex 1 developing countries. (Annex II countries are a subset of Annex I with obligations to facilitate mitigation in developing countries through the provision of financial resources and transfer of technology.)

Scenario C: the introduction of both a GHG tax and a carbon sequestration incentive in all countries – but non Annex 1 countries receive a refund on their tax from Annex II; the additional forest carbon sequestration in non Annex I is again paid by the wealthy Annex II countries.

(nb if you are confused about Annex I and non Annex II countries – Wikipedia is helpful here….)

The cost for GHG emissions (CO2eq) is set at $27/tonne CO₂ eq.

The study considers the effect of these policy scenarios both on changes in emissions intensity (ie. CO₂eq per unit of output) and on absolute output (ie. food).

The study finds the following:

Scenario A: (GHG tax)

Impact on GHGs: In this scenario, output in all three livestock sectors in Annex 1 countries is scaled back, particularly so in the case of ruminant production.  Dairy and nonruminant sectors suffer little reduction in output but significant improvements in emissions intensity.  However this scenario also leads to compensatory increases in production in non Annex 1 (ie. developing countries) – especially in livestock production  -and also to increases in deforestation in these regions.  These, taken together mean that 55% of the GHG savings in land using sectors of Annex I countries are ‘leaked to’  other countries not affected by the tax, so undermining the overall savings achieved.

Impact on livelihoods: Landless agricultural workers and unskilled urban households in Annex 1 (rich) countries suffer under this scenario while those in non-Annex 1countries benefit through the leakage effect – increased income opportunities.  The same is true of non agricultural households.

Impact on food security:  even though incomes rise in non-Annex 1 countries, they are not enough to offset the higher food prices resulting from a drop in food production in Annex 1 countries for some people.  The effects on food security are mixed – there are an almost equal number of winners and lowers.  Food security falls more significantly in Annex 1 countries.

Scenario B: (carbon sequestration incentive)

Impact on GHGs:  the GHG reductions here are much more significant and the vast majority of them come from the forestry sector.

Impact on livelhoods: incomes fall for landless labourers but rise for farm households and those involved in the forestry sector.  Farm households benefit because the constraints on land use lead to higher prices for agricultural commodities.

Impact on food security: a reduction in food output (as land is switched to afforestation etc) means a fall in food consumption except for farming households where income benefits mean they can afford to buy more food

Scenario C (sequestration incentive plus GHG tax, which is paid back in the case of developing countries)

Impact on GHGs: This scenario offers the greatest reductions in emissions, and delivers the greatest reductions in livestock emissions from non-Annex 1 countries – just over half of global livestock abatement.

Impact on livelihoods: in unskilled labour households, incomes fall in rich Annex 1 countries but rise in all but 2 non Annex 1 countries.  Farm households as usual benefit, except in Russia.  (Indeed, across all three scenarios, the most dramatic negative impacts on regional income and consumption are found in Russia. This is because the emissions tax is imposed not only in land using sectors, but in industrial sectors – and for Russia the effects on their industry is the main driver for these large negative impacts in Russia.

Impact on food security: for unskilled labour households, the impacts are mixed but generally negative.  They are positive for farm households. 

Some of the key points / questions to arise from this paper:

Livestock (and agriculture) mitigation options: Mitigation responses represent the range of manure and enteric, cropland, and paddy rice mitigation technologies characterized by U.S. EPA (2006). Fertiliser management is also included but emissions savings are credited to the crop sector – although of course a significant share of crops will be grown for animal feed.  Animal welfare implications are not considered, nor the impacts on water use.

Estimates of mitigation potential: Around 6 GT CO₂ eq can be mitigated if Scenario C ( which holds the most potential) were applied. This is roughly equivalent to the IPCC’s 2007 estimate of agriculture’s direct emissions (ie. methane and nitrous oxide, but excluding agriculturally induced land use change and subsequent CO₂ release).

If agriculturally induced land use change is included as part of agriculture’s ‘responsibility,’ as well as small amounts from on-farm fossil fuel use and fertiliser production (see esimates in this report here), then the mitigation potential accounts for just under half of agriculture’s impacts today, although not necessarily what they are anticipated to be tomorrow, assuming an increase in overall production (the IPCC’s 2007 chapter assumes an increase).

The vast majority of the mitigation potential quantified in this paper comes from the forestry sector.  Within agriculture most of the mitigation potential is from the livestock sector, mainly ruminants.  For example in Scenario C, of the total 5324 MT CO₂ eq savings, only 12% comes directly from the agricultural (including livestock) sector.  The agricultural (excluding forestry) mitigation potential of 667MtCO2 eq amounts to about 10% of the 2007 IPCC’s estimate of GHG emissions from agriculture (and presumably the figure will be higher now).  This estimate of the mitigation potential from agriculture alone (ie. excluding forestry related sequestration) is lower than the lower end of the IPCC’s mitigation estimate.

The table, copied from the web appendix, provides some figures.

Since agriculture and forestry are economically linked to all other sectors, changes in these sectors, as well as the global impact of a GHG  tax, will have impacts on global emissions, which this paper puts at about 10GT of CO₂ eq savings.  For reference, global emissions in 2010 stood at about 50.1GT CO₂ eq.  Note though that the baseline year for emissions in this study is 2001.  At that time total emissions from fossil fuels and non-CO2 were 32GtCO2eq, excluding emissions from deforestation.

Finally the paper doesn’t consider the mitigation potential arising from consider soil C sequestration in either grasslands or croplands. Agriculture’s contribution to mitigation would be much higher if it did and the mitigation potential perhaps aligned, perhaps, with the more optimistic end of the IPCC’s estimates.

Which sector gets the credit for emissions reductions?  The linking of forestry and agriculture together under the category ‘land use sector’ clearly makes sense but also raises interesting questions as to which sectors get the ‘credit’ for mitigation (note that the paper itself doesn’t talk in these terms but it an issue that it, arguably does raise). This paper suggests that substantial reductions in emissions from agriculture are achievable with little impact on food security.  However this only applies if the emission savings from forestry are credited to the agricultural sector – very  little mitigation comes from agriculture itself – and other sectors may have a different view on this.  As noted while the paper doesn't credit forestry savings to agriculture (since both activities are lumped together as 'land using sectors') politically speaking it is easy to see how different interest groups may wish to interpet the findings. As well as the agricultural sector, arguably, the aviation or terrestrial transport sectors may want to claim reductions via paying for carbon offsetting activities in the forestry sector, which means that agriculture’s share of the mitigation credit pie will be reduced.  There are only so many forestry credits to go round.

Long term mitigation: as noted, the bulk of the mitigation potential is related to forestry (avoided deforestation, afforestation, forestry  management). The forest sector modelling baseline projects declining deforestation emissions over time and the mitigation supply reflects this as well as the implications of avoiding deforestation in earlier years on later year mitigation potential.

 Role of developing countries: The paper makes the following important point: “There is the temptation to try to address concerns about food  security and agricultural development by exempting non-Annex I regions from climate policies. However, this exemption can lead to high emissions leakage rates in livestock sectors and agriculture as a whole. The prospect of large leakages because of expansion in non-Annex I production could easily derail the inclusion of agriculture in future mitigation strategies.”

Effects on meat consumption:  as you can see from this table, consumption of ruminant meat declines in scenario B (dark grey parts of the bars) and in Scenario A  for rich countries. In Scenario C (GHG  tax plus sequestration incentive) there is an increase in all forms of animal product consumption in developing countries matched more or less by declines in rich countries.  This is because developing countries benefit from the sequestration incentive as well as a reimbursement of the carbon tax and of course most of the emissions reductions come from the forestry sector.

Effects on consumption overall:  Note that the  baseline against which consumption increases are measured are per capita regional consumption intakes in 2001.  Obviously consumption now is higher and will be higher still in 2030 .  Although the reductions in consumption in poor countries are small, while in rich countries the impacts may not give cause for concern, I’m  not sure whether this is the case for very poor people who are already consuming very inadequate diets.   I’m not sure what the implications of this study are when trends in consumption and production as well as population growth are factored in - clearly land use pressures will be greater and so the potential mitigation-food security trade offs are likely to be higher.

Biofuels: Note that biofuels incentives are not considered in this study nor the consequent impacts on GHGs, incomes/livelihoods, and food security.  However some very simple and preliminary analysis: can be found here: http://ageconsearch.umn.edu/bitstream/61592/2/golubAAEA2010.pdf

You can download the paper and its web appendix here.

A useful summary can also be found on the CCAFS website here.

As mentioned, the long web appendix / supporting document is a very useful source of additional information.

Comments on and further analysis of this paper would be extremely welcome –  please do add these in the comments box below.  Note that you will need to be logged into the website to do so – if you can’t remember your log in details (the ones you used to sign up as an FCRN network member) please get in touch.