The Azimuth Project
Economic growth and labour


This page is about economic growth Economic growth and its effects on employment, wages, health and education and environment.

economic growth, labour and productivity

Currently, the most exhaustive collection on labour statistics may be this:

In particular, the so-called Key Indicator of the Labour Market (KILM) Laborsta provides a tool for assessing the data they gather:

This tool is still in development though. Note that we use here the older 6th edition, which contains the socalled elasticity indicator. Let Lgrowth{Lgrowth} denote the growth in employment, and GDPgrowth{GDPgrowth} the growth of GDP, then following the verbal explanations here:

the elasticity seems to be given by Lgrowth/GDPgrowth{Lgrowth}/{GDPgrowth} where a possible approximation is:

(1)GDPgrowth(year)=GDP(year)GDP(year1)GDP(year1), GDPgrowth(year) = \frac{GDP(year)-GDP(year-1)}{GDP(year-1)},

analogously for employment ( Note: The above formula is an approximation to a smooth curve, the KILM may have used a slightly different formula. ).

For a mathematical less interested audience let’s make a small example which gives a very simplified and partial comparision what exponential growth means here. Lets assume for simplicity that wages were constant in the labour market, so that the value of the labour market was mainly adjusted by its size. Assume a very rich person quadruples his income every year (i.e. multiply with 4) then

IncomeGrowthRich(year) = Income(year)Income(year1)Income(year1) =4Income(year1)Income(year1)Income(year1) =3 \begin{aligned} IncomeGrowthRich(year) &=& \frac{Income(year)-Income(year-1)}{Income(year-1)} \\ =\frac{4 \cdot Income(year-1)-Income(year-1)}{Income(year-1)} \\ &= 3 \end{aligned}

(compare with the formula for the GDPgrowth.) Assume that the rich man doubles the wage of his house maid

IncomeGrowthMaid(year) = Income(year)Income(year1)Income(year1) = 2Income(year1)Income(year1)Income(year1) =1 \begin{aligned} IncomeGrowthMaid(year) &=& \frac{Income(year)-Income(year-1)}{Income(year-1)} \\ &=& \frac{2 \cdot Income(year-1)-Income(year-1)}{Income(year-1)} \\ &= 1 \end{aligned}

so that the ratio of IncomeGrowthMaid versus IncomeGrowthRich is 1/30.31/3 \simeq 0.3 (which is about the average number given by KILM for the global elasticity) Note that the actual growth of GDP is of course rather 3\% than 3, so instead of 1 year it would take rather about 46 years (artifacts of the approximation bluntly taken aside) to quadruple the GDP, i.e. 46ln4/0.0346 \simeq \ln{4}/0.03 (likewise the labour will grow at a factor exp0.01*461.62\exp{0.01*46} \simeq 1.6 \simeq 2 in 46 years). However the principal arguments stay the same, and may appear clearer in this scale. (People are often scared by percentages.) Then the above computation should illustrate that the gap between the income of the maid and the income of the rich person will increase dramatically.

On page 5, table 19b the worldwide elasticity is since 1992 at about 0.3 with even a slight trend of decline (see text to the table). That means that on average the growth in employment is about one third smaller than economic growth. In fact, for East Asia the elasticity is only 0.10.1: while this region had a GDP growth of about 89%8-9\%, the growth in employment was only 0.80.9%0.8-0.9\%.

One can also see this trend if one compares the productivity increase at KILM. Following KILM, for Germany alone the GDP per hour worked increased from 102 points in 1992 to 112 in 1996 and 133 in 2008. This means that economic value is going much less into labour development than into other sectors. Moreover, given the above data it is to be expected that with no or a very small economic growth the job sector would even be in decline (negative elasticity). Unfortunately the KILM doesn’t yet provide elasticities for all countries, so we cannot confirm this claim.

update JUly 9th 2014: Since the beginning of the year all above linked documents of ILO had been taken off the web. Search engine caches are though still holding the kilm19EN-2009.pdf. In particular the ILO claim that worldwide employment was rising can’t be anymore recovered. Given an average GDP of roughly 3 percent then if one follows the kilm19EN-2009.pdf document which claimed that the elasticities where on average 0.3 this would mean that there is on average a global employment point growth of roughly 0.9. A document which was recently still available at the old database, which held the global employment-to-population ratio by sex and age showed however that this ratio is rather constant or especially for the young age groups in decline. The same can however be also established by looking at hte new database at ILOstat. That is apart from exceptions as Hungary and Italy most employment ratios are in decline. There is though not much data available for big parts of Asia, Africa and South America. Please see: Employment-to-population ratio by sex and age (%) By looking at this ratio it thus is not really clear how the ILO claimed employment point growth of 0.9 could be motivated. In particular things are way worse, if one looks at that ratio. In that context it would however be important to establish how much the rather rapid decline in youth employment is partially due to longer education.

update july 24, 2014: The global employment-to-population ratio can now be seen in a diagram. ***********

It would be interesting to assess where the rest of the produced wealth went. As pointed out here:

social conditions such as health and educational conditions do not necessarily improve with economic growth:

One of the most surprising results of human development research in recent years, confirmed in this Report, is the lack of a significant correlation between economic growth and improvements in health and education. Our research shows that this relationship is particularly weak at low and medium levels of the HDI.

Likewise, as the above shows, the improvements in labor development are way smaller than economic development. Moreover wages in the manufacturing sector rise if at all only moderately. Unfortunately the KILM has not yet an automated world index, so let’s look at the example of Germany. The real manufacturing wage index was in Germany in 1996 at 97.6 points, and in 2006 at 100.7 points. The biggest wage jump of 1.5 points was between the years 2002/2003 which gives a growth of 1.5/100.20.0151.5/100.2 \simeq 0.015 which is 1.5%1.5\% in those years. In some other years, however, there was even a decline in wage, despite the above-mentioned large increase in productivity. Likewise the employment-to-population ratio stayed about constant (Germany, 1992: 55.0%55.0\%, 2008: 55.3%55.3\%).

As a comparision: in China the biggest wage jump was between 2006/2007. The index was 189.2 points in 2006 and 209.8 points in 2007, which gives 20.6/189.20.1120.6/189.2 \simeq 0.11, i.e. about 11%11\%. However the wages in previous years grew 6%6\% on average.

economic growth, labour and energy

Growth in productivity may in partial be due to the replacement of human labour by machines. For assessing the possible scope of such a replacement the comparision of economic growth versus energy use is a rather strong indicator. The below is a diagram, which is here a citation from the website of the Federal Ministry for the Environment, Nature Conservation and Nuclear Safety a similar diagram can be e.g. found in Umweltökonomische Gesamtrechnungen 2012 (page 5), which is a publication by the Federal Statistical Office of Germany. It displays how the GDP (green curve) correlates with the use of energy (red curve). The blue curve displays the socalled energy productivity, which is computed from the red and green curve.


Around the year 2009 a sudden drop in GDP, as well as in energy use can be observed (it is actually better visible in the diagrams in the above cited papers), which indicates that the GDP is rather strongly correlated with the use of energy. A similar diagram on wikipedia for Japan displays the correlation even better, however this diagram should be taken with caution as the cited data sources hold only data starting in 1975. Eventually the later supplemented grid to that diagram with the year numbers is erranous. The author will eventually later try to display (west) german data before 1990. The japanese diagram should be reworked and diagrams of other industrial nations should be supplemented as well.

Nevertheless the above given diagrams display quite a strong correlation between GDP and energy use and thus strongly point to an increased use of machines instead of human labour.

That is the population of Germany and Japan stayed roughly constant in the given time, so that a higher use of energy may more or less be mostly due to a higher consumption/use of energy intensive products (i.e. mostly machines) and/or higher energy needs of production. (Remark: For simplicity we call here buildings etc. also "machines") Moreover if one assumes that there is a certain rate of the implementation of higher energy efficiency then the use of energy will be damped by this energy efficiency. It is however to be suspected that the implementation of energy efficiency is mostly on a longer time scale then the ultimate use of energy. That is a more energy efficient machine runs for many years and doesn’t change it’s efficiency if the economy is running bad for a year and a bad economy will have mostly only a delayed impact on efficiency implementations etc. So by regarding the correlation of little “bumps” in the diagrams one sees the interplay of GDP, use of energy and energy efficiency.

A welltempered environment for working may enhance human productivity and this environment may need additional energy (heating/air condition, food), however it is to be expected that this aspect is not constituting a too big component in the investigation of the correlation between GDP and energy use. It is more to be expected that a rather sudden decline in GDP leads to less consumption/use of energy intensive products (i.e. mostly machines) and less use of machines for production. Or in other words in economic bad times one usually hears that companies “have to stop their assembly lines” and that “people make holidays at home” (use less airplanes) etc. and the diagrams give somewhat a rather visible quantification of this fact or in other words the sudden drop of GDP and the sudden decline of energy use around 2009 is visible in the “bumps” in the curves around the year 2009.

Remark: These facts are also very visible in the diagrams 11a (G'‘utertransportintensität) in the german article ‘’Umwelt'’okonomische Gesamtrechnungenwhich displays a sudden decline in the transport of goods in Germany around the year 2009 (the year of a sudden decline in GDP) and in diagram 11b (Personentransportintensität) which displays for that time a sudden increase of the mobility of persons within Germany together with only a very small increase (in relation to the increase of mobility) of energy use forthat purpose, i.e. people made more use of transport within Germany during times of bad economy and the means of transport must have been rather energy efficient, since this additional transport used up -relatively seen- not much energy \url{}