The purpose of my blogs is to estimate Systems Models, specifically Dynamic Component Models, on historical data. My reason for the project are: (1) The historical data is widely available, (2) current analysis is largely based on mental models and (3) no one seems to be applying Systems models. The reason for point three is that historical data are highly correlated and co-linear. My contribution is to use Principal Components Analysis to implement the Satet Space and solve the co-linearity problem using correlations.
The video above explains State Space Systems models which are commonly used in Engineering (not Economics, History or the Social Sciences). What is unique about the DCM models is the definition of System state and how the State Space is constructed:
The state of any system is the collection of independent variables that predict the time path of the system from t1 to tn.
In DCMs, the state space is constructed using Principal Components Analysis. In the video above, the x are the state variables and the y are the indicator variables. For macro-societal analysis, the scope of the system is defined by the Kaya Identity, the same identity used by the Intergovernmental Panel on Climate Change (IPCC).
DCMs can be run using R-code on my Google Site, which is organized basically by regions of the World-System, by topics and by time periods (for example, the Long Twentieth Century).
- WL20 World System
- WL20 Eastern Europe EE
- World System Models
- World System (1950-2000+) Late Modern Period
- WL20 European Union
- World-System (1900-1950) World in Depression
- World in Depression
- World System Forecasts
- Sub-Saharan Africa
- World System Steady State or Collapse
- Middle East Africa (MEA)
- USL20 (and more)
- East Asia Pacific (EAP) Region
- WL20 Asia
- World-System (1900-2000+) The Long Twentieth Century
- Latin American Region (LAC) Long Twentieth Century
- The Long Nineteenth Century (1789-1914)
- World-System Policy
- The Long Twentieth Century (1950-2000+) Geopolitical Alignments
- World-System Technology
- World-System Cycles
- Early Capitalism (1500-1800)
- World-System (500-1800) Middle Ages
- World-System (1300-1450) Late Middle Ages
- World-System (1150-1300) High Middle Ages
- World-System (0-500) Roman Empire
- World-System (1640-1815) Long Seventeenth Century
- World-System Long 16th Century (1450-1640)
- The Long Eighteenth Century (1688-1815)
- World-System (0-2000) Long Duree
- Policy Space and the AD-AS Market Model One useful extension of the Kaya Identity.
- Causal Macro Systems See a simulation of the Kaya Identity and compare it to other macro-models.
- Sub-Saharan Africa (SSA)
- North America (NAC)
- The Classic Economists Interestingly enough, the Classical Economists had more affinity with Systems Theory than do the Neoclassical Economists. Classical Economic models can be easily translated in Systems Theory.
- Dynamic Component Models R-code for estimating, testing and simulating DCM models (all models are implemented in the R programming Language).
Notes
Kalman, R. E. (1982) Dynamic Econometric Models: A System-Theoretic Critique
Annales (2015) Debating the long Duree
Pasdirtz, G.W. Instability and Late Nineteenth Century German Development
Please see the Boiler Plate for more information on periodization, data sources and DCM Construction.
