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 Trade War. Show all posts
Showing posts with label Trade War. Show all posts

Thursday, July 31, 2025

World-System (2015-2100) US Attempts to Dominate India

 


Reuters is reporting (here) that the Trump II Administration is planning to impose 25% Tariffs on India and, not surprisingly, creating Political Chaos in India. My first question is "What is really going on here?" Trump's Trade War makes little economic sense and is opposed to conventional Republican Right-Wing Free-Market Ideology.

If Trump's Trade War isn't about Economics, what is it about? My best guess (and there might not be a rational explanation) is that the Trade War is about Political Power. Trump is intent on the Economic Domination of every country in the World-System and has decided that a Trade War is the way to do it. And, right now, he seems to be winning--most countries appear anxious to capitulate. How the Trade War helps Trump personally or the US Economy, if at all, remains to be seen.


Another (and maybe more important questions) is what kind of Geopolitical Alignment is best for India (or any other country, for that matter)? In terms of the models I have estimate, the World-System Linkage model (W)** is the best for India [-99.33, AIC=-60.09, -28.39]*** while in terms of predicted performance, the BAU model [-9.571, AIC=26.53, 57.61] is best (graphic above). In the long-run, US domination will lead to collapse of the Indian Economy (I would assume that the Trump II Administration isn't thinking this far ahead).

Russian domination (RU) is another interesting case. Russia is itself a cyclical economy (think of the Collapse of the Soviet Union) and that cyclicality would be transmitted Geopolitically to India. There would be periods of success and periods of failure.

In the end, my guess is that Indian (and any other country, really) would prefer the BAU model, that is, to be left alone to find their own attractor path

You can experiment yourself with the IN_LM_BAU model here. You can learn more about how the models were constructed in the Boiler Plate. For more discussion of the Indian Economy, see India as a Small Regional Economy.


Notes

** Other types of Geopolitical Alignments estimated in my models are discussed in the Boiler Plate.

***Akaike Information Criterion (AIC) Smaller is better.

The BRICS (of which India is a member) are actually an intergovernmental organization of ten countries but the name comes from just five: Brazil, Russia, India, China, South Africa. You can experiment with the BRICS BAU state space models on the following sites:
  • Brazil (BR_20 Model) An unstable model that leads to collapse (stabilizing the model also leads to collapse). There are strong historical unemployment and environmental Controllers. See if you can find a steady state for this economy!
  • Russia (RU_LM Model) An unstable model with strong Export-Employment historical controllers. Becomes cyclical when stabilized.
  • India (IN_LM Model) A Malthusian Economy currently in collapse mode. Stabilizing the economy is beneficial.
  • China (CN_LM Model) An unstable economy with strong historical Export-Employment controllers. Stabilizing the economy produces a steady state after 2100.
  • South Africa (ZA_20 Model) An unstable economy with strong Globalization-Environmental historical controllers. Stabilizing the economy would lead to a steady state in the distant future (well after 2100).

The BRICS are not a uniform set of Semi-Peripheral countries. They have different dynamics and different historical patterns of development. Neoclassical Economics makes a mistake when applying the same Exogenous Growth Model to these countries.

Trump doesn't think much of either the Indian or the Russian economies (here) and calls them "dead economies". Obviously, the Trade War is not meant to help other countries in the World-System but rather to dominate them (read more about Trump's Trade War here from a World-System Perspective).

Thursday, February 6, 2025

World-System (1980-2100) Eight Futures for Canada


The Trump II administration has threatened a Trade War with Canada and the imposition of tariffs has been delayed while negotiations proceed (here). Canada, Mexico, China and the EU are the US's largest trading partners. I have presented the MXL20 (Mexico Late 20th Century) model here with eight future growth forecasts. In this post, I will do the same for Canada.

The Canadian forecasts are a little different in that they group clearly into two classes: Growth and Steady State (presented above) or Collapse (presented below) when compared against a the Random Walk (RW)--I give more detail about the classes of models with the MXL20 model here--the best two forecasts are the TECHE (Technical Efficiency) forecast and the Business as Usual (BAU) forecast. The forecasts do not predict growth forever (the Techno-Optimist forecast) but rather a steady-state after 2100 (the type of forecast one would get with a Classical Economic model).

The forecast for Geopolitical Alignment with the United States is very interesting and, I would argue, speaks directly to the threatened Trade War. The forecast with the USL20 model driving inputs to the CAL20 model suggests a Growth-and-Collapse scenario with a peak around 2025 at the start of the Trump II administration. In other words, the benefits of Canada's alignment with US have reached their peak and are not forecast to continue into the future--and my forecasts start in 2010!

Both the IMF (here) and the OECD (here) argue (as of 2024) that Canada needs to focus on internal structural issues and Technology, the BAU and the TECHE (Technical Efficiency) forecasts above. Reorienting the Canadian Economy will be difficult during a Trade War with the US, to say the least.


It would be reasonable to argue that some other Geopolitical Alignment might serve Canada better than alignment with the US, which appears to be ending.  I have estimated three other alignment models: World System (W), North America (NA) and China (CN). All of these models (presented above) predict collapse, some collapses being severe (NA and CN, respectively). Even focusing on Technical Productivity (TECHP) is worse than a Random Walk (RW, tomorrow being like today except for random shocks).

In summary, the CAL20 model suggests that Canada should concentrate on getting it's structural house in order and implementing efficiency-producing technologies, two steps that the IMF and the OCED have been emphasizing for the last twenty years. Let the Trade War with the US play it's way out. Geopolitical Alignment with the US has run it's course.

The IMF (here) and the OECD (here) economic reports about Canada bring up many other macro-economic issues. I'll look at those issues in the future. Setting the big picture for the Canadian Economy will help put detailed policy issues in a better perspective. I want to emphasize again that the future is unknowable. All I am doing is creating scenarios based on state space models. We will have to wait for the future to understand how useful such models might be.

NOTES

The Measurement Matrix for the CAL20 model is:


The first six indicators in standard scores are taken from the World Development Indicators (WDI). KOF = KOF Index of Globalization, EF = Ecological Footprint, HDI = Human Development Index. The second two components: CA2=(0.721 LU - 0.675 EF) and CA3 = (0.586 LU + 0.680 EF - 0.249 CO2 - 0.263 Q - 0.2076 N) describe Environmental and Unemployment Error Correction Controllers (ECCs). I will present the behavior of the Canadian ECCs in a future post.

You can run the CAL20 model here. The Bootstrap confidence intervals for the coefficients are:


You can convert the model into a Techno-Optimist model by setting F[1,1] = 1.0 in the System Matrix, but that would be an improbable value although it is close to the Upper Confidence Interval (UCI) of 0.990595.


Monday, February 3, 2025

World-System (1970-2100) Eight Futures for Mexico

 



The Trump II administration has threatened a Trade War with Mexico and the imposition of tariffs has been delayed while negotiations proceed (here). Canada, Mexico, China and the EU are the US's largest trading partners. Trump has threatened tariffs on all four if his demands on immigration and drug trafficking are not met. The common assumption is that the US "Gorilla" will impose it's will on other nations in the world and will get what it wants. But a World Economic Order where the US imposes it's hegemony on the rest of the World-System is not the only option and it may not be the best for any particular country.

In this post, I will create future paths for the development of Mexico based on different assumptions about possible Geopolitical Alignments. First, there is some history here between Mexico and the two Trump Administrations. In 1994, the US signed the North American Free Trade Agreement (NAFTA), but the impetus for free trade started in the Reagan Administration and is a central assumption of Neoliberalism. In 2017, with the start of the Trump I Administration, NAFTA was renegotiated into the United States-Mexico-Canada Agreement (Trump thought NAFTA was a "bad deal"). Now, in the Trump II Administration, the United States-Mexico-Canada Agreement (that Trump negotiated in 2017) is itself a "bad deal" and Trump thinks that the US is being treated very "unfairly". In the future, trading partners might no longer want to deal with the US and a US Republican party that can't decide whether Neoliberal free trade is a good or bad thing.

Currently, Mexico has an important role to play in the Globalization of its trading partners (CN, CA and US) but does not exert much control over the system (here). Future Geopolitical alignments, to include giving up on the World-System, might be one of the the following (presented visually in the time plot at the beginning of this post):  

  • BAU [119.6 < AIC124 < 128] The Business as Usual (BAU) model is not a bad option (you can experiment with it here--see the bootstrap confidences intervals for coefficients below). It would essentially involve Isolationism and would insulate Mexico from US bullying tactics.

  • LAC [73.36 < AIC = 131.5 < 170.1] The Latin American Regional (LAC) Alignment would direct Mexico to trade with it's Latin American Neighbors. It would not mean isolation from the World-System but rather Latin American Integration or a confederation of Latin American countries that would interface with the World-System (an unrealized idea that goes back to the 19th Century). In the MXL20 model, it would result in a steady-state after 2040.

  • RW [38.14 < AIC = 48.06 < 57.18] The Random Walk (RW) model would open Mexico up to all types of shocks from the World-System, with Mexico becoming a Small Country, possibly dominated by some other World hegemonic leader. In the short-run (year-to-year) it is not a bad description of 20th Century Mexican history as one damned thing after another.

  • W [69.96 < AIC = 91.55 < 104.3] Hierarchy Theory would argue that linking Mexico to the World System would be the best way to control growth, environmental problems and unemployment (see the Measurement Matrix below for the MX1, MX2 and MX3 state variable components). For Mexico, such a linkage would mean Degrowth after 2040.

  • US [104.7 < AIC = 112.8 < 119.5] Hegemony Theory would suggest that linking to the Hegemonic World Leader (the US, in this case) would be the best strategy for a Semi-peripheral country such as Mexico. The US and Mexico have been strengthening ties (off and on) over the last three decades, but the Trump Administrations seem intent on severing the relationship. For Mexico, Hegemonic linkage would also mean Degrowth after 2040.

  • CN [103.2 AIC = 116.5 < 129.1] Mexico has also had a history of strengthening relations with China. If the US cuts Mexico free, China may well move into the Geopolitical vacuum. Of the choices uncovered by the MXL20 model, linking with China would be the worst and would lead to almost immediate collapse. The reason is the predicted collapse of the CN21 Model (to be covered in a future post).

  • TECH1 Productivity [63.11 < AIC = 102.8 < 135.3] Mexico could also choose to improve the technological advancement of their economy. One aspect of Technological change is productivity (output per worker, output per energy input, etc.). Although an attractive alternative suggested by Economic Growth Theory, it would be equivalent to a steady-state in Mexico (see LAC above). 

  • TECH2 Efficiency [81.42 < AIC = 130.8 < 160.6] Another aspect of Technological change is efficiency, for example, decreasing energy use per unit of GDP. In Mexico, increases in Economic Efficiency would not be much better than a Random Walk (RW).

As with the IPCC Emission Scenarios the Geopolitical options for Mexico are not forecasts. The future is unknowable. Using different inputs to the MXL20 model, alternate futures can be extrapolated. We will have to wait to see what direction Mexico actually choses, but shocks from the Trump II administration may push the country into making choices.

Notes

MX1 is the dominant state variable of the MXL20 system with data taken from the World Development Indicators (WDI). The methodology used to create forecasts is similar to the one used by the Atlanta Federal Reserves GDPNow app but produces growth scenarios similar to the IPCC Emission Scenarios. Prediction intervals are generated using a Bootstrap algorithm in the R programming language. The AIC can be used to evaluate the quality of models but does not determine which model provides the best forecast. An explanation of Dynamic Component Models (DCMs) can be found here.

The MX1 state variable was created from the following weighted indicators (the first row of the Measurement Matrix) and explain 98% of the variation. 


The first six indicators in standard scores are taken from the World Development Indicators (WDI). KOF = KOF Index of Globalization, EF = Ecological Footprint, HDI = Human Development Index. The second two components: MX2=(0.892 EF- 0.34 LU - 0.22 HDI) and MX3 = (0.847 LU - 0.305 GDP) describe Environmental and Unemployment Error Correction Controllers (ECCs).

You can run the MXL20-BAU model here. The bootstrap confidence intervals for coefficients are:


You can change parameters in the System Matrix (F) in the MXL20 BAU model to any number you want, but the ranges presented above would not be too extreme and would have reasonable probability values.

IPCC Emission Scenarios


Most of the IPCC Scenarios seem to produce a steady state around 2100 (after controlled growth or Degrowth) but there are some outlier scenarios that seem to grow forever. Te MXL20 BAU model is stable and would eventually also reach a steady state (well after 2100).