Mathematical Analysis of the Historical

Economic Growth
Ron W Nielsen1
Environmental Futures Research Institute, Gold Coast Campus, Griffith University, Qld,
4222, Australia
September, 2015
Historical economic growth is analysed using the method of reciprocal values.
Included in the analysis is the world and regional economic growth. The analysis
demonstrates that the natural tendency for the historical economic growth was to
increase hyperbolically.
Introduction
Historical economic growth is defined as the growth before it started to be diverted to
mathematically distinctly different trajectories, which, in contrast, could be described as
representing modern economic growth. This distinction will become clear when we analyse
the economic growth data (Maddison, 2010). We shall see that the transitions from the
historical to modern growth occurred between around the end of the 1800s and the mid1900s. The analysis of the modern growth will be discussed in separated publications.
The primary objective of the analysis presented here is to find the simplest mathematical
representations of the historical economic growth data. It is not a curve-fitting exercise. It is
not an attempt to fit the historical economic growth data by the incomprehensibly
complicated mathematical formulae just to reproduce their trajectories. Our aim is to find the
simplest and if possible the uniform mathematical representation of the economic growth data
because our ultimate aim is not just to describe the historical economic growth but to explain
it. We shall be looking for a common pattern of growth. In a separate publication we shall
propose mechanism to explain this common pattern. Complicated mathematical distributions
might be impressive if they fit the data perfectly well but it might be also difficult or maybe
even impossible to find a convincing explanation of the mechanism of growth represented by
such complicated mathematical distributions.
Our analysis will be based on using the method of reciprocal values (Nielsen, 2014). This
method is based on the observation (von Foerster, Mora & Amiot, 1960) that the growth of
human population during the AD era was increasing hyperbolically. Recent but limited
analysis (Nielsen, 2014) indicated that historical economic growth was also increasing
hyperbolically.
Hyperbolic distribution describing growth is represented by a reciprocal of a linear function:

S (t )

a kt

(1)

where S (t ) is the size of the growing entity, while a and k are positive constants.

AKA Jan Nurzynski, r.nielsen@griffith.edu.au; ronwnielsen@gmail.com;

http://home.iprimus.com.au/nielsens/ronnielsen.html

The reciprocal of such hyperbolic growth, S 1 (t ) , is represented by a decreasing linear

function:
S 1 (t )

a kt .

(2)

Hyperbolic distributions should not be confused with hyperbolic functions ( sinh(t ) , cosh(t ) ,
etc). Furthermore, in our notation, S 1 (t ) does not represent the inverse function of S (t ) but
its reciprocal, i.e. S 1 (t ) 1/ S (t ) .
Reciprocal values help in an easy and generally unique identification of hyperbolic growth
because in this representation hyperbolic growth is given by a decreasing straight line. Apart
from serving as an alternative way to analyse data, reciprocal values allow also for the
investigation of even small deviations from hyperbolic distributions because deviations from
a straight line can be easily noticed.
Reciprocal values allow also for an easy identification of different components of growth.
This property can be used, for instance, in the investigation of the validity of the Unified
Growth Theory (Galor, 2005, 2011), which is based on the postulate of different regimes of
growth.
When comparing mathematically-calculated distributions with the reciprocal values of data,
we have to remember that the sensitivity of the reciprocal values to small deviations increases
with the decreasing size S of the growing entity.
Suppose we have two values of S at a given time: S1 (t ) and S 2 (t ) , representing, for instance,
the empirical and the calculated values. It is clear that

S
S1S2

S
,
S2

(3)

S 1 increases rapidly with the decreasing S.

The separation of small values of data from calculated distributions are magnified.
where S is either S1 or S 2 . For a given

S ,

It should be also noted that the decreasing reciprocal values describe growth, while a
deviation to larger reciprocal values describes decline. Consequently, a diversion to a faster
trajectory will be indicated by a downward bending of a trajectory of the reciprocal values,
away from an earlier observed trajectory, while the diversion to a slower trajectory will be
indicated by an upward bending.
The data describing the historical economic growth (Maddison, 2001, 2010) do not allow for
a detailed analysis below AD 1500 because there are two large gaps in the data: between AD
1 and 1000 and between AD 1000 and 1500. The best sets of data are from AD 1500.
However, the compilation prepared by Magnuson appears to be the best source of the
historical growth data.
Throughout the analysis, the values of the Gross Domestic Product (GDP) will be expressed
in billions of the 1990 International Geary-Khamis dollars. Furthermore, in order to examine
the quality of fitted distributions to the small values of data, we shall display the GDP values
and their corresponding calculated curves using semilogarithmic scales of reference. All
diagrams are presented in the Appendix.

World economic growth

Results of mathematical analysis of the world economic growth are presented in Figures 1-3.
Reciprocal values of historical data can be fitted using straight line (representing hyperbolic
growth) between AD 1000 and 1955. From around 1955, the world economic growth started
to be diverted to a slower trajectory as indicated by the upward bending of the reciprocal
values, away from the earlier straight line. This part of the data is shown in Figure 2.
Hyperbolic fit to the world GDP data (Maddison, 2010) is shown in Figure 3. The fit to the
data is remarkably good. The point at AD 1 is only 77% away from the fitted curve. We
would need more data between AD 1 and 1000 to decide whether such a difference is of any
significance.
Hyperbolic economic growth of the historical GDP has been uniquely identified by the
straight-line fitting the reciprocal values of data. However, for comparison, we are also
showing the best fit using exponential function for precisely the same data over the same
time.
While the hyperbolic growth can be easily identified as a straight line in the display of the
reciprocal values of data, exponential growth can be easily identified as a straight line in the
direct display of data using the semilogarithmic scales of reference. Furthermore, while the
reciprocal values for the hyperbolic growth decrease linearly, the reciprocal values for the
exponential growth decrease exponentially. Even without trying to fit the data with
exponential distributions we can easily see in all the figures presented in this study that the
historical growth of the GDP, global or regional, was not exponential.
Parameters describing hyperbolic trajectory fitting the data between AD 1000 and 1955 are:
a 1.684 10 2 and k 8.539 10 6 . The point of singularity for this fit is at t 1972 .
From around 1955, the world economic growth started to be diverted to a slower trajectory,
as indicated by the upward bending of the reciprocal values shown in Figure 2. The diversion
to a slower trajectory bypassed the singularity by 17 years (see Table 1).
Western Europe
The growth of the GDP in Western Europe is shown in Figures 4-6. It is the total for 30
countries: Austria, Belgium, Denmark, Finland, France, Germany, Italy, The Netherlands,
Norway, Sweden, Switzerland, United Kingdom, Greece, Portugal, Spain and for 14 small,
but unspecified countries. Ireland is the missing country in this list but it was included from
1921.
The best hyperbolic fit is between AD 1500 and 1900. The parameters of the hyperbolic
distribution are a 9.859 10 2 and k 5.112 10 5 . The point of singularity for this fit is at
t 1929 . Between 1900 and 1910, economic growth started to be diverted to a slower, but
still fast-increasing, trajectory.
To demonstrate that it is possible to improve the fit to the historical economic growth by
using more complicated mathematical expressions than the first-order hyperbolic
distributions defined by the eqn (1), we are showing, in Figure 6, the best fit to the data using
the fourth-order hyperbolic distribution:
1

S (t )

ai t

(4)

i 0

Parameters fitting the data are: a0

6.860 10 2 , a1

6.898 10 4 , a2

1.250 10 6 , ,

a3 7.228 10 10 and a4
1.400 10 13 . The calculated distribution reproduces the data
below AD 1500 but is unappealing. We would need more data in this region but even then
such a complicated description would be still unattractive.

The most complete set of data for Western Europe are for four countries: Denmark, France,
Netherlands and Sweden. These are the only countries that have yearly data from 1820. They
are analysed separately and results are presented in Figures 7 and 8.
The quality of the hyperbolic fit to the data is virtually the same as for the total of the 30
countries but now the fitted curve passes also through the AD 1 point. However, it still does
not reproduce the point at AD 1000.
The historical growth of the GDP in Western Europe is consistent with the hyperbolic
growth. The diversion to a slower trajectory for these four countries appears to have occurred
earlier than for the whole Western Europe, around 1875 rather than around 1900, or 48 years
rather than 29 years before the time of the singularity.
Parameters describing the historical hyperbolic growth of the GDP in Denmark, France,
Netherlands and Sweden are: a 3.821 10 1 and k 1.986 10 4 . The point of singularity for
this fit is at t 1923 . From around 1875 economic growth in Denmark, France, Netherlands
and Sweden was diverted to a slower trajectory.
Eastern Europe
Systematic data for Eastern Europe are available only for seven countries: Albania, Bulgaria,
Czechoslovakia, Hungry, Poland, Rumania and Yugoslavia. For other countries there are no
data until 1990. The analysis of the historical data for Eastern Europe are summarised in
Figures 9 11.
The best hyperbolic fit to the data is between AD 1000 and 1890. Hyperbolic parameters are:
a 7.749 10 1 and k 4.048 10 4 . The point of singularity for this fit is at t 1915 . From
around 1890, the economic growth in Eastern Europe was diverted to a slower trajectory,
bypassing the singularity by 25 years.
Former USSR
The analysis of the data for countries of Former USSR is presented in Figures 12-14. The
best hyperbolic fit is between AD 1 and 1870. Parameters fitting the data are: a 6.547 10 1
and k 3.452 10 4 . The point of singularity for this fit is at t 1897 . From around 1870, or
maybe even a little earlier, the economic growth in the Former USSR was diverted to a
slower trajectory, bypassing the singularity by at least 27 years.
Asia
The analysis of the historical economic growth in Asia is summarised in Figures 15, 16 and
18. The best hyperbolic fit is between AD 1000 and 1950. Parameters fitting the data are:
a 2.303 10 2 and k 1.129 10 5 . The point of singularity for this fit is at t 2040 .

The data and their analysis indicate that the economic growth accelerated around 1950. The
acceleration is indicated by the downward diversion from the original straight line
representing hyperbolic trajectory.
The new trajectory was non-hyperbolic, which was at first increasing significantly faster than
the historical hyperbolic growth, as indicated by the steeper gradient of the reciprocal values.
However, the absolute gradient of the new reciprocal trajectory was gradually decreasing
while the absolute gradient of the historical hyperbolic growth remained constant indicating
that the new trajectory was initially accelerating faster than the hyperbolic trajectory but the
acceleration of the new trajectory was progressively slower.
The decreasing absolute gradient of the reciprocal values should not be interpreted as a
slowing-down growth. For instance, the absolute gradient of the reciprocal values of an
exponential growth are decreasing but the corresponding exponential growth is obviously
accelerating.
The new trajectory, which commenced around 1950, started 90 years before the singularity of
the earlier hyperbolic growth. However, the new growth was initially significantly faster than
the historical hyperbolic trajectory. If continued hyperbolically, its singularity would have
been at t 1965 , i.e. only 15 years from its commencement. Fortunately, the new trajectory
was not hyperbolic.
Preliminary analysis indicates that the new trajectory is exponential, which is almost as bad
as hyperbolic because exponential growth becomes also unsustainable at a certain time and it
does not matter that such a growth is not characterised by singularity. It can lead to an
economic collapse just as effectively as the hyperbolic growth.
Closer inspection of data suggests that the economic growth in Asia might have accelerated
even before 1950. The reciprocal values of data presented in Figure 16 seem to suggest that
the economic growth became slower between 1820 and 1870 but then it followed a faster
hyperbolic trajectory until around 1940, when it again decreased between 1940 and 1950.
From around 1950 the economic growth in Asia accelerated irreversibly along a new but nonhyperbolic trajectory.
This interpretation seems to be supported by the data for Japan presented in Figure 17. This is
the only country in this group that contains the most complete set of data. The four points
between 1870 and 1940 for Asia are now replaced by a series of closely-spaced points
showing clearly that there was a brief but faster hyperbolic growth between around 1870 and
1920. From around 1920, the economic growth became slower until 1944. Between 1944 and
1945 there was a sudden decrease in the GDP value. From 1945, the economic growth in
Japan started to follow an entirely new, non-hyperbolic trajectory.
Africa
Results of analysis of economic growth in 57 African countries are presented in Figures 1921. Reciprocal values of the GDP data presented in Figures 19 and 20 show clearly that the
economic growth was following two hyperbolic distributions. At first it was a slow economic
growth between AD 1 and 1820 characterised by parameters a 1.244 10 1 and
k 5.030 10 5 and by the singularity at t 2473 . Around 1820, this slow hyperbolic
growth was replaced by a significantly faster growth characterised by parameters
a 4.192 10 1 and k 2.126 10 4 and by the singularity at t 1972 . Defined by the
parameter k, this new growth was 4.2 times faster than the earlier hyperbolic growth. From
around 1950, this fast hyperbolic growth was diverted to a slower trajectory, bypassing
singularity by 22 years.
5

Latin America
Results of analysis of the economic growth in Latin America are presented in Figures 22
24. Data for Latin America are difficult to analyse but they also appear to follow two
distinctly different hyperbolic trajectories. However, the identification of the first trajectory is
not as clear as for Africa and is based on the alignment of only three points, which could be
questioned. The identification of the second hyperbolic trajectory is more convincing.
Tentative conclusion is that the economic growth in Latin America was following a slow
hyperbolic distribution between AD 1 and 1500 and a fast distribution between AD 1600 and
around 1870.
The tentatively assigned slow hyperbolic growth between AD 1 and 1500 is characterised by
parameters a 4.421 10 2 and k 2.093 10 5 . Its singularity is at t 2113 . The better
determined fast hyperbolic growth between AD 1600 and 1870 is characterised by parameters
a 1.570 10 1 and k 8.224 10 5 . Its singularity is at t 1910 . Defined by the parameter k,
this growth was 3.9 times faster than the earlier hyperbolic growth. From around 1870, this
fast hyperbolic growth was diverted to a slower trajectory bypassing the singularity by 40
years.
Reciprocal values of data show also clearly that there was a significant decline in the
economic growth between AD 1500 and 1600. The reciprocal value of the GDP in AD 1600
is higher than in 1500. Indeed, the GDP value decreased from $72.88 billion in AD 1500 to
$37.63 in 1600 but this decline was quickly compensated by the faster hyperbolic growth,
which commenced in 1600.
Australia
Australia is one of the four countries belonging to Western Offshoots. The other three are
Canada, New Zealand and the USA. The data for this group are difficult to analyse because
many essential historical data are missing. Fortunately, Australia has excellent set of data
containing the annual list of the GDP values between 1820 and 2008.
Results of the analysis of the economic growth in Australia are presented in Figures 25 27.
The reciprocal values of the data between AD 1 and 1700 can be fitted by a straight line
suggesting a hyperbolic distribution. The parameters describing this distribution are
a 6.986 100 and k 8.739 10 4 . Its singularity is at t 7994 .
Between AD 1700 and 1820, economic growth decreased from $180 million per year to $173
million but then it started to follow an exceptionally fast hyperbolic trajectory (see Figure 26)
described by parameters a 3.215 10 2 and k 1.736 10 1 . Its singularity was at t 1852.7
From around 1848, the economic growth in Australia started to be diverted into a slower
trajectory, bypassing the singularity by only 5 years.
The current economic growth is slower but not without danger because it continues to
accelerate. On average, the annual increase was about $3,000 million around 1950 and
$15,000 million by the end of the last century. In only 50 years, the annual growth
experienced a 5-fold increase.
It appears that the growth is exponential, which is bound to be unsustainable. The danger of
an overshoot is as real for the exponential growth as it is for the hyperbolic growth. It is
delayed but not cancelled.
If the data are of good quality, the type of growth can be determined not only by the analysis
of the reciprocal values of data but also by the analysis of the gradient and of the growth rate.
6

Thus for instance, for the hyperbolic growth, the growth rate is directly proportional to the
size of the growing entity, while for the exponential growth it is constant. We can illustrate
this method by using the good quality of data for Australia.
In Figure 28 we show the growth rate calculated directly from the data and by interpolation.
Direct calculations are sensitive to the local gradients and consequently the calculated values
show strong fluctuations. A far better way is to calculate growth rate using interpolated
gradients. As we can see in Figure 28, calculations using interpolated gradients do not show
such strong fluctuations.
Figure 28 shows the fitted trends to the growth rate calculated directly from the data and for
the growth rate calculated using interpolated gradients. Both fitted trends are similar. Finally,
Figure 28 shows also the growth rate for the hyperbolic distribution given by the equation
R kS , where R is the growth rate, S is the GDP and k 1.736 10 1 as obtained by fitting
hyperbolic distribution to the data between 1820 and 1848.
Whether calculated directly or by interpolation, the trends for the respective growth rates
follow closely the straight line representing hyperbolic growth, confirming that the data were
indeed increasing hyperbolically during that short time. The growth rate increased from
around 3% per year to around 20% per year.
Western Offshoots
As mentioned earlier, Western Offshoots include Australia, Canada, New Zealand and the
USA. Reciprocal values of the combined GDP for these four countries are presented in
Figure 29. The analysis of the economic growth is obviously impossible without data and
there are no data between AD 1700 and 1820 for this group of countries. We can only suggest
that there was a fast hyperbolic growth during that time but we cannot verify it. Our
suggestion is based on observing a similar pattern for Australia, and even for Latin America
and Africa. There is also a confirmation of such a pattern for New Zealand, but we cannot fit
the suggested distribution to the data representing economic growth in Western Offshoots
because there is not even a single point between the apparent commencement of the
suggested fast growth and its apparent end.
Tentatively assigned hyperbolic distributions are presented in Figures 29-31. The data
presented in Figure 29 appear to suggest a slow hyperbolic growth between AD 1 and 1500
and a fast hyperbolic growth between AD 1700 and 1820. The economic growth decreased
between AD 1500 and 1700.
Under these tentative assumptions we interpret the historical economic growth in Western
Offshoots as being represented by two hyperbolic distributions separated by a brief decline.
Parameters for the tentatively assigned slow hyperbolical trajectory between AD 1 and 1500
are a 2.232 100 and k 8.938 10 4 . Its singularity is at t 2494 . For the tentatively
assigned fast trajectory, between 1700 and 1820, the parameters are: a 1.716 101 and
k 9.387 10 3 . Its singularity is at t 1828 . From around 1820 the data follow a welldefined but non-hyperbolic trend (see Figure 30).
Summary and conclusions
Results of the mathematical analysis of the historical economic growth are presented in Table
1. The listed parameters are for the fitted hyperbolic distributions.

Table 1 Summary of the mathematical analysis or the historical economic growth

Region/Countries

Hyperbolic
Range

Singularity

Proximity

World

1.684 10

8.539 10

1000 1955

1972

17

W. Europe

9.859 10

5.112 10

1500 1900

1929

29

W. Europe (4)

3.821 10

1.986 10

1 1875

1923

48

E. Europe

7.749 10

4.048 10

1000 1890

1915

25

Former USSR

6.547 10

3.452 10

1 1870

1897

27

Asia

2.303 10

1.129 10

1000 1950

2040

90

Africa

1.244 10

5.030 10

1 1820

2473

4.192 10

2.126 10

1820 1850

1972

4.421 10

2.093 10

1 1500

2113

1.570 10

8.224 10

1600 1870

1910

6.986 100

8.739 10

1 1700

7994

1.736 10

1820 - 1848

1853

8.938 10

1 1500

1498

9.387 10

1700 1820

1828

Latin America

Australia

3.215 10

W. Offshoots

2.232 100
1.716 10

22

40

a and k Hyperbolic growth parameters [see eqn (1)].

W. Europe (4) Four countries of Western Europe: Denmark, France, Netherlands and Sweden
Proximity Proximity (in years) to the singularity (to the time of the escape to infinity) at the time of the
diversion to a different trajectory (generally slower but faster for Asia).

This analysis demonstrates that the natural tendency for the historical economic growth was
to increase hyperbolically. In general, there is a remarkable agreement between the data and
the calculated hyperbolic distributions. The identification of the preferred hyperbolic growth
between AD 1 and 1500 should be treated as tentative unless the calculated trajectories in this
region includes also the data for the higher values of time.

References
Galor, O. (2005). From stagnation to growth: Unified Growth Theory. In P. Aghion & S.
Durlauf (Eds.), Handbook of Economic Growth (pp. 171-293). Amsterdam: Elsevier.
Galor, O. (2011). Unified Growth Theory. Princeton, New Jersey: Princeton University Press.
Maddison, A. (2001). The World Economy: A Millennial Perspective. Paris: OECD.
Maddison, A. (2010). Historical Statistics of the World Economy: 1-2008 AD.
http://www.ggdc.net/maddison/Historical Statistics/horizontal-file_02-2010.xls.
8

Nielsen, R. W. (2014). Changing the Paradigm. Applied Mathematics, 5, 1950-1963.

http://dx.doi.org/10.4236/am.2014.513188
von Foerster, H., Mora, P., & Amiot, L. (1960). Doomsday: Friday, 13 November, A.D.
2026. Science, 132, 1291-1295.

Appendix
World Economic Growth

Figure 1. Reciprocal values of the GDP (Maddison, 2010) are fitted using straight line
between AD 1000 and 1955 representing hyperbolic growth. Displayed is also the best fit
using exponential function for the same data and over the same range of time. Economic
growth was not exponential.

Figure 2. Reciprocal values of the GDP (Maddison, 2010) showing the diversion of the
economic growth to a slower trajectory.

10

Figure 3. The world GDP data (Maddison, 2010) fitted using hyperbolic distribution. The
point at AD 1 is 77% higher than the calculated distribution. Displayed is also the best fit
using exponential function for the same data and over the same range of time. (Exponential
distributions are shown as straight lines in the semilogarithmic set of reference.) Historical
economic growth was not increasing exponentially.
Western Europe
(Total of 30 countries)

Figure 4. Reciprocal values of the GDP (Maddison, 2010) for Western Europe compared with
the hyperbolic distribution represented by the decreasing straight line.

11

Figure 5. Reciprocal values of the GDP (Maddison, 2010) for Western Europe between AD
1500 and 2008 showing a diversion to a slower trajectory between around 1900 and 1910.

Figure 6. Economic growth in Western Europe. The data (Maddison, 2010) are compared
with the best (first-order) hyperbolic fit. The data point at AD 1 is 42% higher than for the
calculated curve and 48% lower at AD 1000. Displayed is also the best fit to historical data
using the fourth-order hyperbolic distributions (see the text).

12

Western Europe
(Denmark, France, Netherlands and Sweden)

Figure 7. Reciprocal values of the GDP (Maddison, 2010) for four countries (Denmark,
France, Netherlands and Sweden) compared with the straight line representing hyperbolic
growth fitting the data between AD 1 and 1875, when the economic growth started to be
diverted to a slower trajectory.

Figure 8. Economic growth in Denmark, France, Netherlands and Sweden. The data
(Maddison, 2010) are compared with the best hyperbolic fit. From around 1875, the
economic growth started to be diverted to a slower trajectory. The data point at AD 1000 is
41% lower than for the calculated distribution.
13

Eastern Europe

Figure 9. Reciprocal values of the GDP (Maddison, 2010) for Eastern Europe compared with
the hyperbolic distribution represented by the decreasing straight line.

Figure 10. Reciprocal values of the GDP (Maddison, 2010) for Eastern Europe showing that
from around 1890 the economic growth started to be diverted to a slower trajectory.

14

Figure 11. Economic growth in Eastern Europe. The data (Maddison, 2010) are compared
with the best hyperbolic fit. From around 1890, economic growth started to be diverted to a
slower trajectory. The point at AD 1 is 51% higher than for the calculated distribution.

Former USSR

Figure 12. Reciprocal values of the GDP (Maddison, 2010) for the Former USSR compared
with the hyperbolic distribution represented by the decreasing straight line.

15

Figure 13. Reciprocal values of the GDP (Maddison, 2010) for the Former USSR showing
that from around 1890 economic growth started to be diverted to a slower trajectory.

Figure 14. Economic growth in the Former USSR. The data (Maddison, 2010) are compared
with the best hyperbolic fit. From around 1890, economic growth started to be diverted to a
slower trajectory.

16

Asia

Figure 15. Reciprocal values of the GDP (Maddison, 2010) for Asia compared with the
hyperbolic distribution represented by the decreasing straight line.

Figure 16. Reciprocal values of the GDP (Maddison, 2010) for Asia showing that from
around 1950 economic growth accelerated. However, the data also suggest a transient stage
between 1870 and 1940.

17

Figure 17. Reciprocal values of the GDP (Maddison, 2010) for Japan showing a transient
stage of growth between 1870 and 1944, a sudden drop in the GDP value between 1944 and
1945 and a new but non-hyperbolic trajectory from around 1945.

Figure 18. Economic growth in Asia. The data (Maddison, 2010) are compared with the best
hyperbolic fit. From around 1950, economic growth accelerated but by 2008 it was about to
cross the earlier hyperbolic trajectory. The data point at AD 1 is 76% higher than for the
calculated distribution.

18

Africa

Figure 19. Reciprocal values of the GDP (Maddison, 2010) for Africa compared with the
hyperbolic distributions represented by the decreasing straight lines.

Figure 20. Reciprocal values of the GDP (Maddison, 2010) for Africa showing that from
around 1950 economic growth started to be diverted to a slower trajectory.

19

Figure 21. Economic growth in Africa. The data (Maddison, 2010) are compared with
hyperbolic distributions. From around 1950, economic growth started to be diverted to a
slower trajectory.

Latin America

Figure 22. Reciprocal values of the GDP (Maddison, 2010) for Latin America compared with
hyperbolic distributions represented by the decreasing straight lines.

20

Figure 23. Reciprocal values of the GDP (Maddison, 2010) for Latin America showing that
from around 1870 the economic growth started to be diverted to a slower trajectory.

Figure 24. Economic growth in Latin America. The data (Maddison, 2010) are compared with
hyperbolic distributions. From around 1870, economic growth started to be diverted to a
slower trajectory.

21

Australia

Figure 25. Reciprocal values of the GDP (Maddison, 2010) for Australia compared with
hyperbolic distributions represented by the decreasing straight lines.

Figure 26. Reciprocal values of the GDP (Maddison, 2010) for Australia showing that from
around 1820 to 1848, the economic growth was following an exceptionally fast hyperbolic
distribution. From around 1848, the growth started to be diverted to a slower trajectory.

22

Figure 27. Economic growth in Australia. The best fits to the data (Maddison, 2010) suggest a
slow hyperbolic growth between AD 1 and 1700 followed by a fast hyperbolic growth
between 1820 and 1848. From around 1848, economic growth started to be diverted to a
slower, but still increasing trajectory.

Figure 28. Growth rate describing economic growth in Australia between 1820 and 1848
plotted as the function of the GDP. Whether calculated directly from the data, R (Direct), or
by interpolation of the gradient, R (Refined), the trends of the growth rates follow closely the
hyperbolic growth rate given by the equation R kS , where R is the growth rate, S is the
GDP and k 1.736 10 1 as obtained by fitting hyperbolic distribution to the data between
1820 and 1848.

23

Western Offshoots

Figure 29. Reciprocal values of the GDP (Maddison, 2010) for Western Offshoots (Australia,
Canada, New Zealand and the USA) compared with the tentatively assigned hyperbolic
distributions represented by the decreasing straight lines.

Figure 30. Reciprocal values of the GDP (Maddison, 2010) for Western Offshoots (Australia,
Canada, New Zealand and the USA) showing a clear, non-hyperbolic trajectory commencing
around 1820.
24

Figure 31. Economic growth for Western Offshoots (Australia, Canada, New Zealand and the
USA). The data (Maddison, 2010) are compared with the tentatively assigned hyperbolic
trajectories. See the text.

25