Making policy for the reality invariant

I like to call a lot of voters reality invariant. I’m definitely not saying they’re stupid, just that their opinions are reflections of their internal world, and are invariant to reality. Policy makers are treating these signals as true statements of their view of the world and want to change the world to satisfy them. Because some people are reality invariant this is doomed to fail. This is my suggestion for the converse of the Lucas critique.

The Lucas critique says that policy made using past observations won’t be effective. If you change policy from that which produced the predictions you’re using to change policy, the policy change won’t work as expected. Imagine FIFA is corrupt (!) and a team bribed it to make headers worth double, because in past competitions it would have advantaged them. Because rules change in a predictable way any previous advantage might vanish as other teams adjust their playing style.

In contrast, voters being reality invariant mean policy changes won’t work because people don’t care about policy. If you make policy to change a real outcome people say they care about but don’t, then a effective policy won’t be successful. You’ll change the targeted outcome, but not please voters. This can be misinterpreted as failing and so policy must double down on reality, even as the reality invariant remain oblivious.

One of the most obvious is immigration. For many people around the  world national immigration controls are not an issue. Around 10% of people think it is an issue at any one time. In the UK, that figure is around 45%.



This is despite British people agreeing with the rest of the world about their local area. Britons think that immigration is a huge problem nationally, but if everyone went and visited each other they would find out it isn’t.



For over a decade tighter and tighter immigration acts have been passed to the point where the immigration question has been solved. Short of Ukraine or Turkey joining the EU, or the UK leaving, immigration controls cannot get any tighter. And yet the British public are still not satisfied. [1]

It’s that second graph you should worry about if you’re a policy maker. The British don’t care about immigration. They care about the perception of immigration. British politicians have been very effective at reducing immigration. Only 1% of those who want to enter do so. The public don’t care.

They’re reality invariant. You can change reality but it will have no effect on their voting habits or opinions. You can lament the rise of spin, and I do!, but there’s a reason it’s so important. We’ve reached the effective limit of immigration restrictions and the public are still not satisfied.

Inequality is another area where people are reality invariant. Various policy changes have turned the US into a more unequal country, even as European states have held of the worst excesses of liberalisation. The FT have results of a survey that shows nobody understands how unequal their country is. Americans believe their country is more equal than it really is, and Europeans think their countries are more unequal than they really are.

“The results of the study suggest that, in the political debate on income distribution, it is often not the facts that count but [perceptions],” says Professor Michael Hüther, director of the Cologne-based IW economic institute, which carried out the research.


In Hungary belief in widespread poverty is supporting the growth of Jobbik and Hungary’s slide into fascist politics. In America belief in egalitarianism prevents help going to the poor. The voters are reality invariant, they are discussing how they feel about reality, not reality.

Crime is the most famous example. Fear Of Crime™ is a major issue. Crime just isn’t.



The reality invariant voter looks at this and sees fiddled numbers. By 2007 crime was the number one concern for British voters, briefly displacing “immigration” before being engulfed by “the economy.” Despite violent crime falling in half, fear of crime continued to increase.

War is the last area I want to briefly touch on. War, what is it good for? Absolutely nothing, until it’s useful for absolutely anything. Intervening somewhere is a very bad idea, until it isn’t:

It doesn’t matter, how many times your Yazidis get off the mountain without significant UK aid or an entirely theoretical artillery seige of Bengazhi with rapacious house-to-house murders turns into an actual, person-killing, NATO-supported, civilians-running-everywhere seige of Sirte.  There’s always some compelling reason why this war, this intervention, this bombardment, is radically different to all the other invasions and occupations and airstrike campaigns, and why the horrible news always means that there’s no time for rational thought this time before we start blowing that shit up.

I’ve come to you with a problem, and a snappy term to describe it. Traditionally this is where a writer would offer a solution. But not me. Lucas didn’t have a solution, he had a critique. I am also merely pointing out that policy is today being made foolishly. The assumption of policy makers that changing reality will have the desired effect is based on a faulty model. It is a very rational model, but the model is wrong nonetheless. Lots of the public are reality invariant, and you can’t make policy presuming they’re not.


[1] As Chris points out, the reality is that these policies kill and kill intentionally and purposefully. But this is also a reality most people ignore.


5 thoughts on “Making policy for the reality invariant

  1. I don’t understand this post at all. You say, for example, that “fear of crime is a problem…crime just ain’t”. Who defines “fear of crime”? I lock my front door every night because I’m afraid of being burgled, though I’m perfectly well aware that the statistical chances of my being burgled are extremely low. Does that make me “reality invariant”? Or just sensible?

    The idea that risk should be assessed only through the narrow prism of statistical probability seems to me both naïve and unrealistic. Why should it be the case that the fears which preoccupy us are those which are most likely to come to pass? Is it not more likely that our most powerful fears are those attached to our sense of identity and our personal security, regardless of the likelihood of them coming to pass? For example, I’m not too afraid of having my pocket picked (relatively likely) but I’m very afraid of being beaten up (relatively unlikely). So, yes, I’m afraid of “crime”, and I’m unsurprised that most people, not being risk-accounting ciphers, feel the same.

    You seem to think that data trumps feelings, or at least that it should do. As for your idea that “crime ain’t a problem” I defy you to comfort victims of crime by pointing out that they are a statistical minority. It’s a problem for them.

    I would apply similar arguments to the other examples you cite – immigration, for example. I have no idea how you arrive at the conclusion that “the immigration problem has been solved”. Rather like “fear of crime” it all depends on how you define “the immigration problem”. My definition clearly differs from yours.

    But immigration is a big topic, so I’ll leave it there. Apart from one final point. You say that current British immigration policies “kill intentionally”. Are you really suggesting that the present Government introduced these policies with the specific “intention” of killing people? This is surely an overstatement. Or in other words, barking mad.

Comments are closed.