The Third Revolution
Paul Harrison

Appendix: Assessing Population Impact on Environment



Most of the debate on population growth's impact on the environment, or lack of it, has been conducted in an unscientific manner, in a vacuum of evidence, deprived of the oxygen of hard data.

But things become much clearer when they are looked at quantitatively, and in terms of physical mechanisms. Then population tends to take its proper place as one of the key direct factors.

  Paul Ehrlich's Formula: 1.

       Impact = Population x Affluence x Technology.

is often used to discuss the connections. It has the appearance of mathematical rigour, but it does not help towards greater precision. The term `affluence' demands a value judgement about what is `affluent' and what is not. It implies that subsistence consumption has no impact on the environment, when in fact all levels of consumption, even of hunter-gatherers, have some impact.

For more precision, `affluence' needs replacing with `consumption per person'. This covers everything from eating wild roots or building a palm leaf shelter, to walling your bathroom in marble or driving a Porsche at 200 kilometres an hour down the autobahn. The term `technology', too, is not capable of measurement as such and needs replacing, with `environmental impact per unit of consumption.'

Thus we have Formula 2.

Environmental impact =

Population x Consumption per person x Impact per unit of consumption.

Impact can refer to any one of our three main forms of interaction with the environment: our use of primary resources from minerals to water and land; our physical occupation of space; or our output of pollutants.

The formula, which is a rephrased version of Barry Commoner's can be used in many different situations. It can be applied by, or to, high-consuming individuals, to give them some idea of the environmental cost of their lifestyles. It helps to concretize what so often is invisible, and I have tried to do this with the idea of the personal waste mausoleum and the carbon balloon.

It can also be applied at a particular moment in time to countries or regions. The formula makes it clear that, for any given level of consumption per person or technology, higher population means higher environmental impact. Fewer people mean lower total consumption and waste.

To assess what contribution population makes to increases in environmental damage - to try to assign relative blame - we have to look at changes over time. What we need is some measure of the share of each of our three elements in overall impact. If all three are pushing upwards, we can simply express the change in each one in turn as a percentage of the total change. So here:

Formula 3. Population impact =

annual % change in population x 100
annual % change in use of resource or output of pollutant

Sometimes one or more of the three factors may be tending to reduce environmental impact. We can distinguish between upward and downward pressures by `scoring' the upward pressures out of +100 per cent, and the downward pressures out of -100 per cent.

It might help at this point to look at some concrete examples. Take the case of the expansion of agricultural land. This is not merely physical occupation of space and use of the land resource. It is a prime source of environmental damage, cutting down forests, draining wetlands, ploughing up natural grassland, replacing natural ecosystems of high species diversity with artificial ones of low diversity.

Let us see if we can discover what share in the increase in farmland we can attribute to population growth (see Table 1). Between 1961 and 1985, population expanded by 2.3 per cent a year in developing countries. Agricultural production rose by 3.3 per cent. So production per person, which we can take as our consumption factor, rose by 0.9 per cent. Farmland, meanwhile, expanded by only 0.6 per cent a year. The farmland used per unit of agricultural production - the impact per unit of consumption - actually declined by 2.6 per cent annually, because yields were increasing. Technology in this case exerted a downward pressure on environmental impact, at least in terms of land used. Only population and consumption exerted upward pressure.

In this case we can say that population growth accounted for +72 per cent of the growth of farmland, and increase in consumption per person for +28 per cent. Technology gets a score of -100 per cent, since it was the only one of the three factors pushing towards lower use of farmland.

In developed countries, with slower population growth and faster increase in agricultural production per person, the contributions work out quite differently: here population growth accounted for only 46 per cent of the expansion in farmland, increased consumption per person for 54 per cent. Technology again scores -100 per cent of the downward factors.

Another example. Increase in livestock numbers is a factor in several types of environmental damage. Goats eat tree seedlings. Cows compact soil and vent methane. So let's do the sum again for developing countries between 1961 and 1985 (see Table 1). Meat and milk consumption grew by 3.2 per cent a year overall, per capita consumption (given the 2.3 per cent growth in population) by about 1 per cent. The number of livestock, however, rose more slowly than meat and milk production, by only 1.3 per cent a year. So the number of livestock required for each unit of meat and milk dropped by 2 per cent a year, due to technology changes such as better veterinary treatment, better breeds of animal and so on. Once again, technology was responsible for a net downward pressure on livestock numbers and so scores -100 per cent. Population accounted for 69 per cent of the upward pressure on numbers, and consumption for the remaining 31 per cent. (See Table 1 below)

Increasing fertilizer use alters soil composition, contributes to emissions of nitrous oxide which stokes the greenhouse effect, and pollutes rivers, lakes and coastlines. Here we are dealing with a massive and rapid technological change in agriculture. The population impact is correspondingly lower (see Table 2). It accounted for only 22 per cent of the increase in developing countries. Increased consumption - measured by the growth in agricultural production per person - accounted for only 8 per cent. The technology factor - increased fertilizer per unit of agricultural production - accounted for 70 per cent of the increase.

In many cases, as Commoner points out, we don't know the actual damage attributable to each unit of consumption. So we end up measuring the output of pollutants per unit of consumption. Take the output of air pollutants linked with energy consumption in OECD countries (see Table 3 below). In this case I have taken final consumption expenditure as the consumption measure. The data here allow us to break down the technology factor into two components: change in energy efficiency - energy requirements per unit of final consumption; and change in the amount of pollutant emitted for each unit of energy used, which has more to do with the mix of energy sources and the technology of emission cleaning.

Using the same methods, we find that even in the slow growing OECD countries, population was responsible for +25 per cent of the upward pressure on emission of air pollutants, and increased consumption for +75 per cent. But the results varied considerably between countries. In the USA, where population growth was maintained by immigration,, the population factor amounted to +30 per cent of the upward pressure on emissions. In West Germany, where population grew by only 1.3 per cent over the whole eighteen year period, the population impact was only +2.4 per cent.

Technology was a downward pressure. In the case of sulphur dioxide and smoke, total emissions actually fell, respectively by 2.6 per cent and 4 per cent per year. Technological change aimed specifically at reducing pollution per unit of energy consumed (shown as Tp in the table below) appears to have had about twice as much downward impact as the general increase in energy efficiency (shown as Te).

The results for two other pollutants, oxides of nitrogen and carbon dioxide, were quite different. These increased, by 0.7 and 0.8 per cent respectively, though not as fast as the increase in either population or consumption. So once again technology was a downward pressure, though much weaker than in the case of sulphur dioxide and smoke. But in these cases the impact of reduction in emissions per unit of energy used was two to two and a half times less than the increase in energy efficiency.

There are some cases where it is not feasible to measure separately the consumption and technology factors. For example, in the case of expansion of irrigation, it may be difficult or impossible to discover exactly how much foodstuff was produced on irrigated as opposed to non-irrigated land. In the case of chlorofluorocarbons, it is hard to isolate the consumption element, since this would mean aggregating incommensurables like refrigeration, air conditioning, packaging and solvents. There are other cases where consumption and technology are not easily separable: new types of product, for example, such as fridges, video recorders, or computer games, which are not merely providing an old need in a new way (as cars do for transport) but where the product itself creates the consumer `need' which it supplies.

In many of these tricky cases it is still often possible to get some idea of the population contribution, using formula three above. In cases where all three factors are pushing in the same direction, this formula will give the same results as formula two.

When one or other of the three elements are declining, and consumption and technology measures are absent, one can still get some idea of the impact of population growth, compared with the other two elements, using the formula:


Formula 4. Population impact =

absolute value of population growth ratea x 100
absolute [pop growth] + absolute [growth in pollutant - population growth]


This last formula, however, does not assess population's share of the change. Instead it measures the relative strength of the population influence compared to consumption-plus-technology considered together, regardless of whether the influences are positive or negative.

One other caution relates to the geographical level of analysis. Applying these formulae at global level can be misleading, and gives paradoxical results. Take, for example, increases in emissions of carbon dioxide from fossil fuel and cement making. Between 1960 and 1988 these increased in developing countries by 5 per cent a year, while population increased by 2.3 per cent a year. The population impact is therefore 46 per cent (2.3 x 100 ΒΈ 5). In developed countries, emissions grew by 2.41 per cent a year and populations by 0.85 per cent. The population impact is 35 per cent.

At global level we find that emissions rose by 3 per cent, while population rose by 1.9 per cent. The population impact therefore appears to be 63 per cent - much higher than the average of the two separate measures. This happens because the slower growth in the much higher level of emissions in the North results in a lower global figure for emission growth.

Hence global figures for population impact cannot be produced on the basis of global change figures in the three elements. It means that, as a minimum, results should be worked out separately for developing and developed countries. Ideally regional and country level results are needed. We can still arrive at global figures for population impact, by averaging the separate results. The averages can be weighted in various ways. Weighting the regional results by 1988 populations, this gives the outcome that population growth accounted for 44 per cent of the increase in carbon dioxide emissions. Weighted by 1988 emissions, the population impact is 42 per cent. Weighted by total increases between 1960 and 1988, the population impact is 41 per cent.

There is no finality about these methods or results. The subject is young, and fluid, and alternative approaches may emerge. What I hope they do show is the scope for developing detailed quantitative studies in which population, consumption and technology impacts on the environment can be assessed.

Table 1.



FARM AREA Annual % change 1961-1985   Impacts: % of change
    Farm area Pop Cons Tech   Pop Cons Tech  
Developing   0.57 2.3 0.9 -2.6   +72 +28 -100  
Developed   0.13 0.9 1.0 -1.8   +46 +54 -100  
LIVESTOCK Annual % change 1961-1985   Impacts: % of change
  #s of
Pop Cons Tech   Pop Cons Tech  
Developing 1.3 2.3 1.0 -2.0   +69 +31 -100  
Developed 0.3 0.9 0.6 -1.2   +59 +41 -100  


Farm area: area of arable and permanent crops.
P= Population increase.
C= Change in agricultural production per person.
T= Change in farm area per unit of agricultural production. This is the inverse of yield.

Livestock: includes total numbers of cattle, sheep and goats.
P= Population increase.
C= Change in meat and milk consumption per person.
T= Change in number of livestock needed for each unit of meat and milk consumption.

Data source (including population):
Food and Agriculture Organization, Country Tables 1990, FAO, Rome, 1990, pp. 320-21 and 334-5.

Table 2.


    Annual % change 1961-1985   Impacts: % of change  
REGION     Fert Pop Cons Tech   Pop Cons Tech
AFRICA     7.68 2.88 -0.71 5.41   34.7 -100.0 65.3
LATIN AMERICA   8.38 2.45 0.15 5.63   29.8 1.8 68.4  
NEAR EAST   9.74 2.80 0.12 6.62   29.4 1.2 69.4  
FAR EAST   11.14 2.34 0.60 7.95   21.5 5.5 73.1  
ASIAN CPE   13.24 1.90 2.43 8.49   14.9 18.9 66.2  
DEVELOPING   10.99 2.32 0.87 7.53   21.6 8.1 70.2  
DEVELOPED   4.26 0.87 0.75 2.59   20.7 17.7 61.5  
WORLD     5.90 1.90 0.51 3.40   [32.7]b 8.7 58.5

Global population impact:
Weighted by 1988 populations: 22.1 per cent.
Weighted by 1988 fertilizer use: 20.8 per cent.


P= Population increase C= Increase in agricultural production per person.
T= Increase in fertilizer use per unit of agricultural production.

Data source: Food and Agriculture Organization, Fertilizer Yearbook 1989, FAO, Rome, 1990 (for 1988 figures). 1961 data from FAO, Country Tables 1990, FAO, Rome, 1990.

Table 3.


  Annual % change 1961-1985       Impacts: % of change
  Emissions P C Te Tp   P% C% Te% Tp%
SO2 2.6   - 0.8 2.4 -1.7 -4.0   +25 +75 -30 -70
SMOKE -4.0   0.8 2.4 -1.7 -5.4   +25 +75 -24 -76
NOx 0.7   0.8 2.4 -1.7 -0.8   +25 +75 -69 -31
CO2 0.8   0.8 2.4 -1.7 -0.7   +25 +75 -72 -28


SO2 - sulphur dioxide Smoke - particulate matter NOx - nitrogen oxides CO2 - carbon dioxide

Emissions = change in emissions of gas.
P = population
C = change in private final consumer expenditure per person
Te = Energy technology - energy requirement per unit of consumer expenditure.
Tp = Pollution technology - emissions per unit of energy requirements.

All figures are percentages. The first five columns are the annual percentage change over the 1970-88 period. The last four columns represent the percentage contributions - positive (upward) or negative (downward) of each factor to the overall change in emissions.

Data source (including population):Organization for Economic Cooperation and Development, State of the Environment 1991, pp 236, 241, 243, OECD, Paris, 1991.

Table 4

From fossil fuels and cement only.

Region Annual % change 1965-1989 Impacts
  Emissions P C T P% C% T%
Low income 6.7 2.2 3 1.3 33.8 46.0 20.2
China 6.9 1.8 5.8 -0.7 23.7 76.3 -100.0
India 5.8 2.3 1.9 1.5 40.4 33.4 26.2
Mid income 4.5 2.2 2.2 0.0 50.0 50.0 0.0
LDCs (All) 5.4 2.2 2.5 0.6 41.7 47.4 10.8
S.S. Africa 7.0 2.9 0.2 3.8 42.1 2.9 55.0
E. Asia 7.2 1.9 5.3 -0.1 26.7 74.4 0.0
S. Asia 6.2 2.3 1.9 1.9 37.6 31.1 31.3
Near East 7.0 2.9 1.8 2.2 42.2 26.2 31.6
Lat. Amer. 4.2 2.3 1.7 0.1 55.9 41.3 2.8
Other econs 3.0 1.0 1.5 0.5 33.6 50.4 16.1
High Income 1.5 0.8 2.3 -1.6 25.8 74.2 -100.0
World 2.8 1.8 1.5 -0.5 [54.5 45.5 100]
World [weighted by emissions growth] 35.6 64.4 -100.0

P = Annual population growth
C = Annual growth in GNP per person
T = Annual change in CO2 emissions per unit of GNP
NB: Only the weighted world figure should be used.
This is the average of low, mid, high income & other economies res weighted by the absolute growth in CO2 emissions 1965-1989.

Data source: Social Development Indicators 1991-2, World Bank pp xiv-xvii, 374.


Formatted for the web by:
Committee for the National Institute for the Environment
1725 K Street, NW, Suite 212, Washington, D.C. 20006-1401

Phone (202) 530- 5810 [email protected] Fax (202) 628-4311