Carbon flux and N- and M-shaped environmental Kuznets curves: evidence from international land use change

ABSTRACT

Economic growth can affect land use change to release or sequester carbon, intensifying or mitigating the impact of other carbon emissions, and the functional form of that relationship is important to crafting policy responses. Data on land use and land cover change (LULCC) for 14 countries reveal an N – or M-shaped environmental Kuznets curve (EKC) for LULCC carbon flux to/from the atmosphere in some nations, while others display very different relationships. Most nations studied show some variation of the inverted-U EKC. All but one nation display initial turning points ranging from $2,000 to $9,000 per capita GDP (2011 dollars), and half are now net negative carbon emitters with respect to LULCC. For the US, regression analysis of the LULCC EKC indicates a roughly M-shaped quartic EKC function, with local maxima at about $3,700 and $45,700 and a local minimum at about $29,400. Where N-shaped EKCs are observed, the carbon sequestration from increasing forest regrowth is transient, and may be followed by a phase in which rising aggregate emissions dominate slowing sequestration in maturing forests. An M-shaped EKC indicates a third turning point, representing a return to increased net carbon absorption.

KEYWORDS:

• EKC
• carbon emissions
• carbon sequestration
• economic growth

Acknowledgements

The author thanks the Mises Institute for general research support, Tim Bersak, Wesley Downs, Jody Lipford, Joseph Spivey, and an anonymous referee for valuable remarks, and Richard Houghton for sharing data. All errors and omissions are of course the author’s own. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 In some of the literature, this is LULUCF (Land Use, Land Use Change, and Forestry).

2 Most carbon on land is contained in soil (Allmaras et al. 2000). Gebhart et al. (1994) estimated that cultivated croplands in the US account for about 2.7 Tg of carbon emissions per year, out of about 38.1 Tg emitted by U.S. agriculture (including agricultural fossil fuel usage and nitrogen fertilizer production). See also Post and Kwon (2000) on agricultural contributions to carbon loss.

3 More broadly, carbon efficiency (the change in total CO₂ emissions divided by the change in per capita RGDP) tends to increase with economic development; see, e.g., Lipford and Yandle (2010, 434).

4 Data from the National Interagency Fire Center show a decline in acres consumed by wildfire from 1926-2018, and an increase in federal fire suppression expenses from 1985–2018 (National Interagency Fire Center 2018a, 2018b). Pacala et al. (2001) point out that the annual burn area of the United States has decreased by 95 percent since 1850, so that carbon sequestration by woody plants has increased. See McCormick (2004, 180).

5 Houghton and Nassikas (2017) employ a bookkeeping model that tracks carbon sources and sinks, accounting for all changes in carbon in four pools: “living aboveground and belowground biomass; dead biomass, including coarse woody debris; harvested wood products; and soil organic carbon” for twenty types of ecosystems. This involved a two-step process: calculating areas of land use over time, and calculating carbon densities. Effects of environmental change, such as changes in climate, are not considered, but only fluxes “attributable to the direct anthropogenic effects of LULCC,” a useful exclusion for the EKC relationship we are considering here. The 2017 update included several changes:

1. LULCC data were country-specific rather than regional as in earlier work.

2. Rates of LULCC and emissions employed rates of deforestation and reforestation from the UN’s Forest Resources Assessment (FRA) and annual changes in croplands and pastures from Food and Agriculture Organization (FAO).

3. Regional definitions were changed.

4. The “emissions of carbon from draining and burning of peatlands for oil palm plantations in SE Asia” were added.

6 Roots of the derivative of the polynomial estimated in Table 2 are 3,744.95, 29,408.2, and 45,693.

7 This study used US GDP per capita data from Bolt et al. (2018) which is in 2011 dollars. FRED real GDP data for the US is in 2012 dollars (U.S. BEA 2019a), so a conversion factor of 0.9812 obtained from the GDP deflator (U.S. BEA 2019b) was used to convert to 2018’s real GDP per capita ($56,921) to 2011 dollars for this calculation. Nominal GDP for 2018 was $62,853 per capita.

8 EDGAR data for 2016 indicate these 14 countries account for about 21 million of the world’s 36 million kilotons of CO₂ emissions from fossil fuel use and industrial processes. Biomass burning and LULCC are excluded. See Janssens-Maenhout et al. (2017). These nations account for about 54 percent of the flux (in absolute value) from LULCC.