This archive contains the FREE climate-economy model, as documented in my thesis.

The FREE6-corrected subdirectory contains a binary version of the model, with one error corrected – the mixing time of the surface ocean with the atmosphere is set to 1.5 years. The FINAL TIME of the simulation is also set to 2100 to make simulation experiments more user-friendly when optimizations are not being performed. The number of control files included is reduced for simplicity. This is the version users will most likely want to work with.

The FREE6-original subdirectory contains the model and control files exactly as documented in my PhD dissertation.


Feedback Complexity in Integrated Climate-Economy Models

Thomas S. Fiddaman

Submitted to the Alfred P. Sloan School of Management in partial fulfillment of the
requirements for the degree of Doctor of Philosophy in Management

Assessing the economic and ecological impacts of climate change induced by human
activity has become a major activity with a substantial modeling community.
More than 20 climate-economy models have been developed to address different
policy questions. While these integrated models are quite varied, most share some
common assumptions and features. They typically employ a nested structure of
neoclassical production functions to represent the energy-economy system.
Technological potential is represented by elasticities of substitution, exogenous
rates of technological improvement, and backstop energy prices. Factor allocation
is myopically or intertemporally optimal. The impact of a carbon tax on the energy
system at a given time can often be reduced to a simple tradeoff between abatement
costs and emissions (though capital stock rigidities complicate the short-run
picture in some models). The major endogenous dynamics of these models involve
capital accumulation, atmospheric concentrations of greenhouse gases, and the
temperature of the atmosphere and ocean system.

These models draw heavily on the energy-economy models of the 70s and 80s, which
were motivated by energy security issues and explored the potential impacts of
increasing energy prices on economic growth. System dynamics models of that period
shared the same motivation, but sought alternatives to the assumptions of
optimization and equilibrium. They focused instead on disequilibrium dynamics and
feedback complexity, with behavioral decision rules and explicit stocks and flows
of capital, labor, and money.

This research builds on earlier system dynamics models of energy economy
interactions, creating a model that tests the implications of a number of feedback
processes that have not been explored in the climate change context. Among these
are endogenous technological change and boundedly rational decision making, with
perception delays and biases. Energy requirements are embodied in capital, and
energy production capacity depends on explicit capital stocks. The search for
optimal policies is decoupled from other decisions, and uses intertemporally fair
criteria. To enhance the link between this research and other studies, the model
is constructed so that an appropriate parameterization will recover the
neoclassical case found in models like Nordhaus’ DICE (1994).

The principal purpose of the model is to identify the structural features that
have the greatest implications for policy, and thus are worthy of further pursuit.
Experiments with the model indicate that depletion of oil and gas resources has
critical interactions with climate policy. The inclusion of learning-by-doing
and other path-dependent mechanisms suggests that abatement efforts will be more
effective and should be more stringent than models with exogenous technology
forecasts indicate. Inclusion of delays and biases from structural and behavioral
features of the energy system creates higher long-run emissions reduction potential
but imposes substantial constraints that prevent rapid reductions. Fair discounting
and consideration of intangible damages substantially raise the indicated abatement
effort. In both deterministic and uncertain cases, near-term inaction is a poor

Reflections & future directions

My todo list for the model includes the following:

1 – update the data & test calibration
2 – add sequestration (dumb way: extra carbon sinks; smart way: coordinate with explicit forestry and ocean layers)
3 – add age vintaging to the economic capital stock
4 – split into two regions
5 – split into economic sectors (industry, service, transport)
6 – add intermediate energy carriers (liquid fuels, thermal fuels, electricity, hydrogen)
7 – add adjustment costs in energy capital stocks
8 – add backstop liquid fuel (e.g. derived from more-plentiful coal or tar sand resources)
9 – add wage/profit taxes and recycling of carbon tax revenue
10 – upgrade climate, carbon sectors to match ocean thermal layers in climate to layers in carbon cycle, and make radiative balance explicit
11 – improve behavioral representation of depletion rent, e.g. to include explicit, adaptive expectations for backstop energy prices
12 – add learning curves in end use/conservation options
13 – add deliberate R&D (as supplement to existing learning curves and scale economies)
14 – add noise inputs (e.g. business cycle, natural climate variation)
15 – make permits more robust to double-spike emissions trajectories
16 – make retrofits non-costless with an explicit investment decision
17 – rethink/redo energy intensity & think about formal estimates against data, generate behavioral/structural rigidity tradeoffs
18 – split embodied energy requirements by fuel or carrier
19 – pursue robustness to extreme inputs, particularly in carbon uptake and climate radiative balance
20 – add more primary fuels (split oil/gas, add tar sands/kerogen, split hydro/nuclear, split new renewables)
21 – add non-carbon GHGs (methane, nox, aerosols, CFCs, etc.)

A little bit of 1 would be a good idea no matter what.

The goal of some combination of 3, 6, 8, 11, 18 is to eliminate what I perceive to be the one significant flaw in the model – that oil/gas prices rise much too high in the very long term. This flaw doesn’t really percolate through to the economic outcome because the associated volume becomes so low, but nevertheless is bad. The reason this happens is that oil/gas energy requirements are embodied in the capital stock, and thus can at best decrease at some exponential decay rate (set by the capital lifetime and retrofit rate). The resource extraction cost, on the other hand, increases super-exponentially, as it is asymptotically ~infinite due to the fact that the supply of oil/gas is finite and there is no backstop that is fully substitutable in the short term. Thus, even after prices rise high enough to shut down new oil/gas exploitation, the residual driven by the decaying embodied requirements is large enough to drive the resource to extreme exhaustion. In reality, this wouldn’t happen, because the residual oil/gas consuming capital would get abandonded and/or conversion of other primary sources (biofuels, coal) would create perfectly substitutable (though possibly expensive) alternatives.

2 might be quite easy. It would be a lot better if sequestration options were closely linked to the existing carbon cycle (as opposed to just tacking on extra sinks). Sequestration sinks should permit possibility of leakage, displacement of emissions, etc.

4 would be interesting for investigating the value/potential for
(a) strategic depletion – e.g. measuring security benefits of getting away from OPEC dominated fuels
(b) valuing stable climate as a form of foreign aid not easily appropriated by corrupt governments
(c) looking at technology transfer as a principal way of reducing emissions in the developing world – and enriching American companies at the same time.

5, 6, 12, 13, 20 increase the number of explicit technologies that can be discussed/tested in the model. 13 in particular might let you parameterize experiments with short term/incremental policies vs long term/revolutionary technologies.

9 helps look at budget implications of carbon policy, and test options like offsetting income taxes with carbon taxes, reinvesting carbon taxes in R&D, etc.

14, 15 are mainly important for further exploration of permit schemes.

17 is really basic research into what is really going on with the energy system. It’s my contention that the energy sectors of other energy-economy models can’t replicate history (e.g. the OPEC oil spikes) realistically, or, when they can, they do it by assuming that behavior doesn’t matter and that the underlying physics of the energy system is very rigid.

FREE 6 – zip archive of the original model; requires an advanced version of Vensim or the model reader. This archive has not been updated to reflect capabilities of newer Vensim versions, but should work fine. I’ll post a newer copy of the model as time permits.

Leave a Reply

Your email address will not be published. Required fields are marked *