-3

Are public policies by nature unpredictable with certainty? Is it possible to predict with a likelihood that can be trusted, the effects of various proposed legislation and policies and executive orders without applying them and seeing their effects first hand?

Improve this question
3
  • 2
    It might help to narrow the question. Something like raising taxes or the minimum wage can be modeled by economists (the accuracy of economic models may be questionable but is certainly the subject of a lot of research). Other actions are likely to be far less clear - deciding to go to war or to back unproven technology - or it may be hard to know what outcomes to even consider (what are the effects of an abortion ban, beyond fewer abortions? Will corporations relocate, employment rates change, crime rates change, the Christian right feel emboldened and choose to make other demands...)
    – Stuart F
    Commented May 22, 2023 at 12:56
  • What do you mean by "can be trusted"? Like apart from a narrow domain of pure logic nothing can be said with 100% certainty, yet often enough even 60% certainty is enough for people to accept it as given, so where would you draw the line of what "trust worthy" results amount to?
    – haxor789
    Commented May 25, 2023 at 11:15
  • 1
    All predictions are hard, and the systems usually considered in politics are both large and open to outside input. To make any predictions, about the behavior of the system, you need to make assumptions about how the input will develop over time (migration, trade, information, climate, weather, geological and astronomical events,...).
    – Hulk
    Commented May 26, 2023 at 7:44

3 Answers 3

Reset to default
8

One can try. There are many examples of successful predictions, which rarely make the news. There are many examples of unsuccessful predictions, which sometimes make the news.

From time to time, legislatures require that proposed laws or regulations comes with formal assessments of their impact, e.g. in the EU. This may yield slightly better results than informal methods, and it makes sure the assessment is made.

The problem of "trustable likelihood," as you put it, is that history is an experiment with N=1. You try to estimate the number of primary school children next year -- should be easy, how many babies were born five years ago? -- and then Putin invades the Ukraine and women and children have to flee. You try to estimate next year's tax revenue, and a pandemic breaks out and causes lockdowns. That tax estimate might have been actually very good, sound scientific methodology, yet a pandemic was at best in a footnote of a footnote, "low probability events."

Improve this answer
0

History is the teacher of life. Who does not care to learn, repeats the lesson, till the one does.

It is often possible to predict the income and consequences by comparing the effects of similar orders and policies in other countries over the reasonably comparable historical range of time. It is unwise to implement the policy without checking first what happened in other countries, or the own one in the past, after the comparable policy has been implemented.

At very least, if the serious negative consequences are known, the policy proposal should include the measures to avoid the known negative effects.

Improve this answer
-2

A good actuarial response always includes the disclosure: "This prediction is at best only as good as the data we have, and is limited by the degree to which past experience reliably predicts future events".

There will always be unintended and unforeseen consequences of public policies among mortals. The only way to make the effects of policy exactly predictable would be to turn everybody into robots.

One can of course choose what and whom to trust, and make one's own judgments based on knowledge of the track record.

Improve this answer
4
  • Even turning "everybody into robots" would only improve predictability in a limited degree: emergent behaviors of large groups of agents are hard to predict, even if all agents are fully deterministic.
    – Hulk
    Commented May 26, 2023 at 7:15
  • There are a few inspiring simulation experiments with ants, termites and other state-forming insects, demonstrating how the layout of the structures they build can be encoded in very simple deterministic rules. Still, even under laboratory conditions, it is almost impossible to caclulate the outcome in advance, without actually "running" the simulation experiments.
    – Hulk
    Commented May 26, 2023 at 7:31
  • @Hulk I disagree. Emergent behaviors are relatively easy to model via simulation. Certainty might be obviated by such factors as numerical imprecision or inadequate simulation configurations, but software is actually very easy to model, manage and control using software.
    – pygosceles
    Commented May 26, 2023 at 14:21
  • 1
    Well, we actually agree then - it requires simulation. We 'just' disagree how hard that is to get right :-)
    – Hulk
    Commented May 26, 2023 at 14:57

You must log in to answer this question.