A global compass for the great divergence: Emissions versus production centres of gravity 1820–2008

• Author: Jean‐Marie Grether,Nicole A. Mathys,Caspar Sauter
• Date: 01 October 2019
• Published date: 01 October 2019
• DOI: http://doi.org/10.1111/twec.12860

2818

|

World Econ. 2019;42:2818–2834.

wileyonlinelibrary.com/journal/twec

Received: 3 September 2017

|

Revised: 30 July 2019

|

Accepted: 15 August 2019

DOI: 10.1111/twec.12860

ORIGINAL ARTICLE

A global compass for the great divergence:

Emissions versus production centres of gravity

1820–2008

CasparSauter1

|

Jean‐MarieGrether2

|

Nicole A.Mathys3

1UBS SA, Zurich, Switzerland

2University of Neuchatel, Neuchatel, Switzerland

3Federal Office for Spatial Development, Bern, Switzerland

Funding information

Swiss National Science Foundation, Grant/Award Number: Grant 100018‐136625

KEYWORDS

center of gravity, GDP and CO2 emissions, great divergence

1

|

INTRODUCTION

The need for global indicators to address large‐scale socioeconomic issues is today more acute than

ever. Human societies have become so opulent, ubiquitous and interconnected that many of the chal-

lenges they face, from climate change to security, find their roots and/or unravel consequences all

around the planet. In spite of this globalisation of economic problems, most measurement tools used

by analysts and policymakers remain national in scope (e.g. Lindmark, 2004), or simply aggregated

at the world level, without considering the geographical dimension except in specific cases (e.g. in-

dicators of remoteness, Baier & Bergstrand, 2009 or geopolitical importance, Reynaud & Vauday,

2009, but which are country‐specific in nature). In this paper, we propose to revisit and refine the

concept of the world centre of gravity, which encapsulates into a single point the distribution of any

variable upon the earth's surface. We apply it to human population, gross domestic product (GDP)

and anthropogenic carbon dioxide (CO2) emissions, uncovering crucial but not trivial trends across

the 1820–2008 period.

Early applications of the centre of gravity (see Grether & Mathys, 2010; Quah, 2011) focused on

GDP and recent decades. They unveiled a clear Eastern shift of the economic centre of gravity since

1980. Two subsequent papers (Grether & Mathys, 2011; Grether, Mathys, & Lutzelschwab, 2012)

extended the time period backwards, relying on the Maddison (2010) database. They identified a

strong Western shift of the economic centre of gravity during the nineteenth century and a trend

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction

in any medium, provided the original work is properly cited.

© 2019 The Authors. The World Economy published by John Wiley & Sons Ltd.

|

2819

SAUTER ET Al.

reversal towards the east in 1950. Although informative, these early studies are subject to three type

of limitations: first, the reported evidence so far is limited in its scope and its contribution to the

understanding of world challenges; second, the degree of accuracy depends on the availability and

reliability of gridded data on the earth's surface for the relevant variables; third, the centre of gravity

itself is a point beneath the earth's surface, which is not entirely intuitive and leads to distortions

when its position is projected upon a two‐dimensional map. The objective of the present paper is

to provide an appropriate treatment of each of these caveats, thereby illustrating the usefulness of

the approach.

The inclusion of CO2 emissions, along with GDP and population, allows for deeper insights

into the period of the so‐called ‘Great Divergence’. Following Zhao, Stough, and Li (2003), we use

the demographic centre of gravity as a benchmark and construct simple measures of spatial im-

balances to characterise the divergence between world GDP (or emissions) and world population.

As could be expected, the two indices follow a similar inverted‐u pattern over the centuries of the

Great Divergence, but with two important differences: (a) the starting level of spatial imbalances

for emissions is considerably larger than for production and (b) the trend reversal occurs 30years

earlier for emissions (1920) than for GDP (1950). This illustrates the historic responsibility of the

west, a cornerstone of the present negotiations to tackle climate change (e.g. Barrett & Stavins,

2003 or Mattoo & Subramanian, 2012). It also proves how deeply associated are energy use and

transition with the economic divergence or convergence processes (e.g. Bradshaw, 2014). Finally,

it suggests that the industrial revolution was already full steam ahead when it began to materialise

into significant shifts in economic power. This is in line with recent advances in economic history

pointing to early roots of the process, perhaps as far back as the sixteenth century (e.g. Broadberry,

2013 or Studer, 2015).

An important limitation of all previous studies is that, for all years for which gridded data are

unavailable, the assumption is simply that grid shares at country level are unchanged with respect to

the closest available year (e.g. 1990 for the papers based on the G‐Econ database, see Nordhaus et al.,

2006). This is of particular concern for countries like the United States or China, which cover large

areas, represent a significant share of world totals, and where the distribution of people and economic

activity has seen structural changes over the last two centuries. The present paper offers an improve-

ment with respect to that shortcoming, by exploiting the Hyde 3.1 database (Klein Goldewijk, Beusen,

Drecht, & Vos, 2011), which provides gridded population data at a very disaggregated level. This goes

back as far as 1750 and has already been exploited by long run studies of land use by human popu-

lations (Ellis et al., 2013) and its relationship with global warming (Matthews et al., 2014). It allows

to spread national totals regarding GDP (or CO2 emissions) according to varying population shares

rather than by applying fixed shares.

Although the centre of gravity ought to be seen as the tip of a global arrow originating from the

centre of the earth, its representation on a two‐dimensional map is subject to distortions. The method-

ological section provides a thorough discussion of distortions relying on the two projection methods

used so far (i.e. the orthogonal projection on the earth's surface or onto a cylinder wrapping the earth

along the Equator). We then propose a new technique, which is distortion‐free, and consists of using

two maps instead of one to represent the three Cartesian coordinates of the centre of gravity. This

allows more precise tracking of the centre of gravity.

In short, the centre of gravity behaves like a global compass, with its length and direction depend-

ing on the spatial distribution of the corresponding variable upon the earth's surface. Its calculation

and comparisons across key socioeconomic indicators allows unveiling global shifts and spatial un-

balances over the sample period. This may justify further applications of the concept in a number of

different domains, as suggested in the conclusion.