State Space Models

All state space models are written and estimated in the R programming language. The models are available here with instructions and R procedures for manipulating the models here here.
Showing posts with label Cycles. Show all posts
Showing posts with label Cycles. Show all posts

Thursday, October 16, 2025

World-System (1960-2150) La French TECH Cycles



In a prior post (here) and in prior posts on the French Economy (here), I have found that the French Economy (or at least my model of it) is approaching a Steady State. My hypothesis is that the approaching steady state might (in addition to a lot of other forces) be creating Political Instability. However, no one knows the Future and we have to entertain other hypotheses.

The current 2025 Noble Prize in Economics offers another hypothesis. History teaches us that new technologies and Creative Destruction will eventually break steady states and send Economic Systems on to new Attractor Paths. The hypothesis is the essential argument of Kondratiev Wave Theory (here) and is embraced by World-Systems Theory. An alternative forecast for France is that the Steady State will not happen due to La French TECH.

The graphic above plots two historical forms of French Technology: Productivity (TECHP) and Efficiency (TECHE) (see the Notes below). Notice that they are both cyclical (echoes of Schumpeter's Creative Destruction model which is also cyclical but unstable). Notice also that TECHP (productivity) peaks before TECHE (efficiency): once a productivity peak is reached, focus turns toward efficiency until all gains are exhausted.

Unfortunately for France, both of these technological peaks (at least in my model which is estimated from World Development Data) are going to be reached in the near future: (1) reinforcing the steady state or (2) creating more instability and a search for new technologies to put the economy on another growth path. Since we do not know the Future, nothing is guaranteed.

For more on Stable French Technology Cycles see the post hereSchumpeter's Creative Destruction model predicts unstable Creative Destruction cycles and the French Technology Cycles are stable.
 

Notes

FR TECH Measurement Models


The two types of Technological Change  (TECHP and TECHE) are explained in a post on Technology in the United Kingdom which also is a good comparison to the French Model. Essentially, TECHP is productivity (output per capita)  and TECHE is efficiency (for example, CO2 emissions relative to energy consumption).

Friday, October 10, 2025

World-System (1960-2150) Unstable Cycles in Columbia

 



Notes

CO_L20 Measurement Matrix:




CO_L20 BAU System Matrix:






CO_L20 BAU Stabilized Sytem Matrix:














Friday, September 26, 2025

World-System (1960-2500) Unstable System Cycles in Latin America

 


In an earlier post (here and here), I found that the best forecasting model for Argentina was the LA20 Model (Latin America Regional Economy). Out until 2100 (the furthest out the IPCC is willing to go on Global Emission Scenarios), it looks as if Latin American Integration would provide a desirable future, at least for Argentina (here). This post explores what happens after 2100, out to 2500. Of course, no one knows the future (especially out to 2500), but the LA20 Model is a computer program that can be run out to any date. Sometimes it is insightful just to see what happens!


The LA20 Model is unstable and cyclical (the eigenvalues and AIC statistics are presented in the graphic above for the Business as Usual BAU model).  In the Phase Plot, above, time moves from left to right and cycles increase in magnitude and severity.** Although the short-term future might look endurable, the long-term future is not. Luckily, a lot (maybe everything) can change after 2100, especially in the face of Environmental Crisis.

Another reason to look at long-run cycles in Latin America is that cycles (particularly Kondratiev Waves or K-Waves) are discussed extensively and qualitatively in World-Systems Theory. It is refreshingly concrete to see actual historical cycles tested with a statistically estimated model. However, a future of unstable cycles is probably one reason why Latin American Integration has been so problematic.


Notes

** And will require that bailouts increase in magnitude and severity.


LA20 BAU Measurement Matrix:

The State Space of the LA20 model has three components that explain 99% of the variation in the indicators: LA1=(Overall Growth), LA2=(LU-Q-EG) an Unemployment Controller and LA3=(N+L-CO2-Q) a Population Controller. LA1 reinforces the conclusions of Balanced Growth Theory (all major parts of the Economy have to grow together). LA2 balances Unemployment (LU) against overall production (Q) and Energy Use (EG)--reducing LU requires increasing energy-intensive production. LA3 is a Malthusian-Environmental component balancing Population (N) and Labor Force (L) Growth against Emission (CO2) intensive production (Q).

The unstable LA20 BAU model has cyclical periods under a decade but no effective damping time (table above). 
The stable LA20 BAU model also has cyclical periods under a decade but damping times of about 150 years.

NOTE: Periods in both models are longer than those assumed by Kondratiev Waves or K-Wave models (45-60 years). LA2 could be expanded to a Marxian Economic component with (0.901 L - 0.272 Q - 0.239 EG - 0.144 L) where the standard Marxian Economic component is (Q - wL) assuming fixed Ricardian wages. Most of these ideas (MalthusianMarxian EconomicBalanced Growth Theory and Kondratiev Waves or K-Wave) can be combined in Systems Theory models. However, the result for LA does not mean that the ideas can be generalized to all regions, countries and time periods (see Unified Growth Theory).

You can experiment with the LA20 BAU model hereSuggestions are given in the code for how to stabilize the model.

Ex. 1.0 Can you find a way to eliminate cycles once the model has been stabilized? 

The solution to this Exercise can be found in the LA_TECHP model which I will describe in a future post. 

Descriptions of the how the Dynamic Component State Space models are constructed are given in the Boiler Plate.