Evidence-based politics: would such a thing ever be possible?

[butchersapron]

There's a 3rd example.

 

[Paulie Tandoori]

blah. will come back to this tomorrow. but rubbish for now. it's stererotyping by another name, or finding the root cause or something probably.

 

[Pickman's model]

I'm not comparing human society to a system. I'm saying it is a system. Go on and tell me why it's not.

how is it a system?

 

[Pickman's model]

I'm not comparing human society to a system. I'm saying it is a system. Go on and tell me why it's not.

seems to me the only thing about emergence on this thread is the stream of excrement emerging from teuchter's keyboard

 

[teuchter]

how is it a system?

Systematically.

 

[Paulie Tandoori]

 

[IMR]

Evidence-based politics? Nice idea, but can't see it happening any time soon. Does anyone really come to their political beliefs as a result of genuine curiosity about the world?

In the field of social policy, there probably will be a lot of advances in understanding, but over many small, tightly-defined fields, rather than at the level of some Grand Theory of society.

Similar to advances in artificial intelligence maybe. Predictions about computers that can become self-aware, experience emotions and so on still belong to the realm of science fiction. Instead, lots of humbler artifacts are becoming a little bit more artificially intelligent in modest ways.

 

[Jazzz]

Teuchter, your proposition is something of an oxymoron. Wherever we have agreement, a nice neat system - it ceases to be politics. The politics will always be about the things we disagree about. We aren't going to run out of them, there's an endless supply.

Politics also concerns the differing priorities that people will have even where there is agreement on the outcomes of strategy.

 

[teuchter]

I wonder what people make of this:

I’ve spent a lot of time arguing that government should be more evidence based, and that wherever possible, we should do randomised trials to find out which policy intervention works best. We often have no idea whether the things we do in government actually work or not, and achieve their stated goals. This is a disaster.

Paper on randomised trials of government policies

 

[butchersapron]

This isn't evidence based politics. It's evidence based appraisal of potential impacts of various public policy initiatives. All the politics is done beforehand (well a small bit is done afterwards) in the connected process of establishing who gets to decide what policies are put forward for testing, how they came to choose these proposals, how they came to be in a position to be be able to propose and decide on what policies are tested, who gave them the power to be in that position, what requirements (formal and informal) they need to pass to get into that position and so on.

 

[Crispy]

The question is about whether it becomes effectively impossible to usefully predict the behaviour of a system once it exceeds a certain level of complexity, or whether it just becomes more difficult.

It becomes more difficult, but exponentially so. Even though today's satellites and computers and weather stations are an order of magnitude more sophisticated than those 30 years ago, weather forecasting is still just as bad. And compared to human society, weather is an easy problem.

 

[teuchter]

It becomes more difficult, but exponentially so. Even though today's satellites and computers and weather stations are an order of magnitude more sophisticated than those 30 years ago, weather forecasting is still just as bad. And compared to human society, weather is an easy problem.

not sure that's actually true.

 

[teuchter]

From Met Office website:


Comparing forecast accuracy

We continue to increase our forecast accuracy through research, investment in satellite remote sensing and supercomputing technology.
For many years we have verified our forecasts by comparing forecasts of mean sea-level pressure with subsequent model analyses of mean sea-level pressure. These comparisons are made over an area covering the North Atlantic; most of western Europe, and north-eastern parts of North America. From this long-term comparison an average forecast error can be calculated.
The graph shows how many days into a forecast period this average error is reached compared to a baseline in 1980. This graph shows that a three-day forecast today is more accurate than a one-day forecast in 1980.

 

[Crispy]

Well I'll be damned.

 

[teuchter]

This isn't evidence based politics. It's evidence based appraisal of potential impacts of various public policy initiatives. All the politics is done beforehand (well a small bit is done afterwards) in the connected process of establishing who gets to decide what policies are put forward for testing, how they came to choose these proposals, how they came to be in a position to be be able to propose and decide on what policies are tested, who gave them the power to be in that position, what requirements (formal and informal) they need to pass to get into that position and so on.

I get that. But there is also the potential to inform politics, isn't there? There are ideological positions that are based on assumptions that certain things are true, without them really having been tested.

 

[Falcon]

There are ideological positions that are based on assumptions that certain things are true, without them really having been tested.

Tested how, exactly? I guess the problem is the inability to falsify, and unavailability of inductive reasoning. An evidence based politics would attempt the discovery of novel truths about a political system through inductive reasoning and the survival of attempts by controlled experiment to falsify its predictions about reality. Social sciences (such as politics) are compelled to employ deductive reasoning, and can't falsify. So all you are left with is the axiom-proof-theorem format, in which any unsoundness in the (unprovable) axioms propagate into the framework constructed from them, rendering it irreducibly indeterminate.

And the sorts of things we are interested in aren't direct, measurable properties of linear systems. They are emergent properties of chaotic, complex systems so even if you could devise some sort of experiment, it wouldn't be repeatable.

So - no. In my view, evidence based politics will never be possible. Unless I'm misunderstanding what you mean by the term.

 

[inflatable jesus]

I think some of the criticism on the thread is a bit wide of the mark. I don't think that "evidence based politics" would necessarily involve a devising a system were everything in the political world can be predicted and planned for. It could be more of a system of better analyzing the system we have.

Since I'm US based, the example that springs immediately to mind is the Romney-Obama tax question. Romney thinks the economy will recover faster and unemployment will be reduced by keeping the Bush tax cuts in place and possibly cutting taxes further at the expense of Social security. Obama thinks that the way to go is by taxing wealthier Americans more, keeping taxes on everyone else about the same and by using federal funds to employ more people. (Yes, it's horrifically simplistic I know but it's roughly true)

In reality the answer is that nobody really knows how this will play out. Even well collected data is limited and inconclusive. So the correct answer is that all predictions are guesses and nobody should pay attention to them.

Evidence based medicine is about trying to be up front about how certain we can be about the things we do in clinical practice. More often than not it's about finding the things that clinicians have always done that we can't prove actually help anyone. I think there are some fairly analogous areas in political systems and that good data collection is the way to decide what solves problems and what doesn't.

I'd love to see an equivalent of the American Heart Association giving letter grades to political statements. If giving aspirin during a heart attack is a 1a rated plan, well backed up with good trials. I'd love to see how a commission of serious career academic economists would rate the statements on the economy from the politicians.

 

[TruXta]

You can't do controlled experiments with real economies.