Policy Wedges and State Space Simulation Models

 

George W. Pasdirtz’s “Policy Wedges” Approach to U.S. Health Care

George W. Pasdirtz’s 2007 paper “Controlling the US health care system with policy wedges” proposes a state-space modeling framework to address the long-term growth of the U.S. health care sector, which has expanded faster than the economy Springer+1.

Core Methodology

Pasdirtz developed two state-space models:

• U.S. economy model (1950–1999)

• U.S. health care system model (1950–1999)

The economy model’s output was used as a reference input to control the health care model’s growth. This allowed him to simulate a “controlled” scenario where health care growth matched economic growth Springer.

Policy Wedges

A policy wedge refers to a targeted intervention that shifts the growth path of the health care system toward the economy’s growth rate. Pasdirtz’s simulations showed that over the late 20th century, the U.S. health care system grew faster than GDP, with health care spending as a share of GDP rising from 3.4% in 1950 to nearly 14% in 1999 Springer.

To align health care growth with the economy, his model suggested:

• 13% reduction in capital expenditure

• 15% reduction in drug prices

• 32% reduction in physician service prices Springer+1

These wedges represent policy levers—changes in investment, pricing, and service delivery—that could slow health care growth without eliminating care.

Designing Universal Health Care

Pasdirtz also applied the framework to universal health care design:

• Use planning and economic incentives rather than over-engineering benefits

• Avoid centralized, command-and-control approaches

• Balance coverage and cost control through targeted interventions Springer+1
• 
  

Key Takeaways

• Policy wedges are measurable, targeted interventions to slow health care growth.

• They can be applied to both cost control and universal coverage design.

• The approach combines macroeconomic modeling with policy simulation to test counterfactual outcomes.

• It offers a data-driven alternative to ad hoc or politically charged reforms.

In short, Pasdirtz’s “policy wedges” framework provides a quantitative, simulation-based roadmap for aligning health care growth with economic growth, with practical implications for both cost containment and universal coverage policy.

UKL20 Financialization Dynamics

 

Notes

UKL20 FINZ Codes

UKL20 FINZ EU Model

UKL20 FINZ US Model

UKL20 Hardship Dynamics

 

Notes

Hardship Measurement Model (HARD)

Where data were available in the World Development Indicators (WDI), the variables above were used to construct three independent components that explain 97.7% of the variation in the indicators.

The first component, HARD1, explained 75% of the variation and was a relatively equal weighting of all the indicators. The second independent component, HARD2, explaining about 20% of the variation, and was a complex Inequality-Male Infant Mortality-Unemployment, Household Expenditure historical controller. The third independent component, HARD3, was a Family Work-Inequality-Male Infant Mortality historical controller that explained 3% of the remaining variation for a total of 97.7% of the variance explained using Principal Components Analysis (PCA).

From the time plots, all the components were stabilizing by 2014.

UKL20 HARD Model

Brexit, Financialization, Hardship and UK Growth

The United Kingdom (UK) joined the European Union (EU) in 1973 and left in 2020.** From the standpoint of Macro-Systems analysis (rather than Political Philosophy): (1) What were the material economic conditions during the 1973-2020 EU period and (2) What would have happened if Britain had stayed in the EU (the primary counterfactual). In this post, I'm going to approach these two questions in terms of Hardship, Financialization and Overall Growth of the UK Economy as measured by indexes created using data from the World Development Indicators.

From the time plots above, during the EU Membership period (1972-2020) Hardship (HARD1) increased, Financialization (FINZ1)  increased and Overall Economic Growth (UK1) increased. Had Britain not left the EU, Hardship would have been reduced below 1960 levels, Financialization would have peaked in 2030 and declined afterward and a Steady-State Economy would have been reached around 2030.

In other words, Britain gave up on the EU Project before the benefits would have been realized, although there would have been intense debate about the benefits of a Steady-State Economy. 

The next question is what will the future look like for the UK without the EU? We can't know the future, but let us assume that Great Britain does nothing different*** once it has left the EU. The graphic above forecasts the UKL20 EU Input model (see below) eliminating the input from the G matrix (see below). In this forecast, Hardship (HARD1) dips to lower levels, Financialization (FINZ) drops more quickly and  there is a very mild collapse from peak growth (UK1) in 2025.

In either case, Brexit will neither stimulate unlimited exponential growth nor create unmitigated disaster. The looming Steady-State Economy will be interpreted as Economic Stagnation and blamed on the political party in power at the time. 

The Notes below explain how this Counterfactual Conclusion was reached.

You can run both the UKL20 BAU Model and the EUL20 BAU Model on my Google site (here). For more background, see my other posts:

• Blog Roll: United Kingdom The Future is "murky" after Brexit!
• Blog Roll: European Union The EU is forecast to become a Steady-State Economy around 2050.
• Blog Roll: Steady-State Economies More economies are forecast to reach a Steady-State Economy around 2100 than you might think. The Steady-State Economy is a meaningful alternative to Neoliberalism but the Political Philosophy remains to be worked out (here).

For more information about how Historical Controllers**** and how the State Space DCM models were constructed, see the Boiler Plate.

Notes

** One problem with identifying the effects of EU Membership in Great Britain is that it coincides with the rise-and-fall of Neoliberalism. 

*** The are many suggestions from the New Labor Government (see below) about how things could be done differently in Great Britain, but one possibility is that nothing other than leaving the EU will change.

**** In State Space DCM models, there are two types of independent components constructed from Principal Components Analysis (PCA). Overall Growth (usually the first component explaining most of the variation) and Historical Controllers (usually the lower-order components). For example, in the Moving Equilibrium model, the first component is overall growth and the next two components capture negative feedback loops that control growth.

References

Shefner et al. (2015) Austerity and Anti-Systemic Protest: Bringing Hardships Back In

The Conversation (2026) How Can Labor Escape the Doom Loop in 2026

Hardship Measurement Model (HARD)

Where data were available in the World Development Indicators (WDI), the variables above were used to construct three independent components that explain 97.7% of the variation in the indicators.

The first component, HARD1, explained 75% of the variation and was a relatively equal weighting of all the indicators. The second independent component, HARD2, explaining about 20% of the variation, and was a complex Inequality-Male Infant Mortality-Unemployment-Household Expenditure historical controller. The third independent component, HARD3, was a Family Work-Inequality-Male Infant Mortality historical controller that explained 3% of the remaining variation for a total of 97.7% of the variance explained using Principal Components Analysis (PCA).

From the time plots, all the components were stabilizing by 2014.

In a future post (here), I will explore Hardship Dynamics in the UK.

Financialization Measurement Model (FINZ)

Financialization (or financialisation in British English) is a term sometimes used to describe the development of financial capitalism during the period from 1980 to the present, in which debt-to-equity ratios increased and financial services accounted for an increasing share of national income relative to other sectors.

The indicators for Financialization in the UK were: (1) Bank Capital-Asset Ratio, (2) Bank Nonperforming loans, (3) Commercial Bank Borrowers, and (4)  Commercial Bank Branches. FINZ1 (76% of the variation) was an historical controller balancing Bank Capital-Asset Ratio and nonperforming loans with Commercial Bank Borrowers and Branches.  FINZ2 (<20% of the variation)  was an historical controller balancing the Capital-Asset ration with Nonperforming loans. FINZ3 (< 0.05%) was Overall Growth.

Notice that Commercial Banking (FINZ1) takes off after 1975 while FINZ2 and FINZ3 are cyclical.

In a future post, I will explore Financialization Dynamics (here) in the UK.

UKL20 Measurement Model

The UKL20 Measurement model is an implementation of the Kaya Identity, measuring the overall state of the system. UK1=(Overall Growth), UK2=(CO2+EG-LU) and UK3=(LU-Q-L-N). In the UKL20 BAU Model, UK1 and UK2 are unstable; stabilizing the Environmental Controller, UK2, would create a Moving Equilibrium Model.

EUL20 TECHP Model

The EUL20 TECHP (Productivity) state space model is on the edge of stability.

The time plot above shows the EUL20 TECHP (Productivity) model reaching a steady state around 2100.

UKL20 EU Input

The model is stable and cyclical. 

The columns of the Shock Decomposition above show the effects of EU1,  EU2 and EU3 on HARD1, FINZ1 and UK Growth. EU1-3 had (1) positive effects on HARD1 (decreasing unemployment), (2) Delayed positive effects of Financialization and (3)  Mixed effects on UK state variables, UK2=(CO2+EG-LU) being negative.

 

EUL20 Measurement Model

The EUL20 Measurement model is also an implementation of the Kaya Identity, measuring the overall state of the system. EU1=(Overall Growth), EU2=(CO2+EG-LU) an historical environmental controller and EU3=(LU-L-GDP) historical unemployment controller.

AIC Statistics

The Akaike Information Criterion (AIC) Statistics show overlapping confidence intervals and uncertainty about the best model for the UK.

What Should the Labor Party Do?

From Google AI:

The above graphic shows a very long list of things for the Labor Party to do in the future. If the list becomes larger or proves unwieldy, BAU is probably the best prediction for the future. Another option is for the Labor Party to begin preparing for a Steady-State Economy. Unfortunately, such a policy agenda is likely not realistic and will always be associated with Economic Stagnation.

 

Scaling Back Authoritarianism in China

May 19, 2026 China wants A.I. to Flourish but not at the expense of Jobs.

After the 2026 State Visit by Donald Trump to China, Xi announced a new Constructive Stability Strategy to prevent SuperPower confrontations.

We can debate (with experts) what this might actually mean, but let's assume that China has adopted the precise definition of Stability used in Systems Theory (see the Notes below).

The graphic above presents two future forecasts for CN1, the first state variable component of the CNL20 model. The upper graph shows the time path of the unstable system: Exponential growth forever. The lower graph shows the time path for the stabilized system: Steady State after 2152--a long time in the future. 

The graphic makes the point that growth in China can continue well into the future under a stable, less Authoritarian Economic System. 

The next question is: How much effort would it take to stabilize the system? The answer is that a relatively small change in growth rates would put the Economy of China on a stable growth path. Compare the "unstable" system matrix (F) in the Notes below to the "stable" System Matrix. The diagonal elements of F (the growth rates) have been lowered only slightly to create a stable system. And notice that CN2, the Malthusian-Unemployment controller  is still quite reactive (above unity). In other words, the Authoritarian System can still keep a tight control on Unemployment and stabilize overall growth.

China can evolve into a stable system without practical limits on the overall Growth rate of the System!

If this is what the Constructive Stability Strategy actually means, it is quite doable. A similar stability program could be applied to the US_LM model** but the results would be very different, creating a wildly cyclical system. For this (and many other reasons) I would not expect the US to cooperate in Constructive Stability Strategy regime.

Notes

System Stability is determined by the Eigenvalues of the System Matrix, F.

** Instructions for stabilizing the US_LM model are contained in R-code on the Google Site. Also notice that the  US_LM model is not a Malthusian model; control of employment involves the Export Market. But, there are Malthusian overtones (see the US_LM Measurement model below). Similar instructions are available for the CN_LM model and are available in the Google site (here).

For more posts see Blog Roll: China.

CNL20 Measurement Model

Three component state variables explain 99.4% of the variation in the indicators: CN1 = (Growth - LU) is an historical growth controller, meaning that growth is directly controlled in an Authoritarian economy, compared to the USL20W model where overall growth does not have an historical controller. CN2 = (LU - L) is a Malthusian-Unemployment Controller. CN3 = (KOF-E) is a Globalization-Unemployment-Energy controller.

Indicators are taken from the World Development Indicators, KOF is the KOF Globalization Index, EF is the Ecological Footprint and HDI is the Human Development Index (notice that the HDI is not heavily weighted in any of the state variables). The Measurement Model and resulting Historical Controllers are statistically estimated from data using Principal Component Analysis.

CNL20 Unstable System Matrix

The Malthusian-Unemployment Controller CN2 = (LU - L) is the main source of system instability.

CNL20 Stable System Matrix

The system is stabilized by slightly reducing the values of all diagonal elements in F (the growth rates). 

US_LM Measurement Matrix

The US_LM Model Measurement Matrix contains three component state variables: US1 = (Overall Growth), US2 = (X+XREAL-N-L) Export-Population Controller and US3 = (L + XREAL - X- N) Employment-Export-Population Controller. US2 and US3 have Malthusian overtones since Population (N) is involved in each historical controller. Notice that the US_LM Model is not an Authoritarian System because growth does not have an historical controller (as it does in the  CN_LM model).

You can run the US_LM Model on my Google Site (here). Instructions in the R-code explain how to stabilize the model.

 

Has Human Development Stagnated?

 

The 2025 UN Human Development Report notes that the Human Development Index (HDI) has been stagnating for most countries in the World-System. A forecast from the WL20 Model (above) makes the same prediction with narrow 98% Prediction Intervals. This post presents my own version of the HDI index and suggests that a Steady State may not be the best forecast for the future.

The Human Development Index is likely to be driven, in the future, by the World System which is currently in Growth-and-Collapse model.

Stagnation-and-collapse will not be a believable forecasts for optimistic commentators, but it is hard to see how Human Development will stabilize if the World System is in collapse mode. 

First, let me discuss how my version of the HDI (WL20_HDI) was constructed. I felt it was necessary to recreate the index and address criticism of the UN's methodology. One general criticism is that the HDI has not be located globally within the World-System (I do that in this post). From Wikipedia (here):

There have also been various criticism towards the lack of consideration regarding sustainability (later addressed by the planetary pressures-adjusted HDI), social inequality (addressed by the inequality-adjusted HDI), unemployment or democracy).

The removal of literacy from HDI has been criticized because educational attainment evaluates only the quantity of education but not the quality or the outcomes of education and can result in perverse incentives.

The indicators I have chosen to reconstruct the index (see Codes below) address some of these criticisms. From a comparison of WL20_HDI to the UN HDI (see below) both indexes show similar time paths over the Long Twentieth Century, although the UNDP does not not fit a model that can account for the approaching Steady State.

So, yes the HDI has stagnated or reached a steady state. On the other hand, the best model for the HDI is a Random Walk which suggests that future predictions are not certain and collapse is possible.

COMMENT: The intent of the UN HDI was to switch the focus from GDP, as a measure of progress, to "Human" measures of progress. From the standpoint of Systems Theory, however, neither is adequate without a measure of System State. And, what is interesting about the WL20_HD, currently, is that the State of the WL20 Model is not the best predictor. 

Notes

From Wikipedia:

The Human Development Index (HDI) is a statistical composite index of life expectancy, education (mean years of schooling completed and expected years of schooling upon entering the education system), and per capita income indicators, which is used to rank countries into four tiers of human development. 

 

Country Distribution of the HDI Index

The highest level of HDI is in Norway while the lowest is in Niger.  The World System is right in the middle while the US is in the top of the distribution (but has been loosing ground over the Long Twentieth Century). Notice that the UN HDI index is highly correlated with my HDI index and also appears to be reaching a steady state.

WL20 HDI Measurement Model

Two components of my constructed HDI index (constructed with Principal Components)  explain 99.8% of the variation in the indicators. HDI1 is a measure of overall growth in the indicators. HD2 is an historical controller balancing Life Expectancy against growth in the other indicators.

The WL20 HDI Index is largely unidimensional with the second component (HD2) acting as an historical controller before 1980 but not afterward (time plot above).

WL20 HDI Codes

The definition of codes in my WL20 HDI index is presented above.

WL20 HDI BAU Model

The WL20_HDI BAU model is stable and, from the graphic at the beginning of the article, reaches a steady state around 2100. The growth rate for my HDI index has been slowing since after 2000.

WL20 HDI Model AIC Statistics

From the Akaike Information Criterion (AIC) alone, the WL20_HDI Index is best seen as a Random Walk, meaning that in the future, anything can happen.

Since the Growth Component (W1) of the WL20 Model is expected to peak before 2050, the World System is predicted to be in growth-and-collapse mode.  Even though the  World System is not the best predictor of the HDI Index, that could change in the future.

Video