This chapter summarises the most recent trends in ammonia and greenhouse gas emissions (GHG) indicators, the main air pollutants from agricultural activities, from the OECD agri-environmental indicators database. It also conducts an econometric exercise to estimate the relationship between GHG emission intensities and labour productivity, and discusses how New Zealand is tackling GHG emissions intensities.
Trends and Drivers of Agri‑environmental Performance in OECD Countries
Abstract
The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.
Trends in agricultural ammonia and greenhouse gas emissions (GHG) indicate a deterioration of agriculture’s performance in the OECD area. While GHG emissions were practically unchanged in the period 1993-2005, they increased by 0.2% yearly in OECD countries from 2003 to 2015. Ammonia emissions decreased in the period 2003-15, but at a slower rate than during the period 1993-2005.
OECD countries need to address this increase in emissions, which stems primarily from the use of synthetic fertilisers.
The capacity of countries to produce agricultural goods while minimising GHG emissions has weakened. Although GHG emissions per dollar of agricultural production (emission intensities) continued to decline in OECD countries in the period 2003-15, it was at a slower rate than during the period 1993-2005.
In highly productive OECD countries, continued improvements in labour productivity will not necessarily translate into a decrease in GHG emissions intensities. Indeed, these countries may be reaching a productivity level at which further improvements may induce more GHG emissions per unit of output.
The New Zealand case study shows that reducing emission intensities, while maintaining agricultural production is possible when there are policies in place focused on research and development, particularly targeting farm profitability, productivity and emission intensity reductions in tandem with low levels of distortionary support to agriculture.
Agricultural activities affect air quality mainly via greenhouse gas (GHG) and ammonia (NH3) emissions. It is the main emitter of methane (CH4) and nitrous oxide (N2O), two non-CO2 greenhouse gases with more potential to warm the atmosphere than carbon dioxide (CO2), but with a shorter lifespan (IPCC, 2014[1]) GHG emissions from agriculture represent 10-12% of total global GHG emissions (Smith et al., 2014[2]). Worldwide, nearly 40% of agricultural GHG emissions come from ruminants’ digestive process (enteric fermentation) and 30% from agricultural soils; the remaining 30% comes from rice cultivation, biomass burning, and manure management (Tubiello et al., 2013[3]).
Agriculture’s link to greenhouse gas (GHG) emissions and climate change is complex. While the sector is a contributor of GHGs to the atmosphere, agricultural soils can act as carbon sinks depending on how these are managed (OECD, 2008[4]). Agriculture is not only responsible for GHG emissions due to the direct management and operation of farms but also indirectly due to the conversion of natural habitats such as forested lands and peatlands to agricultural fields. The agricultural sector is projected to be the second sector to contribute the most to economic damages from climate change, only after losses associated with health (OECD, 2015[5]). The impacts on the sector are likely to be differentiated by space, time and crop, with some regions, especially in higher latitudes, benefitting from climate change, while regions near the Tropics will suffer the most (Smith et al., 2014[2]; OECD, 2015[5]). In some regions, higher CO2 concentrations in the atmosphere, which tend to improve photosynthesis and increase yields, could more than compensate the potentially negative effects of hotter temperatures (Barros et al., 2015[6]; Murgida et al., 2014[7]).
Agriculture also accounts for 80-90% of total ammonia emissions globally (Bouwman et al., 1997[8]; Zhang et al., 2010[9]; Xu et al., 2019[10]) via volatilisation from livestock manure and synthetic mineral N fertiliser application (Bouwman et al., 1997[8])). Ammonia emissions are associated with two major types of environmental problems: acidification and eutrophication (OECD, 2008[4]). When combined with water in the atmosphere or after deposition, ammonia contributes to acidification of soil and water. Excess soil acidity can harm certain types of terrestrial and aquatic ecosystems. Deposition of ammonia can also increase nitrogen levels in soil and water, which may lead to eutrophication – algal and plant growth due to excess nutrients – in aquatic ecosystems (OECD, 2008[4]). Human exposure to high concentrations of NH3 can affect the respiratory track and lung function (OECD, 2018[11]). NH3 is also a precursor of particulate matter (PM), a potent air pollutant that poses risks to human health (OECD, 2018[12]).
Both GHG and ammonia emissions are transboundary pollutants, affecting areas beyond those where they are emitted. Therefore, international accords are paramount to effectively reducing such emissions.
Agricultural GHG emissions in the OECD area are rising
Agricultural GHG emissions in the OECD area increased by 26 million tonnes of CO2 equivalent, from 1.32 Gt of CO2 equivalent in the period 2003-05 to 1.35 Gt of CO2 equivalent in 2013-15 (Figure 2.1). The average annual growth rate for this period was 0.2%, while the annual growth rate in the period 1993-2005 was slightly negative (-0.02%). Compared to the period 1993-2005, in the most recent period of analysis fewer countries registered negative growth rates and only five countries – Greece, Israel, Italy, Spain and the United Kingdom – had growth rates lower than -0.5%, while 21 countries did in the period 1993-2005.
The share of agriculture in total OECD GHG emissions was 9% in 2013-15. The relative contribution of agriculture in the total of national GHG emissions varies across countries, with six having a share of 15% or higher in 2013-15 (Denmark, France, Ireland, Latvia, Lithuania and New Zealand), although the contribution of these countries to the total OECD agricultural GHG emissions was low except for France (5.7%). The EU15 and the United States accounted for 66% of OECD agricultural GHG emissions in 2013-15.
Higher agricultural soil emissions explain most of the increase in GHG emissions in OECD countries during the period 2003-15 (Figure 2.2). With the exception of Iceland, Korea, Luxembourg, Mexico, Switzerland and Turkey, agricultural soil emissions accounted for more than 50% of the increase in GHG emissions in countries where these emissions increased in the period 2003-15. In the OECD area, the main GHG source that declined during this period was enteric fermentation, while manure management and agricultural soils increased. For half of the countries that saw a decrease in their GHG emissions, enteric fermentation accounted for more than 50% of that decline.
Emission intensities in OECD countries continued to decline in the period 2003-15, but at a lower speed than in the period 1993-2005. Emission intensities were 2 kg of CO2e/USD in 1993-95, 1.8 kg of CO2e/USD in 2003-05, and 1.7 kg of CO2e/USD in 2013 15 (Figure 2.3). The top five countries that saw the largest decreases in emission intensities from 2003 to 2015 were Australia, Israel, Chile, New Zealand, and Spain. While in the period 1993-2005 only three countries – Latvia, Japan and the United Kingdom – increased their intensities, twelve countries did from 2003 to 2015. Moreover, four of the top five largest GHG emitters in the OECD area – France, Germany, Mexico and the United States – slowed the rate of decline in intensities in the period 2003-15; Turkey, the remaining country in the top five, increased its emissions intensity at a rate of 0.4% per year.
Notes: Countries are ranked in ascending order according to average annual percentage change 2003-05 to 2013-05.
1. For Israel, 1993-95 is replaced by 1996. 2. For Chile, 2013-15 is replaced by 2011-13. Source: (OECD, 2018[13]).
Percentage change in GHG emissions from 2003-05 to 2013-15
Notes: The category "other" include liming, urea application, Other carbon-containing fertilisers, Other CO2, Rice cultivation, Prescribed burning of savannas (CH4), Field burning of agricultural residues (CH4), Other CH4, Prescribed burning of savannas (N2O), Field burning of agricultural residues (N2O) and Other N2O emissions sources.
1. The OECD total does not include Chile
Source: (OECD, 2018[13]).
Notes: Countries are ranked in descending order according to average annual percentage change 2003-05 to 2013-05. Greenhouse gas emissions are per gross production value (in constant 2004-06 USD). 1. For Israel, 1993-95 is replaced by 1996. 2. For Chile, 2013-15 is replaced by 2011-13. 3. The OECD and EU15 do not include Belgium and Luxembourg for the period 1993-95.
Sources: Greenhouse gas emissions were obtained from OECD (2018[13]) and Gross Production Value from FAOSTAT (2018[14]).
Ammonia emissions declined in OECD countries
Ammonia emissions in the OECD area decreased in the period 2003-15, but at a slower rate than during the 1993-2005 period. While a majority of countries decreased their emissions in the most recent period of analysis, Austria, Estonia, Germany, Iceland, Latvia, Luxembourg, and Switzerland reversed those trends and increased their emissions in the period 2003-15 (Figure 2.4).
International agreements to reduce emissions have played a critical role for reducing ammonia emissions. The 1999 Gothenburg Protocol to Abate Acidification, Eutrophication and Ground-level Ozone (Gothenburg Protocol) sets national ceilings for 2010/2020 for four major pollutants: sulphur emissions, nitrogen oxides (NOx), volatile organic compounds (VOCs) and ammonia (NH3) (UNECE, 2018[15]). The ceilings were negotiated and agreed to on the basis of scientific assessments of pollution effects and abatement options. The ceilings are more stringent for Parties whose emissions have a severe environmental or health impact and for those whose emissions are relatively cheap to reduce (UNECE, 2018[15]).
To meet the targets at the national level, guidance documents and the Protocol provide a wide range of abatement techniques and measures, as well as economic instruments to reduce emissions in relevant sectors. In the case of agriculture, the Protocol establishes that within a year of the entry into force of the Protocol, signatory Parties need to take the following measures (United Nations, 2013[16]):
establish, publish and disseminate an advisory code of good agricultural practice to control ammonia emissions
take steps to limit ammonia emissions from the use of solid fertilisers based on urea and prohibit the use of ammonium carbonate fertilisers
ensure that low-emissions slurry application techniques are used and that solid manure applied to land shall be ploughed and incorporated into the soil within 24 hours of spreading
for new slurry stores on large pig and poultry farms, low-emissions storage systems will be used and for existing slurry stores on large pig and poultry farms, emissions will be reduced by 40%
new housing systems shown to reduce emissions by 20% will be used for new animal housing on large pig and poultry farms.
Specific abatement guidelines to implement these measures were circulated by the Executive Body to the Convention on Long-range Transboundary Air Pollution. The first set of guidelines was published in 1999 and has since been updated twice as new evidence and technologies become available. The most recent guidelines include abatement recommendations pertaining to the following (UNECE, 2014[17]):
nitrogen management, taking into account the whole N cycle
livestock feeding strategies
animal housing techniques
manure storage techniques
manure application techniques
fertiliser application techniques
other measures related to agricultural N
measures related to non-agricultural and stationary sources.
Abatement strategies are presented with their potential abatement potential and their associated costs. Optimised land application of slurry and improved livestock feeding strategies tend to be the most cost-effective practices (United Nations Economic Commission for Europe, 2015[18]). Communicating practical information to farmers through guidelines has been an important factor in the adoption of such practices (Defra, 2018[19]; UNECE, 2014[17]).
Notes: Countries are ranked in ascending order according to average annual percentage change 2003-05 to 2013-05.
1. For the United States, data for agricultural ammonia emissions have been estimated based on the ratio agricultural ammonia/total ammonia emissions, using the share 90% as recommended by USEPA.
2. The OECD total does not include Australia, Chile, Japan, Mexico and New Zealand for both periods, and does not include Israel for 1993-95.
3. For Korea, for agricultural ammonia emissions, 1993-95 is replaced by 1990 and 2013-15 is replaced by 2012-14. For total ammonia emissions 2013-15 is replaced by 2012-14.
Source: (OECD, 2018[13]).
In May 2012, the UN Economic Commission for Europe agreed on the amendments to the Protocol and set up new national emission reduction commitments for main air pollutants to be achieved in 2020 and beyond (Table 2.1).
Notes: For Spain, figures apply to the continental European territory.
1. The information in this document with reference to “Cyprus” relates to the southern part of the Island. There is no single authority representing both Turkish and Greek Cypriot people on the Island. Turkey recognises the Turkish Republic of Northern Cyprus (TRNC). Until a lasting and equitable solution is found within the context of the United Nations, Turkey shall preserve its position concerning the “Cyprus issue”.
2. The Republic of Cyprus is recognised by all members of the United Nations with the exception of Turkey. The information in this document relates to the area under the effective control of the Government of the Republic of Cyprus.
Source: Annex II of the 1999 Protocol to Abate Acidification, Eutrophication and Ground-level Ozone to the Convention on Long-range Transboundary Air Pollution (United Nations, 2013[16]).
To reach the goal of keeping the increase in global temperature below 2°C this century while maintaining economic growth, countries will need to reduce the emissions per unit of output (emissions intensities) in all sectors of the economy. The agricultural sector is no exception, where lowering emissions must be accompanied by output expansion to meet increasing food demand from a growing and wealthier population. A failure to do this could lead to price hikes and political unrest in certain regions of the world. In the past, productivity growth and agricultural area expansion drove food supply growth (Foley et al., 2011[20]). Demonstrating the relationship between productivity growth and emissions intensity can help to understand the role of productivity growth in tackling global warming.
In OECD countries, the growth of agricultural labour productivity is concomitant with emission intensity reductions but only up to a point; thereafter, emission intensities do not decrease and could even increase when labour productivity increases. Using the greenhouse gas emissions data from the OECD agri-environmental indicators (OECD, 2018[13]), in combination with farm labour statistics from USDA (USDA, 2018[21]) and data on agricultural gross production from FAO (FAOSTAT, 2018[14]), Figure 2.5 plots the estimated1 association between agricultural labour productivity and a) GHG, b) CH4 and c) N2O and emission intensities. Emission intensity is defined as GHG emissions per dollar of value of agricultural production (Annex 2.A) and agricultural labour productivity is defined as the ratio of gross production value to the number of workers economically active in agriculture. While this indicator is only a partial productivity measure as it excludes capital and other variable inputs, it is an appropriate indicator to reflect the long-term evolution of the sector and its structural transformation as it is less responsive to changes in variable inputs (Coderoni and Esposti, 2014[22]).2 The data used for producing the figures represent 33 countries3 during the period 1990-2015.
The point at which the relationship between GHG, CH4 and N2O emission intensities and labour productivity levels off is found at a level of USD 20 000/worker (Figure 2.5). In 2015, the median labour productivity in OECD countries was USD 44 700/worker, indicating that most countries are already beyond the levelling-off point. This may suggest that further improvements in labour productivity will not necessarily translate in a decrease of emission intensities. Therefore, productivity improvements may not be enough to improve emissions intensities; specific policy action may be needed to reduce emissions per unit of output.
For GHG and CH4, the relationship between emissions intensities and labour productivity is nonlinear and after a USD 20 000/worker level in productivity, emission intensities increase up to a given point (USD 54 000/worker), after which emissions intensities tend to decrease again. The shape of the relationship between N2O emission intensities and labour productivity is relatively flat compared to GHG and CH4, and, for N2O emissions intensities, productivity improvements beyond the levelling-off point do not seem to affect emission intensities. This more moderate relationship may be driven by the fact that reductions in N2O emission intensities are mostly driven by a decrease in the use of fertilisers, which may not necessarily translate into a labour force reduction.
Notes: All variables were transformed to non-logged values. Dash lines show the corresponding 95% confidence intervals.
Sources: Gross Production Value was obtained from FAOSTAT (2018[14]), measured in constant 2004-2006 million USD. Agricultural labour was obtained from USDA (2018[21]), measured in 1 000 workers. Greenhouse gas emissions were obtained from the OECD Agri-environmental Indicators database (OECD, 2018[13]).
The negative nonlinear relationship between emission intensities and labour productivity is confirmed via a parametric regression analysis.4 Increases in labour productivity are accompanied by declines in emission intensities (Table 2.2) but this negative relationship becomes less negative as productivity increases (the quadratic term is positive), reaching a turning point (the cubic term is negative) after which the relationship can become negative again. These results are consistent with Figure 2.5. There is also persistence in emission intensities: past emission intensities tend to define current intensities (Lagged Emissions Intensity coefficient is positive and statistically significant).
Notes: Coefficients were estimated using Arellano-Bond one-step GMM estimation and standard errors are shown in parentheses.
*, ** and *** represent statistically significant coefficients at the 10%, 5% and 1% levels, respectively. Due to data availability, Slovenia is excluded. Belgium and Luxembourg, and the Czech Republic and the Slovak Republic are combined, respectively. Difference in observations between GHG and others comes from a lack of data on GHG emission in 2014 and 2015 for Chile. All variables were transformed into logarithms. Year dummies were included.
Sources: Greenhouse gas emissions were obtained from the OECD Agri-environmental Indicators database (OECD, 2018[13]), labour data come from USDA (2018[21]) and value of production from FAOSTAT (2018[14]).
New Zealand registered one of the largest declines in greenhouse gas emissions per value of production in the OECD area, agricultural production growth, and a sharp reduction in agricultural land. This set of events are more notable considering the large share of agriculture in New Zealand’s economy (7%) and its specialisation in livestock production (especially dairy products and sheep meat) (OECD, 2018[23]), a sector characterised by high emission intensities. From 1990 to 2015, the intensity of New Zealand’s agricultural GHG emissions decreased 34%, a negative rate higher than both OECD average (-22%) and the average of the top 10 countries with the largest values of agricultural labour productivity (excluding New Zealand) in 1990 (-19%) (Figure 2.6). Emission intensity reductions were achieved in both N2O and CH4, and, in both cases, were larger than OECD countries as a whole and most productive countries as of 1990. As measured on a per unit of product (kg of meat or milk), emissions intensities have declined 20% in New Zealand’s pastoral agriculture (Parliamentary Commissioner for the Environment, 2016[24]). While total GHG emissions from agriculture increased by 13% from 1990 to 2015, these would have been higher without emission intensities improvements (Ministry for the Environment, 2018[25]).
These achievements are mainly explained by three factors: 1) the adoption of policies focused on research and development, farm profitability, productivity and emissions intensity reductions; 2) changes in the production mix of animal species; and (3) low levels of distortionary support to agriculture (Henderson and Lankoski, 2019[26]). From 1990 to 2016, New Zealand became more specialised in the production of dairy products. The population of sheep decreased by 52.3% and non-dairy livestock by 23.1%, while the size of the dairy herd increased by 92.4% (Ministry for the Environment, 2018[25]). Land use for sheep, beef and deer grazing decreased by 31.6%, whereas it increased by 71.7% for dairy grazing (Ministry for the Environment, 2018[25]). New Zealand’s support to farmers is one of the lowest in the OECD area (below 1% of gross farm receipts) and agricultural policies focus on key general services such as agricultural knowledge, innovation and biosecurity, which represent more than 70% of total support to agriculture (OECD, 2018[12]).
Note: Emissions intensity is the ratio of greenhouse gas emissions to agricultural gross production value.
Sources: GHG emissions were obtained from OECD AEIs (OECD, 2018[13]) and agricultural gross production value was obtained from FAOSTAT (2018[14]).
The government strongly supports innovation and technology transfers to reduce GHG emissions of the agricultural sector and is an international leader in supporting research efforts in this area. New Zealand has established dedicated institutions and R&D funding to reduce agriculture’s GHG emissions, including the New Zealand Agricultural Greenhouse Gas Research Centre (www.nzagrc.org.nz), the Pastoral Greenhouse Gas Research Consortium (www.pggrc.co.nz), and the Sustainable Farming Fund. In addition, the country leads the Global Research Alliance on Agricultural Greenhouse Gases (www.globalresearchalliance.org) which aims to share knowledge and expertise on reducing GHG emissions across 56 member countries. R&D institutions in New Zealand work closely with farmers and industry to develop mitigation technologies and options that are economically attractive (Ministry for the Environment, 2017[27]); they also organise workshops, meetings and presentations with and to relevant stakeholders (Lissaman, Casey and Rowarth, 2013[28]; Kerr et al., 2013[29]; Payne, Turner and Percy, 2018[30]). Since 1990, New Zealand has reduced the emission intensities of the sector primarily by improving pasture management, nutrient management, animal selection and genetics, and animal health.
Urease inhibitors have been used as a mitigation technology since 2001 and their adoption rates have been increasing since 2014. In New Zealand, urea is the main type of nitrogen fertiliser applied to pastures; urease inhibitors restrict the action of the enzyme urease which produces ammonia emissions (Ministry for the Environment, 2018[25]). Inhibitors reduce by half the fraction of nitrogen from synthetic nitrogen fertiliser that volatilises as NH3 (Saggar et al., 2013[31]). Urease inhibitors adoption rates have been relatively low but, since 2014, they have increased. The percentage of urea fertiliser that includes urease inhibitors sold from 2001 to 2013 in New Zealand was 6%. In 2014, the percentage increased sharply to 20% and, from 2014 to 2016, it has been 21% on average.
Looking ahead, New Zealand has clear GHG reduction targets both internationally and nationally. It set a target at -5% below 1990 levels by 2020 under the United Nations Framework Convention on Climate Change (UNFCCC), and at ‑11% below 1990 levels by 2030 under the Paris Agreement. In 2018, the government proposed the Zero Carbon Bill that set the national gazetted target at -50% below 1990 levels by 2050. There are currently ongoing discussions to define the future target under the Zero Carbon Bill policy.
To attain these goals, emission reductions in the agricultural sector are expected to be achieved through a combination of policies and technological improvements. The main policy instrument for reducing GHG emissions in New Zealand is the Emissions Trading Scheme (ETS). Under the ETS, agriculture has reporting obligations but not surrender obligations. The New Zealand government has projected that improvements in emissions intensities will continue and that, in combination with the implementation of the National Policy Statement for Freshwater Management (the main policy to improve water quality) and government schemes to incentivise forestry, the agricultural sector could achieve a 4.8% reduction of the projected emissions in the period 2016-30 as compared to a scenario without policy interventions (Ministry for the Environment, 2017[32]). Additional reductions (up to 10%) may be achieved by increasing adoption of readily available technologies to reduce emissions, but relevant adoption barriers such as lack of education and environmental awareness, risk aversion, and lack of trust in extension services still remain (Ministry for Primary Industries, 2018[33]).
A key question is whether the observed negative trends of emissions intensities can be maintained without affecting productivity growth. In spite of outstanding achievements in emission intensities reductions, productivity growth may be an area of concern. From 1990 to 2015, accumulated total factor productivity (TFP) growth was 40% in New Zealand; such a rate is lower relative to the one that other highly productive countries achieved over the same period (50%) (Figure 2.7). If measured by gross production value per worker, New Zealand ranked 7th (56% increase) in terms of productivity growth among the top 10 most productive countries in 1990, and that rate was almost half the average for OECD countries (109%) (Figure 2.7).
Percentage growth, 1990-2015
Note: Labour productivity is the ratio of agricultural gross production value to number of workers in the agricultural sector.
Sources: Agricultural gross production value was obtained from FAOSTAT (2018[14]), labour and TFP indices were obtained from USDA (2018[21]).
References
[41] Arellano, M. and S. Bond (1991), “Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations”, The Review of Economic Studies, Vol. 58/2, p. 277, http://dx.doi.org/10.2307/2297968.
[6] Barros, V. et al. (2015), “Climate change in Argentina: Trends, projections, impacts and adaptation”, Wiley Interdisciplinary Reviews: Climate Change, Vol. 6/2, pp. 151-169, http://dx.doi.org/10.1002/wcc.316.
[8] Bouwman, A. et al. (1997), “A global high-resolution emission inventory for ammonia”, Global Biogeochemical Cycles, Vol. 11/4, pp. 561-587, http://dx.doi.org/10.1029/97GB02266.
[43] CEIP (2019), 1999 Gothenburg protocol under the LRTAP Convention, http://www.ceip.at/ms/ceip_home1/ceip_home/gothenburg_protocol/.
[38] Choi, I. (2001), “Unit root tests for panel data”, Journal of International Money and Finance, Vol. 20/2, pp. 249-272, http://dx.doi.org/10.1016/S0261-5606(00)00048-6.
[22] Coderoni, S. and R. Esposti (2014), “Is There a Long-Term Relationship Between Agricultural GHG Emissions and Productivity Growth? A Dynamic Panel Data Approach”, Environmental and Resource Economics, Vol. 58/2, pp. 273-302, http://dx.doi.org/10.1007/s10640-013-9703-6.
[19] Defra (2018), Code of Good Agricultural Practice (COGAP) for Reducing Ammonia Emissions, http://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/ (accessed on 22 January 2019).
[2] Edenhofer, O. et al. (eds.) (2014), Agriculture, forestry and other land use (AFOLU), Cambridge University Press.
[14] FAOSTAT (2018), Value of Agricultural Production, http://www.fao.org/faostat/en/#data/QV/metadata.
[20] Foley, J. et al. (2011), “Solutions for a cultivated planet”, Nature, Vol. 478, pp. 337–342, http://dx.doi.org/10.1038/nature10452.
[26] Henderson, B. and J. Lankoski (2019), Evaluating the environmental impact of agicultural policies, http://dx.doi.org/10.1787/add0f27c-en.
[42] Im, K., M. Pesaran and Y. Shin (2003), “Testing for unit roots in heterogeneous panels”, Journal of Econometrics, Vol. 115/1, pp. 53-74, http://dx.doi.org/10.1016/S0304-4076(03)00092-7.
[1] IPCC (2014), Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team, R.K. Pachauri and L.A. Meyer (eds.)].
[29] Kerr, S. et al. (2013), Tackling Agricultural Emissions Potential Leadership from a Small Country, http://www.motu.org.nz/publications/motu-notes (accessed on 21 January 2019).
[28] Lissaman, W., M. Casey and J. Rowarth (2013), Innovation and technology uptake on farm, New Zealand Grassland Association, https://www.grassland.org.nz/publications/nzgrassland_publication_2523.pdf (accessed on 24 January 2019).
[33] Ministry for Primary Industries (2018), Report of the Biological Emissions Reference Group, https://www.mpi.govt.nz/.
[25] Ministry for the Environment (2018), New Zealand’s Greenhouse Gas Inventory 1990–2016, Ministry for the Environment , Wellington, New Zealand, http://www.mfe.govt.nz/sites/default/files/media/Climate%20Change/National%20GHG%20Inventory%20Report%201990-2016-final.pdf (accessed on 17 January 2019).
[32] Ministry for the Environment (2017), New Zealand’s Seventh National Communication - Fulfilling reporting requirements under the United Nations Framework convention on Climate change and the Kyoto Protocol, http://dx.doi.org/www.mfe.govt.nz.
[27] Ministry for the Environment (2017), The Structure of Agricultural greenhouse gas research funding in New Zealand, https://www.nzagrc.org.nz/.
[35] Morán, M. et al. (2016), “Ammonia agriculture emissions: From EMEP to a high resolution inventory”, Atmospheric Pollution Research, Vol. 7/5, pp. 786-798, http://dx.doi.org/10.1016/j.apr.2016.04.001.
[7] Murgida, A. et al. (2014), Evaluación de impactos del cambio climático sobre la producción agrícola en la Argentina, CEPAL, https://repositorio.cepal.org/bitstream/handle/11362/37197/1/LCL3770_es.pdf (accessed on 6 June 2018).
[36] Nguyen Van, P. (2005), “Distribution Dynamics of CO2 Emissions”, Environmental & Resource Economics, Vol. 32/4, pp. 495-508, http://dx.doi.org/10.1007/s10640-005-7687-6.
[12] OECD (2018), Agricultural Policy Monitoring and Evaluation 2018, OECD Publishing, Paris, https://dx.doi.org/10.1787/agr_pol-2018-en.
[13] OECD (2018), Agri-environmental indicators, http://www.oecd.org/tad/sustainable-agriculture/agri-environmentalindicators.htm.
[23] OECD (2018), Chapter 17. New Zealand Support to agriculture, OECD, Paris, http://dx.doi.org/10.1787/agr-pcse-data-en.
[11] OECD (2018), Human Acceleration of the Nitrogen Cycle: Managing Risks and Uncertainty, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264307438-en.
[5] OECD (2015), The Economic Consequences of Climate Change, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264235410-en.
[4] OECD (2008), Environmental Performance of Agriculture in OECD countries since 1990, http://dx.doi.org/www.oecd.org/tad/env/indicators.
[37] Ordás Criado, C. (2008), “Temporal and Spatial Homogeneity in Air Pollutants Panel EKC Estimations Two Nonparametric Tests Applied to Spanish Provinces”, Environ Resource Econ, Vol. 40, pp. 265-283, http://dx.doi.org/10.1007/s10640-007-9152-1.
[24] Parliamentary Commissioner for the Environment (2016), Climate change and agriculture: understanding the biological greehouse gases, http://dx.doi.org/www.pce.parliament.nz.
[30] Payne, P., J. Turner and H. Percy (2018), A Review of the SLMACC Technology Transfer Projects, MPI, Wellington, http://www.mpi.govt.nz/news-and-resources/publications/.
[40] Pedroni, P. (1999), “Critical Values for Cointegration Tests in Heterogeneous Panels with Multiple Regressors”, Oxford Bulletin of Economics and Statistics, Vol. 61/s1, pp. 653-670, http://dx.doi.org/10.1111/1468-0084.0610s1653.
[39] Perman, R. and D. Stern (2003), “Evidence from panel unit root and cointegration tests that the Environmental Kuznets Curve does not exist”, Australian Journal of Agricultural and Resource Economics, Vol. 47/3, pp. 325-347, http://dx.doi.org/10.1111/1467-8489.00216.
[31] Saggar, S. et al. (2013), “Quantification of reductions in ammonia emissions from fertiliser urea and animal urine in grazed pastures with urease inhibitors for agriculture inventory: New Zealand as a case study”, Science of the Total Environment, Vol. 465, pp. 136-146, http://dx.doi.org/10.1016/j.scitotenv.2012.07.088.
[3] Tubiello, F. et al. (2013), “The FAOSTAT database of greenhouse gas emissions from agriculture”, Environmental Research Letters, Vol. 8, http://dx.doi.org/10.1088/1748-9326/8/1/015009.
[15] UNECE (2018), Protocol to Abate Acidification, Eutrophication and Ground-level Ozone, http://www.unece.org/env/lrtap/multi_h1.html.
[17] UNECE (2014), Guidance document on preventing and abating ammonia emissions from agricultural sources Summary, UNECE, http://www.unece.org/fileadmin/DAM/env/documents/2012/EB/ECE_EB.AIR_120_ENG.pdf (accessed on 23 January 2019).
[34] UNFCCC (2018), National Inventory Submissions 2018, https://unfccc.int/process-and-meetings/transparency-and-reporting/reporting-and-review-under-the-convention/greenhouse-gas-inventories-annex-i-parties/national-inventory-submissions-2018.
[16] United Nations (2013), 1999 Protocol to Abate Acidification, Eutrophication and Ground-level Ozone to the Convention on Long-range Transboundary Air Pollution, as amended on 4 May 2012.
[18] United Nations Economic Commission for Europe (2015), Framework Code for Good Agricultural Practice for Reducing Ammonia Emissions.
[21] USDA (2018), International Agricultural productivity, https://www.ers.usda.gov/data-products/international-agricultural-productivity.
[10] Xu, R. et al. (2019), “Global ammonia emissions from synthetic nitrogen fertilizer applications in agricultural systems: Empirical and process‐based estimates and uncertainty”, Global Change Biology, Vol. 25, pp. 314–326, https://doi.org/10.1111/gcb.14499.
[9] Zhang, Y. et al. (2010), “Agricultural ammonia emissions inventory and spatial distribution in the North China Plain”, Environmental Pollution, Vol. 158/2, pp. 490-501, http://dx.doi.org/10.1016/j.envpol.2009.08.033.
The data to create this indicator was obtained from the United Nations Framework Convention on Climate Change (UNFCCC) database on national inventory reports (NIR) (UNFCCC, 2018[34]) for OECD countries included in Annex I of the UNFCCC. For OECD countries not included in Annex 2.B, data were compiled directly by the OECD via a questionnaire. While the UNFCCC requires countries to use a standard reporting format (CRF) for tables to ensure robust and standardised reporting, estimates made by individual countries may vary depending on factors and methods used in their own calculations. In addition, assumptions made in agricultural GHG emission calculations simplify complex agricultural systems, thereby introducing uncertainty into the estimate of GHG emissions. Although the OECD questionnaire for its member countries not included in Annex I follows the CRF tables to facilitate the treatment of the responses, the same caveats apply.
The categories covered according to the IPCC nomenclature are: 3A-Enteric fermentation, 3B-Manure management, 3C-Rice cultivation, 3D-Agricultural soil, 3E-Prescribed burning of savannas, 3F- Field burning of agricultural residues, 3G-Liming, 3H-Urea application, 3I-Other carbon-containing fertilisers, 3J – Others.
This indicator measures agricultural emissions of greenhouse gases per agricultural gross production value. It helps to assess whether agricultural production value is decoupled with greenhouse gas emissions of the sector. Agricultural gross production value measures production in monetary terms at the farm gate level and it is calculated by multiplying gross production quantities by output prices at farm gate (FAOSTAT, 2018[14]). Since intermediate uses within the agricultural sector (seed and feed) have not been subtracted from production data, this value of production aggregate refers to the notion of “gross production” (FAOSTAT, 2018[14]). It is important to recognise that distortionary policies such as market price support may affect the gross production value because it is measured at the farm gate level. A more appropriate measure of value would use non-distorted international prices; however, no such dataset is available at the global level.
Ammonia emissions for OECD countries were obtained from data officially submitted by the Parties to the Convention on Long Range Transboundary Air Pollution (CLRTAP) to the European Monitoring and Evaluation Programme (EMEP) programme via the United Nations Economic Commission for Europe (UNECE). Emissions reported under the CLRTAP tend to follow a bottom-up approach: they are calculated by applying emissions factors to geo-localised farm activities (Morán et al., 2016[35]). While reporting under the CLRTAP ensures standardised formats and facilitates consistency, there could be differences in terms of emissions factors and methodologies used across countries. Moreover, emissions are known to vary through the year and a national figure can mask spatial heterogeneity within countries (OECD, 2018[11]).
This annex provides further details on the empirical analysis of labour productivity and GHG emissions intensities. Table 2.B.1 shows descriptive statistics of the data used, which includes 33 countries in the period 1990-2015.
Sources: Agricultural gross production value was obtained from FAOSTAT (2018[14]), labour data were obtained from USDA (2018[21]) and GHG emissions data from the OECD Agri-environmental Indicators Database (OECD, 2018[13]).
Figure 2.5 was estimated using non-parametric methods. Non-parametric methods are suitable for this analysis because they do not assume a particular shape of the relationship between the outcome and the covariates (Nguyen Van, 2005[36]; Ordás Criado, 2008[37]). The method consists on running a number of local regressions at different values of the covariates with an optimal bandwidth. The density of the outcome is estimated by using the Epanechnikov Kernel function. A rule-of-thumb estimator selects the optimal bandwidth. Only two variables are used, agricultural labour productivity versus emission intensities (GHG, CH4 and N2O), to create the graphs in Figure 2.5.
The parametric model is as follows:
,
,
This model requires that variables are stationary or at least cointegrated so that the relationship obtained in the parametric regression is not merely spurious. First, panel unit root tests are conducted to test for stationarity (Choi, 2001[38]; Perman and Stern, 2003[39]; Coderoni and Esposti, 2014[22]). Three variables – namely labour productivity, GHG emission intensity, and CH4 emission intensity– do not reject the null hypothesis of containing unit roots (hence are not stationary) even at 10% level; they become stationary when first differences are taken (Table 2.B.2).
Notes: Fisher type augmented Dickey-Fuller tests (F-ADF) are conducted. Null hypothesis is containing unit roots in all panels and the alternative is at least one individual in the panel is stationary. We do not include a trend and one lag is used in the ADF regressions. In addition, the Im-Pesaran-Shin test is performed (Im, Pesaran and Shin, 2003[40]) and produces similar results as F-ADF tests.
Provided those three variables have unit roots in levels, cointegration tests are then performed to check for long-term relationships (Pedroni, 1999[40]). According to the results in Table 2.B.3, all the tests, except group ρ, are significant at the 5% level for both GHG and CH4 emission intensities. Hence, there exists a cointegrating relationship between GHG and CH4 emission intensities with first, second and third power of labour productivity.
Notes: Data for 1996 and 2000 of Israel are excluded, as cointegration tests do not allow gaps. All statistics are distributed as N(0,1). Rejecting the null of no cointegration is one-sided. Panel v is non-parametric variance ratio statistic, ρ is non-parametric test statistic, t and ADF (augmented Dickey-Fuller) are parametric statistic. Time dummies included.
Given these results, the preferred model is a dynamic model as it can capture the processes of adjustment to the long-run equilibrium as indicated by the results of the cointegration test. The estimated model is the Arellano-Bond one-step GMM estimation (Arellano and Bond, 1991[41]). The dynamic model generally includes lagged dependent variables as explanatory variable as follows:
,
where variables are indexed over the types of emission , country , and year t. The dependent variable is the log of emission intensity. The independent variables are the log of agricultural labour productivity and its squared and cubed terms. is the intercept. is the error term composed of a panel-level effects component () and an error term i.d.d. over the whole sample () and is the maximum length of lag. A trend () and year dummies () have also been included.
Arellano-Bond GMM uses instruments to deal with endogeneity between the lag of the dependent variable and the error term. We perform the one-step GMM estimation which assumes homoscedasticity on the disturbance term . For model specification, AR(2) test for serial correlation and Sargan test for over-identification.
Since autocorrelation of order 2 was not ruled out, for robustness check, results from a static random-effects model are displayed in Table 2.B.4. Results support the nonlinear and negative relationship between labour productivity and emissions intensities.
Notes: All variables were transformed into logarithms. Coefficients were estimated using a random effect model and robust standard errors are reported in parentheses. *, ** and *** represent statistically significant coefficients at the 1%, 5% and 10% levels, respectively. Year dummies are included.
Sources: Gross Production Value was obtained from FAOSTAT (2018[14]) and agricultural labour was obtained from USDA (2018[21]). GHG emissions data come from the OECD agri-environmental indicators database (OECD, 2018[13]).
← 1. See Annex 2.A for a detailed description of the method used for the estimation.
← 2. An alternative productivity indicator is the total factor productivity index (TFP) that is produced by USDA. The TFP index measures agricultural productivity in relation to a baseline year, so its interpretation is not straightforward and comparisons between the levels of different countries are meaningless. Another drawback for its use in this setting is that it includes variable inputs such as fertiliser and feed which are subject to short-term drivers such as weather and market shocks that are not necessarily relevant to the structural transformation of agriculture (Coderoni and Esposti, 2014[22]). Moreover, the correlation between TFP and labour productivity in our dataset is relatively large (0.7), indicating that although labour productivity may be a partial measure of productivity, it is a good proxy for total factor productivity.
← 3. Australia, Austria, Belgium-Luxembourg (joint due to lack of data availability), Canada, Chile, Czech Republic-Slovakia (joint due to data availability), Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Israel, Italy, Japan, Korea, Latvia, Lithuania, Mexico, Netherlands, New Zealand, Norway, Poland, Portugal, Spain, Sweden, Switzerland, Turkey, United Kingdom, and the United States.
← 4. See Annex 2.A for a detailed description of the method used for the estimation.