Economic cycle math

Denis Lukyanov

May 14, 2020

Abstract

In this short paper I would like to present a rigorous mathematical and yet heuristic description of modelling the
economic cycles. The idea behind the cycles is that its dynamics could be decomposed into oscillation of two wave
functions with low frequency (long-term cycle) and high frequency (short-term cycle). Finding the solution/s for
these equations will be done with a use of a simple trigonometric functions such as sin wt and cos wt that lie at the
heart of the wide amount of physical processes from Newtonian to quantum mechanics. Ultimately, we will arrive at
the energy equation – that is a driving force of this process.

Keywords: economic cycle, superposition of harmonic functions, wave frequencies, energy integral.
Denis V Lukyanov
E-mail address: denluk@list.ru
Phone: +7(910)440-2096

Electronic copy available at: https://ssrn.com/abstract=3601044

General equation. In order to find an equation of the type with sin 𝑤𝑡 function that will satisfy for all w (w is an
angular frequency) we need to solve:

/̇
−w- = /̈ ; by differentiation we get: 0 =(−ẇ - ) = (/ )̈ =
/ /⃛∙/6/̈ ∙/̇
∀w∶ , by solving it we get
/7

⃛ ∙ 𝒙 − 𝒙̈ ∙ 𝒙̇ =0 (∀ w a harmonic function including sin w𝑡 and cos w𝑡 satisfy this eq.)

𝒙

The problem with this equation is that it’s non-linear: superposition of solutions (and we need a sum for two waves)
not always is a solution for this equation because of its nonlinearity.

Searching for the general solution as a superposition. For the dynamics of the economic machine we first approach
the problem by taking the sum of two harmonics with different frequencies (long/short cycles):

𝑨 𝐬𝐢𝐧 w𝒕 + 𝑩 𝐜𝐨𝐬 w𝒕+𝑨 𝐬𝐢𝐧 𝛀𝒕 + 𝑩 𝐜𝐨𝐬 𝛀𝒕 (1)

w-high frequencies, 𝛀-low frequencies

The system that satisfies this equation is:

B𝑥̈ + w- 𝑥 = 0 by solving the system of characteristic equations Bλ- + w- = 0

- - -

𝑥̈ + Ω 𝑥 = 0 λ +Ω =0

we get λH + (w- + Ω- )λ- + w- Ω- = 0

and come to the following linear differential equation:

𝒙̈̈ + (w𝟐 + 𝛀𝟐 )𝑥̈ + w𝟐 𝛀𝟐 𝑥 = 𝟎

For interpretation we subdivide it into the system of two equations:

𝒙̈ = 𝐩
B (2)
𝐩̈ + (w𝟐 + 𝛀𝟐 )𝐩 + w𝟐 𝛀𝟐 = 𝟎

General solution looks as follows: ẍ = −𝑎- 𝑥 + p ; by differentiating twice we get 𝑥̈̈ + aẍ = 𝑝̈
(making first substitution) ê 𝑥̈̈ = −(w- + Ω- )𝑥̈ − w- Ω- 𝑥 Þ (𝑎- − w- − Ω- )𝑥̈ − w- Ω-𝑥 = 𝑝̈ ê𝑥̈ = −𝑎- 𝑥 + p

p̈ = (𝑎- − w- − Ω- )p−𝑎- (𝑎- − w- − Ω- )𝑥 − w- Ω-𝑥

-
p̈ = −(w- + Ω- − 𝑎- )p−(𝑎H − 𝑎- w − 𝑎- Ω- + w- Ω- )𝑥

p̈ = −(w- + Ω- − 𝑎- )p−(𝑎- − w- )( 𝑎- − Ω- )𝑥

So, in general case we get:

𝑥̈ = −𝑎- 𝑥 + p
𝑥̈̈ + (w- + Ω- )𝑥̈ − w- Ω- 𝑥 ⟺ B
p̈ = −(w- + Ω- − 𝑎- )p−(𝑎- − w- )( 𝑎- − Ω- )𝑥

Next we check that the system represent two waves with high and low frequencies. Define 𝑎 = w as a high frequency
(i.e. short cycle) and we come to the system:

Electronic copy available at: https://ssrn.com/abstract=3601044

 𝒙̈ = −w𝟐 𝒙 + 𝐩
R (3)
𝐩̈ = −𝛀𝟐 𝐩

By taking the general solution in x in of the form of 𝑥 = 𝐴 sin w𝑡 + 𝐵 cos w𝑡+𝐴 sin Ω𝑡 + 𝐵 cos Ω𝑡 Þ

𝑥̈ = −𝐴 w- sin w𝑡 −𝐵 w- cos w𝑡 −𝐶 Ω- sin Ω𝑡 −𝐷 Ω- cos Ω𝑡 Þ

p = 𝑥̈ + w- 𝑥 =−𝐶 Ω- sin Ω𝑡 −𝐷 Ω- cos Ω𝑡 and we get p that really corresponds to big waves with low frequencies
(long-term cycles) as could be seen from (3).

One vital correction. As soon as we are modelling an undamped oscillation one correction to equation (1) is essential,
i.e. exclusion of cosine wave that is acting as a damper (and also leave only even degrees of derivative). Starting with:

𝑥 = 𝐴 sin w𝑡 + 𝐵 sin Ω𝑡 Þ

𝑥̈ = −𝐴 w- sin w𝑡 − 𝐵 Ω- sin Ω𝑡 Þ

𝑥̈̈ = 𝐴 wH sin w𝑡 + 𝐵 ΩH sin Ω𝑡

At the end, we get solution identical to (2) in the form: 𝑥̈̈ + α𝑥̈ + β𝑥 = 0 with α = w- + Ω- , β = w- Ω- by Vieta’s
theorem from characteristic equation λH + αλ- + β = 0 with roots λZ = − w- , λ- = − Ω-.

Energy integral. By investigating the system (2) further with standard procedure:

𝑥̈ = p ½×𝑥̇
R Þ R 𝑥 ̈ 𝑥̇ − p𝑥̇ = 0½×β and subtracting one from another Þ
p̈ = −αp − β𝑥½×ṗ p̈ ṗ + αpṗ + βṗ 𝑥 = 0

p̈ ṗ + αpṗ + βṗ 𝑥 − β𝑥̈ 𝑥̇ + βp𝑥̇ = 0 Þ we get Energy equation:

𝐩̇ 𝟐 𝒙̇ 𝟐 \ ̇ 𝒑𝟐 𝛂
[𝟐 − 𝟐
+ 𝟐
+ βp𝑥_ = 0

𝐩̇ 𝟐 𝒙̇ 𝟐 𝛃 𝐩𝟐 𝛂
Kinetic energy − − (depends on derivatives), Potential energy + 𝛃𝐩𝒙
𝟐 𝟐 𝟐

By finishing these calculations, we arrive at the physical interpretation of the economic machine’s process with two
cycle waves in terms of kinetic and potential energy. That means the process is not only a mathematical manipulation
of harmonic function formulas, but has a real physical force behind it. The concrete form of force and more generally
a physical process needs a further investigation.

R code. At the end I attach a simple version of a toy model in R for visualizing cycles with frequencies acting as
independent variables/parameters: 100 years corresponding to long-term cycle, and 10 years to short-term. The values
of phase and amplitude are subject to initial conditions and are taken arbitrary. Further extensive research should be
performed on actual data.

library(xts)

t=(1:200) #period

stc = 0.7*sin(2*pi*t/10); #short-term cycle

ltc = 2*sin(2*pi*t/100); #long-term cycle

lt=0.015*t #trend, because as in economic machine the level is growing in time

plot.ts(stc+ltc+lt, main="The cycle dynamics")

Electronic copy available at: https://ssrn.com/abstract=3601044

References
[1] Dalio Ray, Big Debt Crises, Bridgewater, 2018.
[2] A.A. Andronov, A.A. Vitt ,S. E. Khaikin, Theory of oscillators, Physmathlit, 2nd edition, 1959.
[3] L.D. Landau, E.M. Lifshitz, Mechanics: Volume 1 (Course of Theoretical Physics S), Physmathlit, 5th edition 2007.
[4] U.M. Grigoriev, I.S. Kychkin, Oscillations and waves, Physmathlit, 1st edition, 2018.

Electronic copy available at: https://ssrn.com/abstract=3601044