Computer Simulations of Participatory Planning


Is it possible for workplaces and neighbourhoods to democratically plan an economy themselves instead of relying on market competition or central planning?

Our annual participatory planning procedure provides a possible answer to this question. We describe how worker and consumer councils might cooperate to coordinate their interrelated economic activities themselves — fairly, sustainably and efficiently (read more about Participatory Planning here). But since no country has yet tried our alternative to markets and central planning, we had to resort to computer simulations to explore if participatory annual planning would work in practice and not only in theory.

Project Goals

This project is an effort to simulate the participatory annual planning procedure to test how practical, and how robust it is likely to be. To be “practical” a feasible plan must be reached without requiring councils to make too many proposals and revisions. To be “robust” it must reach a feasible plan even when productive technologies and consumer preferences violate assumptions economists commonly use in our theoretical work.

Project Members

  • Robin Hahnel (Professor of Economics)
  • Mitchell Szczepanczyk (Computer Programmer)
  • Michael Weisdorf (Computer Programmer)

About the Simulation

Initially we used Netlogo to simulate annual participatory planning. But since Netlogo limited the number of councils and goods we could model, we moved on to use Clojure and Clojurescript programming languages to simulate an economy that can approximate the size and complexity of a modern, national economy.

The agents in the simulations are:

  • Self-managing Workers Councils
  • Self-managing Neighborhood Consumer Councils
  • Federations of Worker Councils and Consumer councils

Worker councils and consumer councils live as data structures whose production functions and utility functions are randomly generated. The simulation proceeds as follows:

  1. To begin, an Iteration Facilitation Board announces an initial estimate of the indicative price of each kind of capital stock, each consumption good, each intermediate good, each natural resource, and each category of labor.
  2. Neighborhood councils then attempt to maximize their satisfaction from consumption subject to an income constraint of its members.
  3. And worker councils attempt to maximize their satisfaction from work subject to a constraint requiring them to produce outputs whose social value is at least as great as the social cost of the inputs they use.
  4. The Iteration Facilitation Board then calculates demand and supply for each good, and raises the “indicative price” when demand is greater than supply and lowers the “indicative price” when supply is greater than demand. 

This process repeats for as many “iterations” as are needed to reduce the excess demand for every good below a specified “threshold.”

Key Results

We generated forty “experiments” with:

  • 30,000 worker councils,
  • 30,000 consumer councils, and
  • 100 “goods” (20 private consumption goods, 20 intermediate goods, 20 capital goods, 20 different inputs from the natural environment, and 20 different categories of labor).

For a full explanation of our procedure and results interested readers should see chapter nine in Democratic Economic Planning (Hahnel, 2021).

To make a long story short, so far the results have been highly encouraging. Under standard assumptions about preferences and technologies it took only six and a half iterations to reach a feasible plan on average – which could easily be done in the month of December, yielding a new annual plan ready to go on January first.

But perhaps what was even more encouraging was that it never took more than 8 iterations before a feasible was reached. And, even when we relaxed assumptions about technologies and permitted twenty percent of worker councils to experience increasing returns to scale, not only did the planning procedure never break down – which was our principle concern – but it converged just as quickly. 

Further Experimentation

1. Other functions besides Cobb Douglas. Cobb Douglas functions  are convenient for several reasons, which is why they are commonly used by economists. But they generate constant elasticities of demands for all goods, in contrast to real world economies where different goods can have very different elasticities.  We plan to replace our Cobb-Douglas production and utility functions with different functions to see what difference this makes.

2. Improved price adjustment rule. The price adjustment rule we used was arrived at with little effort.  Despite the encouraging results, we are searching for an even more efficient price adjustment rule. 

3. Environmental impacts.  We have already incorporated an incentive compatible pollution demand revealing mechanism into our theoretical model. It remains to do so in our simulation work.

4. Further robustness testing.  We will continue to test for robustness by violating even more assumptions to see at what point the procedure breaks down.


The fullest test of a democratically planned economy would be to implement one in some country in the real world.  Absent that, we have pursued the next best option: Implementing such an economy as a computer program to address reasonable concerns about our proposal’s practicality and robustness. 

The computer simulation results builds confidence that our annual decentralised democratic planning procedure is both robust and practical, taking only six and a half iterations to reach a feasible plan on average.

So far the assertion voiced first by Alec Nove in 1980, and repeated by many others since that “there is no alternative to markets or central planning… there is no ‘third way’” … seems to be very premature, and quite possibly very misleading.

Watch a video presentation of the project — a detailed exposition of the program and its full context:

Researching “third way” democratically-run economies with Clojure (Script)

Further Information

  • Data Files – The experiments referenced in the book “Democratic Economic Planning” are available as gzipped Clojure data files online: