Abstract:In this paper, we study a neoclassical growth model with a (productivity inhibiting) pollution effect. In particular, we obtain a necessary and sufficient condition for the existence of a topological chaos. We investigate how the condition changes as the strength of the pollution effect changes. This is a new application of a recent result characterising the existence of a topological chaos for a unimodal interval map by Deng, Khan, Mitra (2022).
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: Under the condition of introducing the pollution effect, determine the necessary and sufficient conditions for the existence of chaotic dynamics in the neoclassical growth model, and study how these conditions change as the intensity of the pollution effect changes.
### Problem Background
The paper starts from a simplified neoclassical growth model, which is represented in discrete time and has production lags. Its difference equation is:
\[
k_{t + 1}=s\cdot f(k_t)
\]
where \(k_t\) is the capital - labor ratio \(K_t/L_t\), \(s\) is the fixed savings rate, and \(f\) is the per - capita production function. Assume that there is no population growth, and the production function \(f(k) = A k^\gamma\) is a Cobb - Douglas production function, where \(A>0\) and \(0 <\gamma\leqslant1\).
According to the above equation, we can conclude that:
1. If \(0 <\gamma< 1\), then \(k\) converges to the steady state \(k^*=(sA)^{1/(1 - \gamma)}\).
2. If \(\gamma = 1\), then \(k\) may converge to 0, remain at the initial value \(k_0\), or diverge to infinity, depending on the value of \(sA\).
The key point is that in any case, chaotic behavior will not occur. In order to introduce chaotic behavior, the author introduces a productivity - suppressing effect (i.e., "pollution effect"), so that the production function becomes:
\[
f(k)=A k(1 - k)^\beta
\]
where \(\beta\geqslant0\). This new production function means that when \(k\) is close to the maximum value (set as 1), the term \((1 - k)^\beta\) is close to 0, thus simulating the pollution effect.
Finally, the research object obtained is the difference equation:
\[
k_{t + 1}=f(k_t)=s A k_t(1 - k_t)^\beta=\alpha k_t(1 - k_t)^\beta
\]
where \(\alpha>0\) and \(0\leqslant k_t\leqslant1\).
### Main Research Content
The main objective of the paper is to find the necessary and sufficient conditions for the existence of chaotic dynamics and study how these conditions change with the change of the pollution effect intensity \(\beta\). Specifically, the author uses a combination of algebraic and numerical calculation methods to explore the relationship between the parameters \(\alpha\) and \(\beta\) to determine the existence conditions of topological chaos.
### Research Results
By applying the characterization results of Deng, Khan, Mitra (2022) on the existence of topological chaos in unimodal interval mappings, the author obtains the following main conclusions (Theorem 1.3):
For different \(\beta\) values (1, 2, 3, 10), if \(\alpha\) satisfies a specific range, then the mapping \(f\) has the following properties:
1. \(f\in G\).
2. The necessary and sufficient conditions for the existence of odd - period cycles of \(f\).
3. The necessary and sufficient conditions for \(f^2\) to be turbulent.
For example, when \(\beta = 1\), if \(3.679 <\alpha\leqslant4\), then the mapping \(f\) has an odd - period cycle; if \(3.679\leqslant\alpha\leqslant4\), then \(f^2\) is turbulent.
### Conclusion
The paper shows that as the pollution effect intensity \(\beta\) increases, the lower limit of \(\alpha\) required to generate chaos becomes looser, that is, the value range of \(\alpha\) becomes wider. In other words, to generate chaotic behavior, a larger \(\alpha\) (controlling the vertical stretching of the graph) or a larger \(\beta\) (strong pollution effect) is required.