How to resolve conflicting production plans?

Questions
1

Dear all,

i wanted to follow a suggestion and (re)post a discussion that arose following Anders @Anders @Jason workshop during the The crisis of nation-states event.

Essentially, i was wondering about the following the situation:

Given are two production federation that submit production plans that produce two roughly equally important, but possibly very distinct, products. However, both rely on a critical capital good (say a factory), whose use is mutually exclusive. Hence, there is a conflict over the use of a scarce capital good. How do we resolve such conflicts? How do we decide, who is assigned the use of the factory in the planning period? Is there any formal procedure how to adress such issues?

My feeling is that such situation are relatively frequent and normal. If, we do not have a gross oversupply of production facilities, the means of production should be operating near their full capacity. In such a case it could easily happen that such a resource is requested beyond capacity and decisions on their utilization have to be made. Hence, i think this should be process that should follow some kind of streamlined procedure. I have a few opinions on this myself but id be happy to hear the forums feedback, how people see the issue here

2 Likes
2

The only way for everything to be accepted in the end is if all supply covers demand. Prices will update and another round of planning begins if demand is higher than supply for any good.

So let’s say the demand for the capital good is higher than supply. When the IFB analyzes the current round’s data it will see this and increase the price of the capital good. Producers will then respond to this in the next round of planning. Either enough people commit to using something else so supply covers demand, or the price of the capital good will eventually become so high that no one can use it. It would probably take many rounds of planning for the last scenario to happen. In most cases, enough people will move away from the capital good so that supply covers demand.

Also, remember that higher input costs make it more difficult for production proposals to be approved. Workplaces that are asking for this capital good would either need to increase the value of their outputs or use less costly inputs to be approved.

Does that make sense? I hope it helps!

1 Like
3

This is how I could see it working:

One possibility could be that for any capital assets which are unique (cannot be grouped into categories with other substitutable assets), like a factory that you mention as an example, they could be bidded on, on an individual basis, during the annual planning. The price would go up until only one workplace is left with a Social Benefit > Social Cost requesting that asset that then gets access to it. But I don’t think that is practical for several reasons.

I think each workplace that currently uses a unique asset by default could simply have the right to continue using it. Other workplaces then would not be able to bid for specific capital assets that are already in use by other workplaces. For example, if I am in a workplace and we are using a building or factory unit, then we continue to use it and nobody else can bid on it. As long as we submit a production proposal which has a social benefit greater than the social cost, our proposal is accepted and we retain use of the building.

This does raise the issue if there is no other demand placed on the asset, how is access to these capital assets priced? Perhaps there could be fixed term leases, and there is bidding when the lease term is up. That’s one possibility but I think Anders has a better suggestion in his book on how to calculate a fee.

The other question is what happens with new buildings which become available. I think this where the industry federations have a role to play. New startups, if approved by the federation, could be allocated an available capital asset, e.g. a building unit, by the federation. The procedure could depend on the specific industry.

Anyway, some of my thoughts.

1 Like
4

Dear all, sorry for the late reply, but it always takes some time for me to collect my thoughts and give a proper answer.

@MichaelHicks, i agree with your argument that everything being accepted is if supply covers demand. Clearly, iterative updates of prices can help to guide society to this point over time. However, this process may take some time and i think that it does not entirely adress the question of how-to resolve these conflicts in the short-run. It may be the case, that two conflicting production plans can be fulfilled in future planning periods with updated prices. Nontheless, the current planning period needs to be resolved before that and we need a way to decide how to allocate the resources in the short-term too. Id be especially interested in thoughts about this short-term conflict resolution, since its least clear to me how we would proceed on that one.

I am kind of keen on this question, since i am convinced that these type of conflicting usage proposals will rather be the norm than the exception.

The reason for this being, is that one of the stated goals of participatory planning is efficiency.

I think, that any efficient production plan will have to result from a mathematical optimization calculation. Now, depending on the exact planning procedure the involved optimization may differ. However, most proposals for such procedures argue for the usage of linear optimization (see a quick intro video on linear programming to get an idea of what im talking about). The theory of linear optimization directly tells us a few interesting things, about how a efficient production plan has to look like.

A mathematical optimization problem consists of two parts:

1a. A quantifiable “Goal”, usually called the objective function.
1b. A specification wether want to “maximize” or “minimize” our goal. A sensible goal would be to minimize worked hours or CO2 emission, today it frequently is to maximize profits, but it could also be to maximize the output of production.
2. A set of quantifiable “Limits” to the goal, usually called the constraints. Constraints, can be a large variety of things like how much machines can produce, but also minimum requirements of the planning such as that each person requires at least a minimum provision of nutrition or housing.

Now, given any formulation of goals and limits into a set of equations the theory of linear optimization allows you to calculate how-to balance those most efficiently.
However, it also tells us that any efficient solution has at least one, most likely several, constraints that need to be fully exhausted. Otherwise, we are not making full use of our resources.

If these constraints model productive resources, then it is sensible to think that we should expect some level or conflict about it. Also, if efficient production plans need to involve some level of exhaustive use of limitng factor, id expect this question arise in every planning step by design.

1 Like
5

Hey, no worries about the late reply! Thanks for sharing your thoughts!

The iterative updates happen within a small period of time. So when you talk about solving these conflicts in the short-term, the whole process we’re talking about would occur over a few weeks.

There have been computer simulations that address the practicality issues (check the news section of this site for a presentation and source code), and if I remember correctly the results were averaging around 10 – 20 iterations. So to be clear, all conflicts are being resolved in these simulations fairly quickly.

Regarding the efficiency part of your post, each workplace is required to submit a production proposal that has a social benefit/social cost ratio greater or equal to one. Because of this requirement, efficiency is achieved since workplaces must maximize the benefits of what they’re producing to have their plan accepted.

I’m a programmer myself, and I’m familiar with linear programming… but I feel like linear programming is more relevant to centrally planned economies where you have one actor (the government) trying to calculate the most efficient plan and then assign production targets to workplaces. The goal of democratic planning is to simply provide info actors need to make efficient decisions themselves.

Best,

Michael

6

This sounds feasible to me. It’s sort of a back end issue, that is, it becomes a question of which Worker Council(s) will get access to the factory or capital asset. If there is excess demand for the factory then it would drive the cost of this means of production higher. This input is factored into the production cost for each WC that wants access to it. There would then be corresponding increase in the Indicative price of the good(s) being produced by the WC(s). This IP goes back up to the IFB, then to the consumers who are requesting these goods. When consumers see the prices go up, some may opt out and spend their money on some other good. This information is then fed back to the IFB and then the WC(s). At some point, there should be an equilibrium where WC1 may end up accessing 40% of the facility and the other 60% is accessed by WC2, depending on how demand plays out on the consumer side.

This scenario assumes that there is no alternative way of producing the good.

If one of the WCs does find an alternative way of producing the good, then WC1 may get 100% access to the factory and WC2 would be accessing some other capital asset. In that case, then I would see the Indicative Price be more at the breakeven for WC1 where SB=SC. If the alternate production facility for WC2 is more expensive then this would also be factored into its Indicative Price. If the IP is too high, demand for the product may drop.

Regardless of the scenarios, long term planners may want to keep an eye on this and consider investing resources to build additional production facilities to meet overall demand for both products.

I agree with CCC that this would likely be a common occurrence.

C

1 Like
7

CCC said…

Nontheless, the current planning period needs to be resolved before that and we need a way to decide how to allocate the resources in the short-term too. Id be especially interested in thoughts about this short-term conflict resolution, since its least clear to me how we would proceed on that one.

Dear CCC

see my post below, the Indicative Prices would go up likely given the capital asset cost would go up due to demand, and consumer demand may go down for a product if the Indicate Price gets too high. this is how it would be resolved in the current planning session.

C

1 Like
8

Thanks for bearing with my slow responses :wink:

In my opinion mathematical programming is useful to any type of planning wether it is central or democratic. Hence, i wouldnt really see it as particularly characteristic to central planning. Personally, i would even believe it to be necessary for any kind of efficient planning on a larger scale. How else would individual actors process the information about a whole economy?

Id say the difference between central and democratic planing should rather be in the way to use mathematical programming. For instance, i would say that the difference should be that instead of the government federations would rather be defining objective functions and constraints.

Id say that there are even advantages for participation. Fo instance, we could simply phrase goals (“Our goal is to work little”, “Our goal is to increase comfort of living”, etc…) and limitations (“We dont want to work longer than 8 hours”, “We only want to emit 10t of CO2”…) rather then requiring us to submit detailed micro-plans.

Moreover, my feeling is that simply calculating wether a plans social benefit is larger than social costs might miss out on some more efficient outcomes. For instance, two federations might propose two plans that are overall beneficial. Maybe one yields a higher cost-benefit ratio and is then chosen, but it is thinkable that an even better plan, that works for everyone, was overseen and not proposed. Things like this might be avoided by the usage of optimization.

All the best and thanks for listening,
ccc

2 Likes
9

No worries, you always seem to reply on my lunch break so I’m quick to reply :stuck_out_tongue:

I agree, but it depends how it’s being used I think. I don’t think there should be any kind of higher level agency making anyone do anything. The goal of the higher level agency should be to provide the information people need to make efficient decisions themselves. The way individual actors would process information about the whole economy would be through price signals (which are basically generated by computers crunching data).

If the workplaces want to utilize technology to generate their own plan, I think that’s great! I’m also in favor of using technology to automate these things as much as possible. I also think it’s fine if the IFB generated suggestions for the actors based on calculations for what they think is most efficient. However, I think it’s very important that the only thing we’re doing here is providing information or making suggestions. Having an agency that is assigning or forcing people to do what they think is best is a core characteristic of central planning.

Thanks for the great discussion,

Michael

1 Like
10

Dear Michael,

We are absolutely on the same page here. The thought of an LP making the final economic decisions within a society seems deeply authoritarian.

Yes, i definetly think optimization algorithms (or IFBs employing them) should have the possibility to propose plans alongside federations, since i think that they might traverse a broader range of possible plans than individual federations.

I can see how for instance specialized tools or raw materials that are relevant only to a few key industries can well be managed by an individual federation.
However, some resources (like electricity or labour) are required virtually everywhere.
Properly, analyzing the tradeoffs of such resources will benefit from an overall optimization and may generate new plans that probably wouldnt occur from the perspective of a unique federation.

The right to accept proposed plans, however should definietly not lie in the hands of an IFP or algorithm.

The role of mathematical programming should be solely to (i) allow formulation of goals and limits as objective function and constraints, (ii) critically evaluate proposals,
(iii) generate information, (iv) highlight conflicts and possible solutions
and (v) propose plans.

I think the workflow of an IFB advisory round should be roughly like this

  1. Federations and indviduals use some external decision making process to select constraints and objective functions. (Note: Most likely we might have a multi-objective problem making things less easy to model)

  2. Federations optionally submit production plans (It would also be fine to submit no plan, but some general guidelines like have 100 people working up to 8 hours and leave planning to others!)

  3. IFB collects the submitted objective(s) and constraints and tries to calculates optimal plan(s) according to this specifications
    3a.The problem may be infeasible (no proposed plan can meet the formulated conditions), meaning that society had unrealistic expectations. The IFB reports this back.
    3b.The problem may be feasible (the problem can be solved optimally by at least one plan). In linear &convex programming, uniqueness is quarunteed. Most likely, we will have to face problems with integer or non-convex constraints with possibly many solutions. The IFP computes (if computationally possible) a set of optimal plans and proposes them.

  4. The IFP computes the objective value of all feasible plans (including proposals from federation and its own). This quantifies to what extent differing plans attained the set goals. IFP reports this to everyone, so we can make an informed decision on the plans.

  5. Moreover, the IFP determines all so-called “active” constraints for both feasible and infeasible plans. These are the constraints that effecitvely limit our goals. This is valuable information that needs to be fed-back to society. For a feasible plan, this effectively tells us things like: “Our goal is to work little. We can do everything we want with 6hrs of work. The reason we can not work even less, is that our constraint to limit CO2 emissions and the lack of a factory of type xy limit our goal.” For an infeasible plan, we know that some of the active constraints need to be relaxed to reach feasibility. In this case, we had unrealistic expectations and IFP can tell us which expectations might be necessary to drop. Furthermore, all active constraints constitute a production plan conflict to some extent.
    This automatically lists all federations that may need to compromise and who they have to compromise with.

  6. Each active constraints is associated with a so-called dual variable. The IFP computes these and reports them. This is very interesting information, since it tells us how much we would gain if a constraint would allow for one more extra unit. Essentially, dual variables quantify how bad a conflict is. Large dual variables should be considered as an indication, that a given constraint is relevant and needs extra attention.

  7. The IFP then publishes all active constraints and dual variables. It may recommend constraints (that essentially constitute a resource conflict) for negotiation and informs the submitting federations what society could gain if they can compromise on a given constraints. There may or may not be a formal mediation process to accompany this.

  8. This information can be used to either accept a feasible plan through some kind of society wide decision making process, or to make an informed decision when proposing a new plan in case of either (i) infeasibility or (ii) too large dual variables.

In summary, i think that mathematical programming gives a very nice framework to structure the work of an IFB. It even automatically highlights whose needs are in conflict with whom and to what extent. I always felt this can be the basis for a nice “mediation” procedure. @Jason, i think this might also be of interest to your project of modelling the specific roles within a PARECON UI.

all the best,
ccc

11

Great discussion everyone. Robin Hahnel received a grant to build a computer model to simulate the IFB process. He contracted out the work to Mitchell Szczepanczyk who produced a simulation model. you can view and try it out by following this link…

http://www.szcz.org/pecs.html

Also, at some point, decisions about what to produce won’t solely be made by using mathematical models or algorithms, but recall that qualitative data is collected from consumer proposals and probably from worker council proposals. This qualitative data can be used to enhance decision making. At whatever level in society that we have IFBs, they will use this additional information to take decisions. The manner in which they reach decisions would be agreed to by each IFB which may be one person one vote, consent or consensus. It remains a democratic process nonetheless and not an autocratic, top down model.

Cheers

C

2 Likes
12

That’s a good summary I think, and recommendation from here as well to check the very interesting simulation work compiled by Mitchell.